Home

Report - AAA Foundation for Traffic Safety

1. 10 8 6 4 l ll 0 q T Not Conversing Reference Conversing m No Hands m Eyes In Adverse Events Hour Conversing Conversing with another passenger in the vehicle had little impact on our three outcome measures Drivers were only slightly more likely to take their eyes off the road or to experience an adverse vehicle event 57 Figure 17 Internal distraction effects on driving performance 25 20 KK 15 7 10 5 0 nn No Internal Manipulating Reaching etc Other Internal Distraction Controls Reference E No Hands W Eyes In Adverse Events Hour p lt 01 Internal distractions Reaching for objects inside the vehicle manipulating vehicle con trols and other forms of internal distraction were all associated with higher levels of drivers taking their hands off the steering wheel and eyes off the road Only reaching distractions however were also associated with a significantly higher rate of adverse vehicle events wandering encroaching into another travel lane sudden braking Of the three categories of internal driving distractions reaching for objects inside the ve hicle appears to pose the greatest risk Figure 18 External distraction effects on driving performance
2. 20 15 10 5 o m _ B No External Distraction Reference External Distraction m No Hands Eyes In Adverse Events Hour External distractions Although external distractions were associated with higher rates of adverse vehicle events the results were not significant statistically and there was no measurable effect on the other two outcome measures 58 Taken together these results suggest that a number of activities commonly occur ring while driving may adversely affect driving performance Dialing or answering cell phones are one such activity but eating and drinking manipulating music controls reading grooming reaching for objects inside the vehicle and manipulating other con trols in the vehicle all were associated with negative driving outcomes What is not known is the extent to which the driving outcomes measured by this study hands not on the steering wheel eyes not on the road and adverse driving events such as wandering in the travel lane or encroaching into another travel lane translate into actual increases in crashes The issue of crash risk is also addressed in the following section although again without drawing any definitive conclusions Comparison of Phase and Phase II Findings One of the original goals of the project was to combine the results of the crash data analysis carried out in Phase I of the study with the exposure data collected in Phase
3. 48 Table 12 Percentage of time vehicle stopped within levels of distraction variables of Time Potential Distracting Event Vehicle Stopped Cell phone pager Phone not in use Dialing phone Answering phone Talking listening Eating or drinking Not eating or drinking Preparing to eat drink Eating drinking spilling Music audio Music audio not on Music audio on Manipulating audio controls Smoking Not smoking Lighting or extinguishing Smoking Reading writing Not reading writing Reading writing Grooming Not grooming Grooming Occupant distraction No occupant distraction Distracted by baby Distracted by child Distracted by adult Conversing Not conversing Conversing Internal distraction No internal distraction Manipulating vehicle controls Reach lean look for pick up etc Other internal distraction External distraction No external distraction External distraction Consequences of Distractions on Driving Performance Since our primary concern is the effect of distractions on driving performance the analyses carried out and presented in this section are restricted to only those segments of videotaped driving where the vehicle was in motion As described in the methodol ogy section and in Appendix D confidence intervals for testing significant differences in estimated proportions were developed using bootstrap percentile methods There were three primary outcome variables of interest that could be ass
4. 100 Table F 3 Percentage of time engaged by a potential distraction within categories of weather conditions Weather Conditions Distraction Overall Good Bad Talking on cell phone 1 2 3 3 1 3 Dialing answering cell phone 0 2 0 6 0 2 Eating drinking 1 4 3 5 1 5 Preparing to eat or drink 2 8 9 9 3 1 Music audio on 71 5 70 0 71 4 Manipulating audio etc 1 1 1 9 1 1 Smoking 1 9 0 8 19 Lighting extinguishing 0 04 0 03 0 04 Reading 0 8 0 7 0 8 Grooming 0 3 0 7 0 4 Internal distraction Reaching 2 2 4 2 2 3 Manipulating controls 1 3 1 6 1 4 Other internal distraction 1 3 3 1 1 4 External distraction 2 4 0 9 2 3 Other vehicle occupant Baby 0 4 0 5 0 4 Child 0 3 0 1 0 3 Adult 0 3 0 1 0 3 Conversing 15 6 13 2 15 5 Overall 95 4 4 6 100 101 102 Table F 4 Percentage of time engaged by a potential distraction within categories of road type travel lanes Road Type Distraction Neigh 2 3 Multi Inter Other Overall borhood Lane Lane Divided state Talking on cell phone 0 9 0 7 08 14 31 2 0 1 3 Dialing answering cell phone 0 3 0 2 02 02 03 04 0 2 Eating drinking 0 7 1 4 20 10 20 1 8 1 5 Preparing to eat or drink 1 1 4 2 26 2 3 19 2 1 3 1 Music audio on 66 0 71 9 68 9 75 2 75 4 67 6 71 4 Manipulating audio etc 1 5 1 2 0 7 1 0 1 3 0 9 1 1 Smoking 1 3 1 2 14 3 2 1 8 1 0 1 5 Lighting extinguishing 0 1 0 03 0 1 0 04 0 1 0 02 0 04 Reading 1 3 0 4 08 05 02 2 3 0 8 Grooming 0 2 0 3 0 8 0 2
5. DISTRACTIONS EVERYDAY Prepared by Jane Stutts John Feaganes Eric Rodgman Charles Hamlett Thomas Meadows Donald Reinturt University of North Carolina at Chapel Hill Highway Safety Research Center Prepared for Kenneth Gish AAA Foundation for Traffic Safety Michael Mercadante 607 14th Street NW Suite 201 Loren Staplin Washington DC 20005 Tel 202 638 5944 Fax 202 638 5943 www caafoundation org TransAnalytics LLC June 2003 This study was funded by the AAA Foundation for Traffic Safety in Washington DC Founded in 1947 the AAA Foundation is a notfor profit publicly supported charitable research and education organization dedicated to saving lives by preventing traffic crashes and reducing injuries when crashes occur Funding for this study was provided by voluntary contributions from the American Automobile Association and its affiliated motor clubs from individual AAA members and from AAA affiliated insurance companies This publication is distributed by the AAA Foundation for Traffic Safety at no charge as a public service It may not be resold or used for commercial purposes without the explicit permission of the Foundation It may however be copied in whole or in part and distributed for free via any medium provided the AAA Foundation is given appropriate credit as the source of the material The opinions findings and conclusions expressed in this publication are those of the authors and are not
6. and reaching for objects inside their moving vehicle Nearly as many were observed manipulating music audio controls or had their attention drawn to something outside the vehicle Approximately three fourths ate or drank something while driving or con versed with a passenger Reading writing and grooming activities were also relatively common but declined to less than half of the participants when observations were restricted to moving vehicles only About a third of the subjects used a cell phone while driving and nearly as many were distracted by passengers riding in their vehicle Tak ing into account the shorter amount of time that children and especially babies were carried in vehicles children were about four times and infants almost eight times more likely than adults to be a source of distraction to the driver based on number of dis tracting events per hour of driving Percentage of drivers engaging in potentially distracting activities during three hours of driving and percentage duration of these activities when their vehicles were moving Potential Distraction of Subjects of Total Driving Time Talking on cell phone 30 0 Answering cell phone 15 7 1 30 Dialing cell phone 27 1 Eating drinking spilling 71 4 1 45 Preparing to eat or drink 58 6 3 16 Manipulating music audio controls 91 4 1 35 Smoking includes lighting and extinguishing 7 1 1 55 Reading or writing 40 0 0 67 Grooming 45 7 0 28 Bab
7. into the datafile so that it could accurately track the durations of the various states being coded for each of the 70 subjects The larger problem was how to deal with the event codes which were single point in time occurrences as opposed to the state codes which were toggled on and off over some duration of time The initial SAS datafile created was actually a file of nearly 7 5 million observa tions corresponding to 207 2 hours of data coded in 1 10 second intervals In examin ing one of the coded states for example cell phone use we simply tallied up all of the various 1 10 second intervals in which subjects were either not talking on the cell phone talking on the cell phone dialing or answering a call This variable distribution could be crosstabulated by other variables that were coded as states for example hands on or off the steering wheel and by driver age and sex However the event outcomes vehicle wandering vehicle encroachment into another travel lane and sudden brak ing had to be handled differently essentially these were treated as count data without any associated duration Eventually we were able to reduce the overall size of the file by removing duplicate observations i e successive 1 10 second intervals in which nothing changed and in corporating a weighting factor to retain the original event duration data In developing the analysis files our approach was always to begin by produci
8. 80 6 yes 19 4 no If yes about how far is your work from your home one way x 17 8 miles median 12 0 8 About what percent of your driving is during rush hour morning or afternoon traffic x 42 0 9 About what percent of your driving is done when it is dark outside 26 0 median 20 0 10 How often do you transport a Infants or pre school age children in your vehicle 10 1 every day 1 5 several times a week 8 7 once or twice a week 79 7 almost never never b Older children or teenagers in your vehicle 4 4 every day 10 1 several times a week 13 0 once or twice a week 72 5 almost never never c Other adults in your vehicle 7 3 every day 30 4 several times a week 47 8 once or twice a week 14 5 almost never never THANK YOU How Traffic and Roadway Conditions Affect Driving Behavior POST DRIVING SURVEY Name Date 1 Did the equipment installed in your vehicle cause any problems for you 14 5 yes 85 5 no If yes please explain See comments attached 2 Was the driving you did this past week typical of your usual driving 78 3 yes 21 7 no If no please explain See comments attached 3 Did having the equipment in your car change your driving in any way 21 7 yes 78 3 no If yes in what way was your driving changed See comments attached 4 Did you have any accidents or close calls while driving this past week 30 4 yes 69 6
9. APPENDIX E VIDEO PRO SIMPLE STATISTICS FOR CODED DATA Observational data file 70 files Statistics across observations gt From aaan Start Of Observation TO senna End Of observation Class phonepgr Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur phonenot 168 734759 4 98 53 4373 57 4273 48 329 71 2 3 11027 7 phoneuse handheld 100 9264 8 1 24 92 65 176 29 17 63 1 2 1264 2 phonring handheld 15 117 9 0 02 7 86 4 24 1 10 1 3 19 7 phondial handheld 122 1567 7 0 21 12 85 13 41 1 21 1 0 65 7 Total 405 745709 8 100 00 1841 26 3480 61 172 95 1 0 11027 7 Class hands Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur handboth 11522 237440 5 31 84 20 61 44 32 0 41 0 0 1820 0 handone 15147 447957 8 60 07 29 57 68 91 0 56 0 0 1921 1 handnone 4576 60311 5 8 09 13 18 24 63 0 36 0 0 364 7 Total 31245 745709 8 100 00 23 87 56 14 0 32 0 0 1921 1 Class eyehead Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur eyesout 7295 700598 0 93 95 96 04 248 51 2 91 0 0 5823 6 eyesin 7244 45111 8 6 05 6 23 14 68 0 17 0 0 276 4 yawn 623 z anger 10 drowsy 22 Total 15194 745709 8 100 00 51 29 181 96 1 51 0 0 5823 6 Class lanes Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur lane2 1506 363242 4 48 71 241 20 299 91 7 73 0 0 2889 9 laneintr 130 77517 8 10 40 596 29 700 61 61 45 13 5 3604 7 lane
10. I forgot that I had it in my car If I did remember it was usually guilt after I ran a yellow light Only the first day aware of equipment self conscious I was probably more aware of how I was driving safety wise and I was a bit more careful Just that I was a bit more attentive to how I drove knowing that it was being moni tored I don t believe my driving changed much though A little was sometimes aware of it I never really forgot that it was there I don t think my driving changed but I found myself holding back from things like singing with the radio or talking to myself Question 4 Did you have any accidents or close calls while driving this past week Comments to yes responses One car and separately one bus pulled out from side street in front of oncoming traffic which I was part of Both somewhat short of close calls but could have been dangerous Close call While driving in Cary on a 4 lane road I was in the right lane when an SUV in the left lane suddenly changed lanes cutting me off and almost hitting my car I hit the brakes I was coming down Duke Street and a car from the right lane crossed over two lanes getting in front of me I had to hit my brakes suddenly to avoid hitting the car in the back People pulling out in front of me and passing very close to me Yes no comments 10 minutes ago on Franklin Street a van changed lanes in front of me that made me brake hard
11. Key study findings that include descriptive results on how often drivers engage in distracting behavior under what conditions they engage in such behavior differences among drivers by age and gender and some consequences of the distracting behavior on driving safety appear in the Results section A final Discus sion section summarizes the study findings and their significance METHODS The AAA Foundation for Traffic Safety s Request for Proposal RFP for the study of the role of driver distraction in traffic crashes called for the development and valida tion of a driving log methodology to estimate the rate of occurrence in the overall driving population of the distractions identified in Phase I of the study In vehicle video cameras were to serve as a basis for the validation Thus the Phase II tasks were to include e Development and validation of a driving log methodology e Refinement of a driver distraction taxonomy e Development of sampling strategies for the base rate study e Collection of base rate data e Coding and analysis of base rate data In our response to the RFP we proposed that a multi camera in vehicle system packaged with an inside vehicle audio source plus accelerometer and other vehicle sen sors be developed to serve as the gold standard of comparison and that another less expensive and easier to install single camera system be developed and validated against this standard In developing the validation system
12. My name will not be connected with the data or used in the research in any way When reporting the findings of the study researchers will be combining data from all of the study participants and reporting overall results If portions of the data are used in a presentation of the study results my identity will not be revealed All data will become the property of the project sponsors and will not be released except as might be required by law Incidences of potential child abuse will be reported as required by North Carolina state law I understand that if I am in a crash during the time that the equip ment is operating that data will be destroyed If I have any questions about the research I can call the project director Dr Jane Stutts at 962 8717 in Chapel Hill or at 1 800 672 4527 if outside of Chapel Hill If she is not available to take my call I can leave a message and she will call me back I also understand that I am free to withdraw from the study at any time even after already participating in some or all of the scheduled data collection This research has been reviewed and approved by the UNC School of Public Health Institutional Review Board on Research Involving Human Subjects Name Date 730 Airport Road Campus Box 3430 Chapel Hill NC 27599 3430 Phone 919 962 2202 e Fax 919 962 8710 APPENDIX C PRE AND POST SUBJECT SURVEYS Subject ID How Traffic and Roadway Conditions Affect Driving Behavior PRE
13. The average time required to light a cigarette was 4 1 seconds and the average smoking event lasted 3 4 minutes the latter increased by a 17 4 minute interval in which a subject was smoking a cigar There were only 17 recorded instances of finishing smok ing averaging 7 3 seconds each Reading writing and grooming Although separate activities reading or writing and grooming were coded in the same class of distractions because of the limitation on total number of classes that could be coded The actual coding scheme allowed for reading writing and grooming to be coded simultaneously but this did not occur in the data Subjects 45 of the total 70 from Table 6 were observed reading or writing on 303 occasions A review of the descriptive comments revealed about equal instances of each activity sample comments included reading a map reading a piece of paper opening and reading mail writing on an envelope writing in a check book reading the newspaper writing in a notebook etc While the average duration of these events was relatively long at 18 4 seconds it should be emphasized that these results do not differentiate between whether the subject s vehicle was moving or stopped at the time However movement status of the vehicle is taken into account in the analyses that follow Grooming noted on 229 occasions included looking at self in mirror combing hair putting on lipstick using a toothpick putting in eye drops taki
14. however it became apparent that technology had advanced to such a level and prices declined that creating multiple gold standard units for placement in subjects vehicles became a viable option At the same time that the data collection and validation process seemed to become more straightforward the process of data reduction and analysis became more compli cated The question became one of how to efficiently code and analyze multiple event data over time Not only were we interested in how often people were distracted when driving and the sources of the distraction but also the durations of behaviors associated with the distractions the operating contexts in which they occurred and their effects on driving and vehicle control To address such data reduction needs we explored com puter software specially designed for the management and analysis of video data The following sections describe in greater detail the methods adopted and how they were used to study distractions in everyday driving Development of Driving Log Methodology The system that was developed for continuous unobtrusive recording of in vehicle driving behavior in subjects own cars consisted of a camera unit camera cable record ing unit and trigger cable Figures 1 and 2 taken from the Driver Distraction Logging Unit User s Manual included in Appendix A show diagrams of the camera and record ing units Further details of the various system components ar
15. standard deviation and standard error and minimum and maximum durations It also performs nested analyses which are essentially multiway crosstabu lations of the data However it does not provide tests of statistical significance As an initial step in our data analysis we examined descriptive results for each of the coded variables on each of the 70 datafiles This allowed us to check for any incon sistencies in the data coding For example a file with an unusually long maximum duration for eyes looking inward or cell phone dialing might mean that these codes were inadvertently left on after the activity had ended If a questionable result was identified it was checked and verified by reviewing the videotapes 29 30 The individual file summaries also allowed us to identify how many of the 70 subjects had displayed each of the various distracting behaviors for example how many used a cell phone smoked or transported young children All subsequent analyses involved all 70 data files combined or a specially selected subset of these datafiles e g all files in which the driver used a cell phone A major challenge was to convert the Video Pro data to a SAS data file so that more in depth analyses including statistical testing could be performed This process was not straightforward One complicating factor resulted from breaking the coding for each subject into half hour segments This necessitated programming stops
16. the CAMERA CABLE and the TRIG GER CABLE The CAMERA BOX mounts to the host vehicles windshield and the RECORD BOX is secured inside the trunk cargo area The CAMERA CABLE connects between the two units and the TRIGGER CABLE interfaces be tween the host vehicle and the RECORD BOX Both cables are hidden upon completion of in stallation What follows is a description of each component THE CAMERA BOX The CAMERA BOX is a small black plastic box measuring 5 X 2 X 2 5 It houses three monochrome microvideo cameras and a micro phone and is designed to be attached to the wind shield below and behind the rearview mirror Two CCD cameras record the cabin of the host vehicle One is a close up view of the driver s head the other is a wide angle view of the interior of the vehicle Both offer a clear 420 TV lines of resolution The third camera records the view of the road ahead of the host vehicle ACMOS cam era is used here to compensate for frequent dy namic changes in light This camera offers 280 TV lines of resolution THE CAMERA BOX MICROPHONE WINDSHIELD MOUNTS CAMERA CABLE CONNECTOR Near infrared filters cover openings on both faces of the CAMERA BOX and allow the cam eras to see out while preventing anyone from see ing inside The near infrared filter also helps reduce solar glare on the road camera The CAMERA BOX is also equipped with a tiny high gain omni directional microp
17. 0 1 0 6 0 4 Internal distraction Reaching 2 2 1 5 20 20 2 7 5 1 2 3 Manipulating controls 1 8 0 8 0 7 06 04 50 1 4 Other internal distraction 1 2 0 9 1 1 23 22 2 3 1 4 External distraction 3 0 1 6 1 2 2 1 17 64 2 4 Other vehicle occupant Baby 1 2 0 3 0 3 0 1 0 8 0 3 0 4 Child 1 5 0 2 0 2 0 02 06 0 2 0 3 Adult 0 0 0 4 0 3 02 01 04 0 3 Conversing 15 5 144 21 4 17 9 11 8 15 7 15 5 Overall 5 2 48 7 10 4 11 5 104 13 8 100 Table F 5 Percentage of time engaged by a potential distraction within categories of vehicle occupant characteristics Other Occupants Distraction None Adult Any Overall Only Child ren Talking on cell phone 1 4 0 6 0 8 1 5 Dialing answering cell phone 0 3 0 1 0 1 Eating drinking 1 6 0 9 1 6 1 5 Preparing to eat or drink 3 1 3 4 2 6 3 1 Music audio on 80 2 41 3 65 9 71 4 Manipulating audio etc 1 3 0 7 1 0 1 1 Smoking 2 0 0 4 0 0 1 5 Lighting extinguishing 0 1 0 01 0 0 0 04 Reading 0 9 0 3 0 6 0 8 Grooming 0 5 0 1 0 2 0 4 Internal distraction Reaching 2 4 1 9 2 4 2 3 Manipulating controls 1 4 1 1 1 4 1 4 Other internal distraction 1 4 2 1 0 5 1 4 External distraction 2 2 2 4 2 9 2 3 Other vehicle occupant Baby 0 0 0 0 2 9 0 4 Child 0 0 0 0 2 3 0 3 Adult 0 0 1 7 0 01 0 3 Conversing 0 1 62 7 33 9 15 5 Overall 69 7 17 9 12 4 100 103 Table F 6 Percentage of time engaged by a potential distraction within categories of vehicle movement Vehicle Movement Distracti
18. 1518 73 1492 74 861 83 604 1 3241 3 occboth 3 allages 4 4492 3 0 60 1123 07 518 02 259 01 589 1 1806 3 Total 340 745709 8 100 00 2193 26 2658 56 144 18 0 3 11098 6 Class vmovemnt Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur vehstop 6332 113964 9 15 28 18 00 27 07 0 34 0 0 479 2 vehwndr 900 vehencro 444 vehmove 6357 631744 9 84 72 99 38 181 58 2 28 0 2 4649 7 vehbrake 22 vehint 4436 vehturn 3158 Total 21649 745709 8 100 00 58 77 136 15 1 21 0 0 4649 7 Class traffic Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur traflght 907 511719 9 68 62 564 19 958 08 31 81 0 8 10364 8 trafmod 934 208861 7 28 01 223 62 322 49 10 55 3 8 3516 1 trafhvy 102 25128 2 3 37 246 35 257 97 25 54 0 0 1353 2 Total 1943 745709 8 100 00 383 79 714 28 16 20 0 0 10364 8 97 98 Class music Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur musicoff 299 204521 1 27 43 684 02 1438 95 83 22 0 5 11027 7 musicon cd 65 21360 2 2 86 328 62 462 08 57 31 2 7 2723 4 musicon tape 127 24480 8 3 28 192 76 296 78 26 33 1 4 2209 4 musicon radio 1215 458730 4 61 52 377 56 720 32 20 67 0 3 10601 6 musicon unknown 106 28210 2 3 78 266 13 471 15 45 76 0 9 3412 9 musicman 1539 8407 1 1 13 5 46 8 63 0 22 0 0 80 3 Total 3351 745709 8 100 00 222 53 660 52 11 41 0 0 11027 7 Class extdstr Behavioral Elements Freq TotDur TotDur M
19. DRIVING SURVEY Name Date Address City Zip Home phone Work phone Age Sex Drivers License No State Vehicle Make Model Year e g 1992 Ford Taurus 1 How many years have you been driving x 28 1 years 2 Approximately how many miles do you drive in a typical week x 244 6 miles 3 Next week do you expect to drive more less or about the same number of miles not coded 4 Are there other people who may drive your vehicle this next week ___ yes ___ no not coded If yes please provide information on each driver s age and sex a Age Sex ___ c Age Sex b Age Sex___ d Age Sex 5 Do you have any medical conditions or take medications that can affect driving 2 9 yes 97 1 no If yes please explain 6 How often do you drive a On local roads or city streets 92 9 every day 5 7 several times a week 1 4 once or twice a week 0 0 almost never never 85 86 b On rural two lane roadways outside of urban areas 50 0 every day 24 3 several times a week 17 1 once or twice a week 8 6 almost never never c On Interstate highways 17 7 every day 25 0 several times a week 44 1 once or twice a week 13 2 almost never never d On other multi lane roadways between urban areas such as US 15 501 or U S 70 34 3 every day 43 3 several times a week 14 9 once or twice a week 7 5 almost never never 7 Do you drive to and from work on a regular basis
20. Light Conditions Distraction Overall Light Gray Dark Talking on cell phone 0 9 2 5 1 2 Dialing answering cell phone 0 2 0 3 0 2 Eating drinking 1 4 1 9 1 5 Preparing to eat or drink 2 9 3 7 3 1 Music audio on 69 7 78 1 71 4 Manipulating audio etc 1 1 1 2 1 1 Smoking 1 6 1 2 1 5 Lighting extinguishing 0 04 0 02 0 04 Reading 0 8 0 6 0 8 Grooming 0 3 0 5 0 4 Internal distraction Reaching 2 0 3 3 2 3 Manipulating controls 1 4 1 3 1 4 Other internal distraction 1 5 1 3 1 4 External distraction 2 5 1 9 2 3 Other vehicle occupant Baby 0 4 0 1 0 4 Child 0 3 0 04 0 3 Adult 0 3 0 2 0 3 Conversing 16 2 12 8 15 5 Overall 79 6 20 4 100 Table F 2 Percentage of time engaged by a potential distraction within categories of traffic conditions Traffic Conditions Distraction Light Moderate Overall or Heavy Talking on cell phone 121 1 5 1 3 Dialing answering cell phone 0 3 0 2 0 2 Eating drinking 1 4 1 8 1 5 Preparing to eat or drink 3 1 3 1 3 1 Music audio on 68 8 77 2 71 4 Manipulating audio etc 1 3 0 9 1 1 Smoking 1 5 1 5 1 5 Lighting extinguishing 0 04 0 04 0 04 Reading 0 8 0 7 0 8 Grooming 0 4 0 3 0 4 Internal distraction Reaching 2 4 2 0 2 3 Manipulating controls 1 6 0 8 1 4 Other internal distraction 1 3 1 7 1 4 External distraction 2 2 2 6 2 3 Other vehicle occupant Baby 0 4 0 2 0 4 Child 0 3 0 2 0 3 Adult 0 3 0 3 0 3 Conversing 16 7 12 7 15 5 Overall 68 6 31 4 100
21. MONI TOR VIDEO cables back to the trunk Run them along the exterior of the car if possible Open the RECORD BOX if not already open The VCR is the top unit inside Examine the back of the VCR There are two leads cur rently connected to it AUDIO IN and VIDEO IN Directly below these are AUDIO OUT and VIDEO OUT The OUT plugs are what the MONITOR cables connect to Connect the MONITOR AUDIO cable to the AUDIO OUT RCA plug on the back of the VCR You may have to lift the VCR up slightly to reach the plugs prop erly Connect the MONITOR VIDEO cable to the VIDEO OUT BNC plug on the back of the VCR Do this by lining up the plugs so that they slide together then twist the cable onto the VCR plug until it locks into place The monitor is now installed STEP 2 Activate the System Turn the vehicle key to the Accessories posi tion Turn the monitor power on You should see a four way split screen on the monitor If you do not consult the TROUBLESHOOTING AP PENDIX Test the audio This is best done using a pair of headphones Plug the headphones into the monitor and tap on the plastic cover to the CAM ERA BOX Speak clearly If you do not hear audio consult the TROUBLESHOOTING APPENDIX Loosen the wing nuts on the CAMERA BOX and angle the CAMERA BOX so that you can remove the screws holding the infrared filter to the front of the box Angle the CAMERA BOX back to where it was and retighten the wing nuts Place the infra
22. The driver I believe is still unaware Car stalled in front of me caused me to have a close call A little surprised when I was facing a truck that crossed the center line to pass a recy cling truck I was worried that the second car would follow Yes no comments There was one incident where I changed traffic lanes had checked rear view mirror to determine safety and it was OK and then a car that must have been speeding came up suddenly I did have a couple close encounters with deer while driving at night and a sudden hard brake in traffic while going to Raleigh Several folks pulling out in front of me One lane change into my lane probably didn t see me since I was in their blind spot 2 cars almost drove right into my passenger side while trying to change lanes Many people pulled out in front of me A driver took a left turn in front of me on a wicked turn No panic stops but a car making a turn While making a left turn on the inside lane of a four lane highway a vehicle changed lanes in front of me very close I was turning and car almost hit me 91 92 Several times people pulled in front of me or traffic slowed down rapidly where I almost rear ended but did not SEPTA bus merged into my lane without any signal Witnessed woman driving through red light using cell phone Car in front of me just missed being hit Someone crossed an intersection right in front of me Question 7 Do you h
23. The larger category of other internal distraction captured such activities as opening one s purse to get out change cleaning sunglasses putting on gloves taking off a tie or jacket using a garage door opener adjusting the sun visor etc External distractions As noted in the Methods section no attempt was made to identify a priori specific external distractions since the potential list was so long and since in many instances the source or nature of the distraction might not be revealed by the outside camera Typical external distractions identified in the comment field included waving or talking Table 10 Duration of potential distractions while vehicle was moving overall and adjusted for number of drivers engaging in the activity of Adjusted of Drivers Total Time Engaging in While Vehicle Activity Moving of Total Potential Distraction Time While Vehicle Moving Using cell phone includes talking dialing answering Eating drinking spilling Preparing to eat or drink Manipulating audio controls Smoking includes lighting and extinguishing Reading or writing Grooming Other occupants Baby distraction Child distraction Adult distraction Conversing Internal distraction 3 External distraction TOTAL without conversing with conversing Based on total sample of 70 drivers Adjusted to reflect the percentage of drivers engaging in that activity i e of total time while vehicle movi
24. _ POWER INVER INVERTER EaP T WHE WOR N N BATTERY PACK j ue GUAD IMAGE PRODESSOR a SUA EC TOMS TO CAMERA BOT RCCER THE RECORD BOX e Record Box casing is extremely rugged and designed for mobile applications e Internal components from bottom to top are Battery Pack Image Processor Video Recorder A Power Inverter is attached to one end panel to convert power for the Video Recorder BOX and supplies the CAMERA BOX with power THE TRIGGER CABLE The TRIGGER CABLE interfaces with the host vehicle allowing the in car recording sys tem to activate only when the host vehicle is turned on This conserves videotape and power resources The TRIGGER CABLE operates via a wired connection to the vehicle s fuse box 73 74 INSTALLATION Installation of the in car recording system is fairly simple and should take approximately 30 min utes in most passenger cars You will need the following tools in order to perform the installa tion e set of jeweler s screwdrivers e regular straight screwdriver e security screwdriver e rubbing alcohol e paper towels e scissors or sharp knife Jeweler s screwdrivers can be purchased as a set from most hardware stores The security screw driver has been provided The RECORD BOX will reside in the vehicle s trunk cargo area The CAMERA BOX will be attached to the windshield by industrial grade suction cups The CAMERA C
25. alignment is the same as the fuse that was removed This new fuse is connected to a wire which has a quick disconnect plug at its end Connect this plug to the one on the red wire coming from the TRIGGER CABLE The black wire coming from the TRIGGER CABLE isa ground wire It needs to be connected to a ground point on the vehicle This can be almost any part of the frame or even a metal brace under the dashboard Locate an appropriate point preferably a screw and tighten the ground wire onto it Hide the TRIGGER CABLE in the same manner as the CAMERA CABLE by tucking it in moldings and securing it with Gaffer s Tape 75 76 ALIGNMENT STEP 1 Monitor Set up Sit the monitor on the passenger s seat If the vehicle s cigarette lighter is available connect it to the monitor using the included adapter Oth erwise the monitor will require ten 10 C size batteries to function There are two RCA plugs on the back of the monitor The RED one is labeled AUDIO IN and the YELLOW one is labeled VIDEO IN The MONITOR AUDIO and MONITOR VIDEO cables are labeled as such The MONITOR VIDEO cable has an adapter on one end to con vert from RCA to BNC This end will attach to the RECORD BOX so attach the opposite end to the monitor now The ends of the MONITOR AUDIO cable are undifferentiated attach either end to the monitor now Make sure the TV AV button on the back of the monitor is set to AV String the MONITOR AUDIO and
26. angle as neces sary then retighten the screw The road camera can be adjusted VERTI CALLY only Horizontal adjustment will require repositioning the CAMERA BOX To adjust the VERTICAL angle of the road camera slightly loosen the mounting screw on the side of the CAMERA BOX which corresponds to the road camera Adjust the camera manually with your fingers then retighten the screw When all three cameras are properly aligned and focused check all screws for tightness then replace the infrared filter Verify all corrections on the monitor then disconnect and remove the monitor monitor power supply if used and MONITOR AUDIO and MONITOR VIDEO cables STEP 5 Secure Camera Cable The CAMERA CABLE needs to be secured to the vehicle both to protect it from damage and to minimize the obtrusiveness of the in car re cording system Start with the CAMERA BOX and work back to the trunk access point Use the natural secure points of the vehicle as much as possible Tuck the cable under molding when possible Other wise use Gaffer s Tape to secure the CAMERA CABLE to the vehicle BE NEAT Using small identical strips of tape evenly spaced is easier and has a cleaner appearance than long easily tangled strips along parts of the car FINAL CHECKLIST Confirm that the record box has power Confirm that there is a tape in the VCR Confirm the record box is locked e Remove any trash tools excess equipment fro
27. is designed for easy installation in a majority of modern cars see the Appendix for a compatibility list This product operates on its own power supply and is autonomous from the host vehicle except for the TRIGGER CABLE which allows the product to record only while the host vehicle is on in order to conserve resources such as videotape and battery power For technical support and replacement parts please contact the following Michael Mercadante 215 855 5380 mercadante transanalytics com 71 72 THE LAW REGARDING AUDIO SURVEILLANCE United States Codes Title 18 Section 2510 2 states Oral communication means any oral com munication uttered by a person exhibiting an expectation that such communication is not sub ject to interception under circumstances justify ing such expectation By definition of the code section a person cannot have an expectation of privacy nor can he or she expect that communication will not be intercepted if there are public signs posted indi cating that the communication is being moni tored The First Amendment of the Constitution provides that any conversation between individu als is private unless otherwise notified In simple terms this means that any overhear ing or recording of a conversation is illegal unless both parties are aware that it is being done COMPONENTS The in car recording system consists of four 4 main parts The CAMERA BOX the RECORD BOX
28. low cost consumer VCRs but they were rejected because they did not have built in battery backup In the absence of this feature there was no way to control the settings for the tape at startup and at the end of the tape without the added expense and uncertainty of building custom circuits The Sanyo SRC 800 VCR en abled us to circumvent the problem and also allowed positioning of the time date stamp in its unused fourth quadrant The battery packs were designed and built by a local battery vendor specifically to fit the available enclosures which were purchased at the local WalMart In the end the system integrated by the research team was very much an off the shelf product Figures 3 5 contain photographs of the camera and recording units and Fig ure 6 shows how the camera unit appeared when mounted in a subject s ve hicle The camera unit housed three cameras plus a microphone One camera was focused on the driver s face another captured a more wide angle view of the vehicle s interior and a third was directed outside at the roadway immediately in front of the vehicle The mi Figure 3 Photo of exterior of camera unit crophone and cameras were hidden from the driver s view by near infrared filters that covered openings on both sides of the camera box The recording unit was positioned in the vehicle s trunk and connected to the camera box and on off trigger via cables that were routed along the lower door fram
29. necessarily those of the AAA Foundation for Traffic Safety or of any individual who peer reviewed this report The AAA Foundation for Traffic Safety assumes no liability for the use or misuse of any information opinions findings or conclusions contained in this report If trade or manufacturers names or products are mentioned it is only because they are considered essential to the object of this report and their mention should not be construed as an endorsement The AAA Foundation for Traffic Safety does not endorse products or manufacturers 2003 AAA Foundation for Traffic Safety TABLE OF CONTENTS IST OF TABLES APPENDIX A PART A 69 Design Specifications for In Vehicle LIST OF FIGURES 4 insrumenaton ACKNOWLEDGMENTS 5 Unobtrusive Driver Distraction XECUTIVE SUMMARY Recording System pa NDIX A PART B NTRODUCTION 11 APPENDIX B 8 Driver Consent for Participation MEI HOD A N DIX Pre and Post Subject Surveys How Traffic and Roadway Conditions Affect Driving Behavior 2 Description of Bootstrap Percentile Method A NDIX Video Pro Simple Statistics for Coded Data APPENDIXF 99 Contextual Factors and Driving Distractions LIST OF TABLES Table Pg 1 12 2 21 3 23 4 25 5 31 6 35 7 36 8 37 9 38 10 42 11 45 12 48 13 50 Percentage distribution of specific driver dis tractions linked to crashes based on national Crashworthiness Data System
30. or calling and all but four of these drivers used it when their vehicle was moving When restricted to moving vehicles only the percentage of subjects who dialed a cell phone decreased from 36 to 27 but there was only a slight decrease in the percentage of subjects who talked on the phone and no change in the percentage answering a phone Table 6 Number and percent of subjects who engaged in potentially distracting activities at any level during the three hours of recorded driving overall and when vehicle was moving Overall Potential Distraction Moving or Stopped Talking on cell phone Answering cell phone Dialing cell phone Any use of cell phone Eating or drinking 51 50 72 9 71 4 Preparing to eat or drink 43 41 61 4 58 6 Manipulating music audio 66 64 controls radio CD etc 94 3 91 4 Smoking 5 5 7 1 7 1 Reading or writing 45 28 64 3 40 0 Grooming 40 32 57 1 45 7 Other vehicle occupant Baby distracting Child distracting Adult distracting i 80 0 Internal distraction Reaching Manipulating vehicle controls All other internal External distraction 1Percent of total subjects n 70 Table 7 Number and percent of subjects by age who engaged in potentially distracting activities at any level during the three hours of recorded driving while their vehicle was moving Potential Distraction Talking on cell phone Answering cell phone Dialing cell
31. periods of time have a higher risk of crashing we do not know this to be true Neither do we know that increased wandering in the travel lane is associated with higher crash risks in real world driving Most importantly we were unable to capture any measure of cognitive distraction which has been linked in the literature to poorer driving performance and increased likelihood of crashing Such studies have typically been carried out in more controlled settings using driving simulators or instrumented vehicles or drivers on test tracks Often they have involved some sort of peripheral detection task to measure driver workload An example is a recent driving simulator study in which users of hands free cellular telephones were found to have reduced recognition of billboards encountered while driving The authors conclude that talking on a cell phone produces a form of inattention blindness Strayer Drews and Johnston 2003 Other studies have suggested that drivers fixed gaze may be an indicator of cogni tive distraction These studies distinguish two types of eye movements that can indicate a driver is distracted either short glances away from the driving task or the longer fixed gazes that signify a cognitive distraction In our less controlled naturalistic driving study we could not differentiate fixed gazes from the desirable category of eyes directed at the roadway for example when cell phone users were carrying on a conver
32. process for both the hour and the minutes Press the PAUSE SEARCH button when finished to save time and date TO DISPLAY THE TIME DATE ONSCREEN DURING RECORDING Press the MENU button until SET UP 2 is displayed Press 16 until the desired item for which the display function will be set is flashing Press 13 14 to set to Y anything that is sup posed to be recorded onto the tape Press PAUSE SEARCH to save and exit Switch the ON SCREEN button 18 to ON Use 15 and 16 to move the display into the lower right quad rant of the display This is the field of the image processor which has no video signal SETTING AUTO RECORD The VCR can be programmed to automati cally enter record mode as soon as a new tape is loaded This setting is necessary to insure that the VCR automatically goes into record mode every time it is powered up To accomplish this press the MENU button until SET UP 2 is dis played Press 16 until TAPE IN MODE is flashing Press 13 14 to set to REC Press PAUSE SEARCH to save and exit SETTING AUTO STOP The VCR can also be programmed to stop when it reaches the end of the tape rather than rewind This is vital to ensure that it does not record over previously collected data To set this press the MENU button until SET UP 2 is flash FRONT PANEL OF VCR 79 80 ing Press 16 until TAPE END MODE is flash ing Press 13 14 to set to STOP Press PAUSE SEARCH t
33. rear in a Porsche SAAB all models have trunk access SATURN all models have trunk access SUBARU Impreza has option of trunk access all others have trunk access TOYOTA Echo has no trunk access all others have trunk access VOLKSWAGEN all models have trunk access VOLVO all models have trunk access APPENDIX TROUBLESHOOTING PROBLEM The monitor is not showing the camera images SOLUTION Check to make sure the monitor cables are con nected properly Make sure the TV AV switch on the back of the monitor is set to AV Make sure the monitor has power Make sure the CAMERA CABLE is connecting properly Make sure the power supply is plugged into the rest of the RECORD BOX PROBLEM The monitor has no power SOLUTION The monitor batteries are dead or the cigarette lighter port is non functional Run the monitor off fresh batteries PROBLEM The in car recording system stays on all the time It does not shut off with the vehicle ignition SOLUTION The TRIGGER CABLE is plugged into an in correct fuse Connect the TRIGGER CABLE to the Accessories or ACC Relay fuse Consult the vehicles Owners Manual for assistance in find ing the correct fuse PROBLEM I can hear a high pitched whine coming from the RECORD BOX SOLUTION This likely is the Loss of Signal alarm It means the RECORD BOX is not receiving a signal from the CAMERA BOX Make sure the CAMERA BOX is connected and reset the R
34. research requires that my vehicle be equipped with certain equipment including video and audio recorders I under stand that the equipment and its installation will not harm my vehicle in any way Installation of the equipment should take no more than 30 minutes I will then drive my vehicle just as I usually do for a period of approximately one week I will return to the location where the equipment was installed to have it removed Re moval of the equipment should take no more than 20 minutes During this time I will be given a brief questionnaire to complete and will be paid 100 for my participation The questionnaire will ask about any crash or near crash situations I may have encountered while driving and my thoughts about participating in the study If someone other than myself drives this vehicle during the time that the equipment is installed their driving behavior will also be monitored However this portion of the data will not be used unless that person has indicated their consent to participate by also signing and returning a consent form Additional drivers will not be compensated I understand that this study is being conducted by researchers at the University of North Carolina Highway Safety Research Center to learn more about conditions affecting road safety and ways of helping drivers avoid crashes There are no costs to me for participating in the study other than costs associated with my travel to and from the installation site
35. results The Video Pro software provides information on the percentage of exact matches i e coded the same way at exactly the same 0 1 second point in time win dow matches coded the same but within some preset interval of time such as plus or minus two seconds window errors coded the same but not within the predeter mined time and coding errors i e none of the above More importantly the soft ware can be used to generate a printout showing at each point in time what each coder has entered This printout especially when examined alongside the videotape itself can be used to pinpoint specific areas of discrepancy In our own coding the overall percentage of agreement number of matches total number matches plus errors only reached about 65 to 70 after training This is low and is a reflection of the subjective nature of many of the behaviors coded At what point does the light condition shift from light to grey When does a reach stop if the driver s hand does not return to the steering wheel And what exactly constitutes grooming Even those behaviors which were more narrowly defined e g lighting a cigarette or dialing a cell phone could be ambiguous in terms of exactly when they started and stopped With an almost infinite possibility of behaviors and conditions it was not possible to develop a single set of objective coding criteria to cover all the behaviors observed
36. use of hand held cell phones while driving Regardless of its eventual safety impact such legislation can be expected to produce a change in many drivers behavior Distractions are a large part of everyday driving and contribute to a large percent age of crashes With all of the new technologies that future vehicles will afford learning how to better manage these distractions is of critical importance to the safety of our roadways References Dingus T A 2002 Driver distractions and design related antidotes Presentation at the Transportation Research Board Human Factors Workshop on Driver Distraction Assessment and Antidote January 13 2002 Washington D C Dingus T A 1995 Moving from measures of performance to measures of effective ness in safety evaluation of ITS products or demonstrations ITS America Safety Evalu ation Workshop May 1995 Llaneras R E 2000 NHTSA Driver Distraction Internet Forum Summary and Pro ceedings Rockville MD Westat Final Report for NHTSA Contract No DTNH22 99 D 07005 Available http www nrd nhtsa dot gov pdf nrd 13 FinallnternetForumReport pdf Pennsylvania Joint State Government Commission 2001 Driver Distractions and Traffic Safety Staff Report of the General Assembly of the Commonwealth of Pennsylvania Joint State Government Commission Harrisburg PA Shelton L R 2001 Statement before the Subcommittee on Highways and Transit Committee on Transportatio
37. vali dated statistically however due to small expected counts in many of the table cells Table 8 with the results by driver sex shows that females were more likely than males to have engaged in some form of grooming activity while their vehicle was mov ing p 004 and were also more likely to have been distracted by something outside the vehicle p 03 Women were also marginally more likely than men to have been distracted by another adult riding as a passenger in their vehicle p 08 Results with respect to baby distractions were inconclusive due to small expected counts in the indi vidual table cells FREQUENCY AND DURATION OF DISTRACTION OCCURRENCES The Video Pro software is designed to provide elementary statistics for each coded variable Included is the variable s frequency or number of times coded total duration in tenths of seconds for the event percent of total duration mean duration and stan dard deviation standard error and minimum and maximum durations This informa tion is reproduced in Appendix E Table 9 contains a simplified version of the printout showing the frequency or number of recorded occurrences total time and percent of total time and mean minimum and maximum durations for each of the identified distractions These results based on the combined data for all 70 subjects provide basic descriptive information on potential driving distractions The results do not differenti ate between moving
38. 1980 1990 vehicles MINI BLADE FUSES are found in most new cars after 1990 Amp ratings for blade and mini blade are printed on top of the fuse MINI BLADE APPENDIX THE VCR Although both the VCR and Image Processor will arrive programmed instructions are provided here to reprogram to original specifications should the need arise The VCR is wired to record only the quad image from the image processor This is done to prevent changes in image appearance from the accidental pressing of any of the image processor s front panel buttons SETTING THE VCR CLOCK Press the MENU button to display the SET UP 1 menu Press the up down arrow buttons 13 amp 14 on illustration below to set the auto daylight standard time adjustment Press the down track ing button 16 Set the day the daylight savings time adjust ment is made Press 13 Or 14 to set the week then press the right tracking button 15 Press 13 14 to set the day of the week then press 15 Press 13 14 to set the month the adjustment is made then press 15 Press 13 14 to set the time the adjustment takes place then press 16 Use the same procedure described above to set the time when daylight saving time is changed back to standard time Press 13 14 to set the current month then press 15 Press 13 14 to set the current day then press 15 The day of the week will set automati cally Press 13 14 to set the year then press 15 Repeat this
39. 1996 The role of driver inattention in crashes New statistics from the 1995 Crashworthiness Data System In 40th Annual Proceedings of the Association for the Advancement of Automotive Medicine 377 392 Vancouver British Columbia APPENDIX A PART A Design Specifications for In Vehicle Instrumentation The video logging system consisted of the camera unit camera cable recording unit and trigger cable CAMERA UNIT The camera unit contained 3 cameras 2 directed inside and 1 outside the vehicle a microphone and a near infrared NIR source A camera cable with a removable lockable connector was used to supply power to the camera unit and relay signals back to the recording unit The entire camera unit was mounted to the windshield just below the rear view mirror using suction cups sealed with a thin film of KY jelly Driver Camera A high sensitivity high resolution miniature CCD charge coupled device camera configured with a 8 mm lens was used to record the driver s face This provided a narrow about 30 degree horizontal field of view HFOV which had reso lution sufficient to permit a crude assessment of glance location Cabin Camera A high sensitivity CCD camera was configured with a 2 5 mm lens which provided a 150 degree HFOV The left edge was aligned roughly with the head rest of the passenger seat The right edge aligned with the driver s door handle to try to capture the driver s hands on the steering wheel T
40. 3 755 77143 8 10 35 102 18 110 81 4 03 4 3 1021 2 lanedivd 506 85857 8 11 51 169 68 255 56 11 36 3 1 2298 5 laneothr 1157 103213 2 13 84 89 21 89 75 2 64 0 2 735 3 lanenh 520 38734 8 5 19 74 49 68 26 2 99 0 3 442 7 Total 4574 745709 8 100 00 163 03 255 90 3 78 0 0 3604 7 Class wlight Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur light good 159 580816 5 77 89 3652 93 3029 33 240 24 19 6 11027 7 light bad 21 12963 2 1 74 617 30 563 48 122 96 38 6 1818 4 dark good 29 12347 4 1 66 425 77 384 04 71 31 10 9 1800 4 dark bad 1 926 1 0 12 926 10 0 00 0 00 926 1 926 1 gray good 100 118542 2 15 90 1185 42 1512 52 151 25 33 3 8768 5 gray bad 26 20114 4 2 70 773 63 969 58 190 15 48 0 4686 6 Total 336 745709 8 100 00 2219 37 2647 04 144 41 10 9 11027 7 Class readgrm Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur rdgrmnot 597 737419 2 98 89 1235 21 2355 33 96 40 0 0 11027 7 reading 303 5583 9 0 75 18 43 29 70 1 71 0 0 282 4 grooming 229 2706 7 0 36 11 82 29 77 1 97 1 0 340 0 Total 1129 745709 8 100 00 660 50 1817 30 54 09 0 0 11027 7 Class smoking Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur smoknot 112 734368 8 98 48 6556 86 4978 93 470 46 12 0 11098 6 smoking 45 11061 3 1 48 245 81 162 95 24 29 13 2 838 1 smoklite 38 154 6 0 02 4 07 2 39 0 39 0 9 10 2 smokfnsh 18 125 1 0 02 6 95 6 90 1 63 0 8 23 7 Total 213 745709 8 100 00 3500 98 4836 96 331 42 0 8 11098
41. 6 Class food Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur foodnot 962 711385 9 95 40 739 49 1891 09 60 97 0 0 11027 7 food 904 5750 2 0 77 6 36 18 19 0 61 0 0 350 0 foodprep 1503 23146 3 3 10 15 40 34 70 0 90 0 0 755 5 drinking 1028 5378 5 0 72 5 23 7 40 0 23 0 3 104 9 foodspil 12 48 9 0 01 4 07 5 22 1 51 0 2 17 6 Total 4409 745709 8 100 00 169 13 933 26 14 06 0 0 11027 7 Class occupant Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur occnone 162 519922 2 69 72 3209 40 3310 17 260 07 0 3 11098 6 occfront 1 chidonly 9 6576 9 0 88 730 77 601 06 200 35 53 3 1802 0 occfront 1 adltonly 92 127211 3 17 06 1382 73 1455 48 151 74 5 9 8083 7 occfront 2 babychld 1 206 0 0 03 206 00 0 00 0 00 206 0 206 0 occrear 1 babyonly 13 17186 6 2 30 1322 05 1249 74 346 62 18 9 4567 8 occrear 1 chidonly 19 19248 5 2 58 1013 08 1025 36 235 23 11 8 3598 9 occrear 2 chidonly 5 3964 4 0 53 792 88 1123 56 502 47 8 8 2710 2 occrear 2 babychld 3 4208 3 0 56 1402 77 1925 04 1111 42 16 4 3600 7 occrear 3 babychld 5 13400 5 1 80 2680 10 1370 16 612 75 737 9 3935 6 occboth 1 babyonly 1 27 6 0 00 27 60 0 00 0 00 27 6 27 6 occboth 2 chldonly 4 3007 1 0 40 751 77 35466 177 33 433 7 1244 1 occboth 2 adltonly 5 6531 3 0 88 1306 26 1478 19 661 07 138 9 3604 2 occboth 2 babyadlt 6 9654 7 1 29 1609 12 1973 75 805 78 2086 5456 3 occboth 2 chldadlt 8 5515 9 0 74 689 49 482 92 170 74 160 3 1611 1 occboth 3 chldadlt 3 4556 2 0 61
42. ABLE will con nect the two units and will be hidden as much as possible as it is routed through the interior of the vehicle The TRIGGER CABLE will connect between the RECORD BOX and the vehicle s fuse box Installation of the in car recording system re quires temporarily replacing an ordinary fuse with a specially rigged fuse of the same size and type This will allow the in car recording system to turn itself off with the vehicle and also preserves all electrical functionality of the host vehicle It is important to note that this is a noninvasive procedure and that the original fuse should be replaced promptly when the in car recording sys tem is uninstalled An appropriate host vehicle requires two de sign elements an interior fuse box and interior access to the trunk Verify the presence of both of these design elements prior to installation An interior fuse box is simply one where the fuse box is installed into the interior of the vehicle rather than under the hood Under the hood fuse boxes are useless unless a way can be found to safely and securely route the TRIGGER CABLE into the interior of the vehicle and if the fuse box cover can be completely sealed after the installa tion of the TRIGGER CABLE fuse For more information on Trunk Access see the Appendix Here are step by step detailed instructions for the installation of the in car recording system STEP 1 Clean Windshield Box Mounts The CAMERA BOX a
43. ANK YOU VERY MUCH FOR PARTICIPATING IN THE STUDY Please use space below for any other comments or suggestions you may have RESPONDENT COMMENTS Question 1 Did the equipment installed in your vehicle cause any problems for you Comments for yes responses Suction cups came loose One day there was a beeping until I came over and techni cians checked equipment After that no more beeping Had to continue to explain Other than that nothing Don t know for sure but wireless key entry panic alarm did not function Tape came loose over windshield a few times Added another slight blind spot to a car that already has too many but I got used to it Slight obstruction of view I was refused entrance on a military base as a military dependent since the equipment was locked and could not be checked by security without breaking the lock Seemed to be a drain on the battery Had a little trouble starting car Wire was loose on windshield Question 2 Was the driving you did this past week typical of your usual driving Comments for no responses Almost no weekend pleasure driving due to very cold rainy weather No no comments Traveled twice to and from Fayetteville Went to a concert Christmas shopping I usually drive more I frequently drive on I 85 and I 95 but had just completed a long distance trip on those roads prior to participating in this study There was one incident where I changed
44. CDS and Pennsylvania crash data Taxonomy of driver distractions Data coding scheme for use with Observer Video Pro software Operational definitions used in coding the video data Licensed U S drivers age 18 and above by age group and sex Number and percent of subjects who en gaged in potentially distracting activities at any level during the three hours of recorded driving overall and when vehicle was mov ing Number and percent of subjects by age who engaged in potentially distracting ac tivities at any level during their three hours of recorded driving while their vehicle was moving Number and percent of subjects by sex who engaged in potentially distracting activities at any level during their three hours of re corded driving while their vehicle was mov ing Frequency and duration of distracting events Duration of potential distractions while ve hicle was moving overall and adjusted for number of drivers engaging in the activity Distribution of total recorded driving time by context variables Percentage of time vehicle stopped within levels of distraction variables Distribution of outcome measures for exam ining the effects of distractions on driving performance 14 51 15 52 16 53 Results of bootstrap analyses for hands on steering wheel as a function of each distrac tion event when vehicle was moving Results of bootstrap analyses for eye direc tion as a function of each dist
45. ECORD BOX by turning it off for a moment If the whine con tinues the CAMERA CABLE may be damaged Replace and try again PROBLEM The suction cups will not stick to the windshield or they will not stay in place SOLUTION Clean both the suction cups and the windshield thoroughly again We recommend licking the cups because saliva is the best lubricant weve tested The key is to apply something that will not evaporate creating a strong suction Try Vaseline or olive oil but apply sparingly other wise cups may slide PROBLEM The buttons on the VCR do not respond SOLUTION Check to make sure a videotape is loaded If dddd is flashing on the display all functions are suspended due to condensation inside the machine Wait for the dddd indicator to go off Check to make sure the security lock is disen gaged OTHER PROBLEMS QUESTIONS COMMENTS CONTACT TECHNICAL SUPPORT Michael Mercadante 215 855 5380 mercadante transanalytics com TransAnalytics LLC 1722 Sumneytown Pike P O Box 328 Kulpsville PA 19443 www transanalytics com 83 84 APPENDIX B THE URISERSTY OF ORTH CAROL HIGHWAY SAFETY _ RESEARCH CENTER SAFETY RAOELITY INFORMATION How Traffic and Roadway Conditions Affect Driving Behavior CONSENT TO PARTICIPATE I understand that I have been asked to participate in a research study to learn more about how traffic and roadway conditions affect driving behavior The
46. ES INDEPENDENTLY USER S MANUAL UPDATED APRIL 2001 2001 TransAnalytics LLC 1722 Sumneytown Pike P O Box 328 Kulpsville PA 19443 215 855 5380 phone 215 855 5381 fax www transanalytics com A GUIDEBOOK FOR INSTALLATION AND OPERATION INTRODUCTION TABLE OF CONTENTS INTRODUCTION ncritesavvetivmimipiccsnemmnanaunen 7 THE LAW REGARDING AUDIO SURVEILLANCE 12 COMPONENTS s soos1sos00ss0sss0sssoessotesstssseoesnoessseesseesse 12 INSTALLATION isce dicrircctencaieinntetnatcnnsens 74 ALIGNMENT ssa jacinta east arasctetcainatan aati 76 FINAL CHECKLIST nininini 11 CHARGING THE BATTERY PACK sseeccccosssseseeeeeeees 11 THE BATTERY CHARGER sssssssssssccescssssseeeceeesseees 78 APPENDIX AUTOMOTIVE FUSES ssosssssesccceeessees 78 APPENDIX THE VR zirci nians 78 APPENDIX THE IMAGE PROCESSOR eeessesosnseees 80 APPENDIX TRUNK ACCESS ssseecscccccssssssecceeesssees 8 APPENDIX TROUBLESHOOTING o oo ssssescccnenssees 83 This product has been designed to meet the needs of a driver distraction research study being con ducted by the Highway Safety Research Center at the University of North Carolina in coordination with the AAA Foundation for Traffic Safety This product is designed to continuously and unobtrusively record video and audio within the cabin of a vehicle and record video of the road ahead of the vehicle The in car recording system
47. II to provide some indication of the relative crash risk associated with various driving distractions The Phase I analysis was based on five years 1995 1999 of the National Automotive Sampling System NASS Crashworthiness Data System CDS data Over all 8 3 percent of drivers on the file were identified as distracted at the time of their crash When adjusted for drivers with unknown attention status this percentage rose to 12 9 percent The primary sources of driver distraction identified on the CDS data were shown in Table 1 The three most often cited distractions identified as contribut ing to crashes were outside objects persons or events adjusting the radio or other audio equipment and other occupants in the vehicle Eating and drinking and cell phone use were both less frequently cited although it should be noted that cell phones were not as popular during the time period covered by the CDS data as they are today For the Phase II exposure study external distractions played a less prominent role than did the various internal distractions including manipulating vehicle controls and reaching for things inside the vehicle It is not immediately known where such reach ing distractions would be coded in the CDS data but perhaps only in the other distraction category since they encompassed a wide range of behaviors Eating and drinking was another very frequently observed category especially when considered along with preparin
48. MW all models have trunk access BUICK all models have trunk access CADILLAC only SeVille De Ville and Catera have trunk access CHEVROLET offered as an option on Malibu and Impala all others have trunk access CHRYSLER all models have trunk access DODGE standard on Avenger all others available as an option FORD not available on Crown Victoria and Thunderbird available as an option on all others HONDA all models have trunk access HYUNDAI all models have trunk access INFINITI Q 45 has no trunk access all others have trunk access JAGUAR only the 2000 S Type offers any trunk access No other models or years offer trunk access KIA all models have trunk access LEXUS only ES300 has trunk access no other models offer trunk access LINCOLN Towncar and Continental have no trunk access LS has trunk access MAZDA Prot g and 626 have trunk access no other models offer trunk access MERCEDES C Class have trunk access no other models offer trunk access MITSUBISHI Galant Mirage and Eclipse have trunk access MERCURY Grand Marquis has no trunk access all others have trunk access NISSAN all models have trunk access OLDSMOBILE all models have trunk access PLYMOUTH all models have trunk access PONTIAC Grand Am offers trunk access as an option Bonneville Grand Prix and Sunfire have trunk access no other models offer trunk access PORSCHE No models offer trunk access the engine is in the
49. Our approach to dealing with this situation primarily involved having the coders continuously review each other s coding and also having them work ing together in the same office so that if questions arose they could get a second opin ion Altogether a total of 207 2 hours of video data were coded This is just shy of the targeted 210 hours 3 hours for each of the 70 subjects For two subjects we could not code the full quota of data because of poor tape quality e g too dark and or problems with the recording equipment Since most tapes contained much more than the requi site three hours we divided the tape into half hour segments for coding For example if there were six hours of usable data on the tape we would skip the first half hour code the next half hour skip the half hour after this etc until the end of the tape If there were less than six hours of total video we would skip less time between the half hour segments and if there were more than six hours of total video we would skip more time between segments Except for the two cases with too little data the first half hour was always skipped to allow the driver some period of adjustment to the presence of the recording equipment in the vehicle Video Data File Development and Analysis The Observer Video Pro software is designed to provide basic descriptive data on observational data files This includes summaries of event frequency total duration mean duration
50. RC 800 VCR All VCRs were programmed to automatically record in SLP mode when powered on The time date stamp was positioned in the lower right quad rant where it was clearly legible by the data coders A 50 Watt power inverter was used to provide AC power to the VCR Power Supply Power was provided by a custom battery pack Four 12 volt DC each providing 7 2 AH sealed gel cell batteries were connected in parallel to provide a total of 28 8 AH The total current draw in record mode was 2 64 amps As such total record time for one full charge was approximately 10 hours TRIGGER CABLE A small amount of power from the vehicle was used to activate a relay The relay was used to complete the power circuit and was triggered using a fuse adapter that was plugged into the vehicle s fuse box The fuse adapter replaced the existing fuse with matching amp rating In most instances the accessory fuse was used because it sup plied continuous power only while the vehicle was running In some cases the acces sory fuse was constantly on In these instances a voltmeter was used to find an appropriate fuse location The trigger cable ground was connected to a bolt underneath the instrument panel using an alligator clip The normally open relay contacts were connected to the battery pack which sup plied power to the recording system APPENDIX A PART B UNOBTRUSIVE DRIVER DISTRACTION RECORDING SYSTEM DESIGNED FOR MODERN CARS OPERAT
51. a coding and analysis system and also converted to a SAS data file for more detailed analyses including non parametric bootstrap procedures for constructing confidence intervals around propor tion estimates Following are highlights of the results e Distractions are a common component of everyday driving During their three hours of coded driving time almost all of the volunteer subjects were observed manipulating vehicle controls such as air conditioning or window controls and reaching for objects inside their moving vehicle Nearly as many were observed manipulating audio controls or had their attention drawn to something outside the vehicle Approximately three fourths ate or drank something while driving or conversed with a passenger Reading writing and grooming activities were also relatively common but declined to less than half the participants when observa tions were restricted to moving vehicles only In addition about a third of the subjects used a cellular telephone while driving and nearly as many were dis tracted by passengers in their vehicle e Age differences in the likelihood of engaging in a particular distraction were gener ally small Compared to older drivers younger drivers appeared more likely to eat or drink while driving to manipulate audio controls and to be distracted by other 61 62 occupants in the vehicle Sample sizes within age groups however were small prohibiting valid statistical testing Wi
52. a unit was mounted inside the vehicle just below the vehicle s rear view mirror Two of the cameras were directed inside at the driver and front seat area of the vehicle and the third was directed outside the vehicle straight ahead The recording unit was generally placed in the trunk of the vehicle and cables discretely run between the units The recording equipment was installed in the vehicles of 70 volunteer subjects equally distributed among males and females in five age groups 18 29 30 39 40 49 50 59 and 60 Half the subjects were recruited from a base in Chapel Hill North Carolina and half from Kulpsville Pennsylvania just outside of Philadelphia Installa tion of the equipment generally took 30 minutes or less Subjects were informed only that the study was being conducted to learn how traffic and roadway conditions affect driving behavior They were instructed to drive normally and scheduled to return one week later for removal of the equipment The resulting videotape data was coded using software The Observer Video Pro specially designed for the coding and analysis of videotaped data A coding scheme was developed that included a full range of driver distractions from the Phase I analysis along with selected contextual and outcome variables Contextual variables included the presence of passengers in the vehicle light conditions weather conditions roadway type traffic level and whether the vehicle was stopped
53. affic level Light Moderate Heavy Passenger presence in vehicle age No passengers Baby in vehicle Child in vehicle Adult in vehicle Passengers in vehicle seating position No passengers Front seat passenger s only Rear seat passenger s only Both front and rear seat passengers 1 Driving under dark lighting conditions is underreported since it could not be reliably coded 2 Percentages total more than 100 because of overlap among the occupant categories e g both another adult and a baby in the vehicle 45 46 e With regard to eating and drinking subjects were more likely to be eating and drinking when lighting conditions were grey or dark e g when driving to work during the winter months and when weather conditions were bad Otherwise the different contextual variables had little impact on the eating and drinking be havior of our subjects Music or audio was more likely to be on when light conditions were gray or dark traffic conditions were moderate or heavy and when no passengers were present in the vehicle It was especially less likely to be on when another adult was present in the vehicle Interestingly drivers were not more likely to manipulate the music controls when their vehicle was stopped compared to when it was moving The five subjects in our study who smoked were more likely to do so during day light driving conditions in light traffic and when weather conditions were good They were also
54. and two variables describing any passengers in the vehicle whether a baby child or adult and whether seated in the front or rear of the vehicle The distribution of recorded driving time by these contextual variables is shown in Table 11 Iwo other context variables whether passing through or turning at an intersection are not included in the table since they were considered to be events without associated durations Beyond providing some basic descriptive information about the recorded driving data e g passengers were present in the vehicle less than a third of the time weather conditions were deemed good 95 percent of the time traffic levels were typically light and very seldom were they heavy etc our primary interest in coding the contextual variables was to learn whether distractions were more or less likely to occur under cer tain conditions For example are people more or less likely to talk on cell phones when driving in heavy traffic Does having a passenger in the vehicle make certain distrac tions less likely Does it make some more likely How do people behave when their vehicle is stopped in traffic compared to when it is moving Do weather conditions affect drivers willingness to engage in certain distracting behaviors etc To address these questions Appendix F contains tables showing the percentage of time drivers engaged in each type of distraction within levels of each contextual vari able Because of the l
55. and stopped vehicles since under real world driving conditions distractions do not start and stop whenever the vehicle starts and stops For example a cell phone conversation does not end simply because a vehicle stops at a stop light To present accurate descriptive data about the occurrence and duration of cell phone con versations one needs to include both the time the vehicle is moving and when it is stopped Key descriptive findings from Table 9 are highlighted in the individual descrip tions that follow Cell phone pager The 28 subjects in our sample of 70 who used a cell phone placed 122 calls received 15 calls and carried on 100 phone conversations Assuming three hours of coded data per subject this translates into 1 5 placed calls 0 2 incoming calls and 1 2 conversa tions per hour for those 28 subjects who used a cell phone at all in their vehicles The 39 40 average time required to place a call was 12 9 seconds and to answer a call 7 9 seconds the average conversation lasted 1 5 minutes 92 7 seconds but ranged from only a second or two to over 20 minutes 1264 2 seconds in length Eating and drinking Eating or drinking was coded whenever food or drink was brought to the mouth it was not coded during chewing or swallowing or if the food or drink was simply being held in the hand or lap In the latter situation preparing to eat or drink would be coded Preparing to eat or drink was also coded
56. ave suggestions for improving data collection with future participants Comments for yes responses Was not sure equipment was recording It was very unobtrusive and quiet Is there some way to check if equipment is working properly Also suction cups on windshield would come off at night I would stick them back on Make large box smaller so it wouldn t be as inconvenient Extend the length of time the measurements are taken so that one becomes less conscious of the devices I would do the study to not just see the conditions but really study people and their driving as well If voice is recorded and I assume it is it would be nice to disable it Remind participants that it is only audio and then at some point video would also be helpful to see what the driver sees and you can adjust data to it audio to video Try to collect behavior data such as Interstate speed slows on inclines causing a backup in heavy traffic DOT could allow for this by adding additional lane where necessary Also parking lights why do people use them when headlights are needed Need to tell us you need space in the trunk so we can clear up first If you leave the equipment in more than the 7 day contract you should compensate for the extra days I knew there was camera but somehow missed the point that it was road focused only Smaller digital equipment might be more costly but also more efficient Smaller video recorder Mount a came
57. cluding one by NHTSA Executive Director L Robert Shelton Shelton 2001 In March 2002 the National Conference of State Legislatures NCSL released a report documenting an eight month study that brought together state legislators and staff wireless service providers auto manufacturers other interested companies and industry safety groups federal agencies researchers and other stakeholders to address the issue of technology in motor vehicles Sundeen 2002 The report identified 14 principles to assist state legislatures in addressing legislation about technology in motor vehicles Among other things the report recommended that states rather than local jurisdictions should decide whether to regulate the use of wireless telephones and other technologies that no regulation should prevent a driver s use of a wireless telephone in 11 12 emergency situations that all drivers should receive driver distraction educational ma terials and that all states should collect data about the involvement of driver distrac tions on crash report forms The forum failed to agree however on whether legislation should be passed restricting the use of specific technologies including wireless tele phones in motor vehicles and on whether hands free phones might be allowed but hand held phones prohibited It also failed to agree on whether wireless phone use as opposed to all potential driving distractions should be singled out for reporting on
58. coding station etc to aid in coding Even so with such a complicated data coding profile the actual time for coding an hour of video data generally ranged from six to eight hours Because of this and because we had varying amounts of available data for each of our subjects a decision was made to code only three hours of total data per subject Three employees of the HSRC completed all of the coding All had extensive prior experience in coding video data though not in using the Video Pro software Two of the employees were directly involved in developing the coding scheme and testing the various coding profiles As such they routinely practice coded iterations of the final coding scheme over a period of several months By the time formal data coding was initiated they were both well experienced with the system A third data coder was brought in to assist about midway through the coding and was trained by the other two coders This primarily involved studying the coding scheme watching the other two coders as they coded the data and then practice coding on some of the same tapes and comparing results Formal inter observer reliability checks were made at several stages in the coding process Early on these were primarily used as a training tool to ensure that the coders were all interpreting the video data consistently This typically involved having the coders all code the same one half to one hour segment of data and comparing the
59. controls Smoking Not smoking Ref Lighting or extinguishing Smoking Reading writing Ref Not reading writing Reading writing Grooming Not grooming Ref Grooming Occupant distraction No occupant distraction Ref Distracted by baby Distracted by child Distracted by adult Conversing Not conversing Ref Conversing Internal distraction No internal distraction Ref Manipulating vehicle controls Reach lean look for etc Other internal distraction External distraction No external distraction Ref External distraction Each variable level compared to reference Ref level e g talking listening on cell phone compared to phone not in use dialing answering cell phone compared to phone not in use etc 2N S non significant 53 54 Figure 9 Cell phone effects on driving performance 70 z 60 50 40 30 20 10 o L No Phone Reference Dialing Answering Talking Listening E No Hands E Eyes in Adverse Events Hour p lt 05 p lt 01 Cell phone pager Dialing or answering a cell phone or in two instances pager was associated with significantly higher levels of no hands on the steering wheel and eyes directed inward The rate of adverse vehicle events was also higher although not signifi cant statistically Talking on a cell phone was also associated with higher levels of no hands on the steering wh
60. created a partnership with wireless service providers auto manufacturers other interested com panies and industry safety groups federal agencies academics and other stakeholders specifically to discuss the use of technology in motor vehicles and to provide guidance to state legislatures on how they should respond to this emerging safety issue Sundeen 2002 A large and growing body of literature addresses the benefits and potential dan gers of new technologies that auto manufacturers are installing in vehicles and others that people simply carry into their vehicles While these newer technologies are compelling areas for study the results of the current project both the Phase I analysis of NASS Crashworthiness Data System CDS data and the Phase II field data study have demonstrated that the majority of driver distractions are neither new nor technological Rather they are aspects of every day driving that people are likely to seldom think about sipping a cup of coffee 65 66 reaching into the glove compartment changing a tape or CD tending to a small child or rubbernecking to get a better view of something outside the vehicle It has long been recognized that drivers themselves cause the vast majority of crashes Indiana University s landmark in depth crash causation study conducted in the mid 1970s revealed that human factors including improper lookout excessive speed inat tention improper evasive ac
61. creen Use the lt and gt keys to set the SECURITY LOCK to ON This dis ables all front panel buttons except MENU APPENDIX TRUNK ACCESS The RECORD BOX in the in car recording sys tem is designed to be placed in the trunk cargo area of the vehicle This is for practical and security purposes but also to minimize the obtrusiveness of the in car recording system by removing all but the neces sary equipment from the driver s field of vision This in car recording system is NOT WEATHERPROOF and is not intended for use in pickup trucks or other vehicles that lack suffi cient trunk space The in car recording system may be used in sport utility vehicles and vans by placing the RECORD BOX in the rear cargo area Because the CAMERA CABLE needs to con nect from the interior of the vehicle to the trunk a host vehicle must have some sort of access be tween the two areas The most common ways to do this are a fold down rear seat a split rear seat half the seat folds down and a center console pass through Convertibles have sealed wells be tween the rear seat and trunk to accommodate the top therefore convertibles offer no trunk ac cess from the interior of the vehicle A list of major car manufacturers follows with a note as to the availability of trunk access from the interior on newer model years 1995 2000 81 82 ACURA only the TL and 4 door models have trunk access AUDI all models have trunk access B
62. ctor on the lower right side of the RECORD BOX It will be identical to the one on the CAMERA BOX Connect the CAMERA CABLE to the RECORD BOX following the instructions in Step 3 STEP 6 Install the Trigger Cable The TRIGGER CABLE allows the in car record ing system to interface with the host vehicle so that the system is only functioning when the host vehicle is on This allows for more efficient data collection and better conservation of resources Feed the TRIGGER CABLE through the trunk and into the cabin Locate the vehicle s fuse box This is typically in the driver s side dashboard below the steering wheel It probably has a cover over it Consult the Owner s Manual if available Gently remove the cover Locate the fuse diagram usually on the inside of the fuse box cover Locate the ACCESSO RIES fuse It may also be labeled ACC RELAY or something similar It is the fuse that controls power to the stereo windshield wipers and other accessories It is off this fuse that the TRIGGER CABLE functions Make sure the key is not in the ignition Re move the Accessories fuse and examine it to de termine the type of fuse and amp rating see APPENDIX for more information on fuses Store it in a safe place It will need to be replaced when the in car recording system is uninstalled Locate an identical fuse among the ones sent with the in car recording system Install it into the vehicle making sure the
63. d Figure 12 Smoking effects on driving performance 35 30 25 20 15 10 0 E l m Not Smoking Lighting Extinguishing Smoking E No Hands Eyes In Adverse Events Hour p lt 05 Smoking Even though lighting or extinguishing cigarettes was associated with higher levels of all three negative outcomes only the results for eyes looking in were significant statistically p lt 05 due at least in part to the small sample size only five smokers and less than five minutes of total time spent lighting and extinguishing In contrast when smoking drivers were significantly Jess likely to experience an adverse vehicle event p lt 05 They also were no more likely to have no hands on the steering wheel or their eyes directed inward 55 56 Figure 13 Reading writing effects on driving performance 100 80 60 40 20 a Not Reading or Writing Reference Reading or Writing m No Hands m Eyes In Adverse Events Hour p lt 01 Reading writing Attempting to read or write something while their vehicle was in motion was associated with significantly higher levels of no hands on the steering wheel and especially and not surprisingly eyes looking inward The fact that reading writing was not also associated with statistically higher level
64. d etc No associated duration Code for any apparent distraction by an insect including swatting at an insect brushing an insect away opening window to let insect out etc No associated duration Start when distraction first occurs and stop when it ends Distractions might include sudden movements or noises barking climbing into the driver s lap etc Start coding when arm reaches out and or body position shifts and stop when it returns to a normal position Includes accessing glove compartment and reaching into pockets of clothing Code for any other events occurring inside the vehicle that appear to distract the driver Identify event in comments None of the below default setting Code for any apparent distraction outside the vehicle Examples might include looking at passing scenery sudden movements by other vehicles traffic congestion vehicle occupants or pedestrians searching for street names etc Coding should be based on an observed change in eye direction and or intensity of gaze Code even if appears to be distracted i e looking away says something but source of distraction cannot be determined No other occupants in vehicle default setting One or more other occupants in the front seat only One or more other occupants in the rear seat s only One or more other occupants in both the front and rear seats Normal daylight driving conditions default setting Nighttime driving conditions lighted or u
65. de the vehicle along with higher rates of adverse vehicle events associated with each of the identified driving distractions Al though in the anticipated direction however the results frequently do not attain statis tical significance This likely reflects a combination of factors including small sample sizes e g smokers other occupants in the vehicle rare events as outcomes and any lack of precision in our coding There are also some notable exceptions to the trend of higher levels of potentially dangerous driving behaviors for a few of the identified dis tractions Results for the individual distractions are highlighted below aided by a series of histograms summarizing the results in the tables Table 13 Distribution of outcome measures for examining the effects of distractions on driving performance Measure of Occurrence Outcome Measure i Overall Vehicle Moving Hands on Steering Wheel Both hands on wheel One hand on wheel No hands on wheel Focus of Eyes Eyes directed outside vehicle 93 9 Eyes directed inside vehicle 6 1 Adverse Vehicle Events Wanderings within travel lane Encroachments into another lane Sudden brakings 1 For hands and eyes measure is percent of total driving time For adverse vehicle events it is the total number of recorded events by definition all occurred when vehicle was moving Table 14 Results of bootstrap analyses for hands on steering wheel as a function of each distrac
66. e also included in the Appendix 15 16 WINDSHIELD MOUNTS CAMERA CABLE CABIN CAMERAS Figure 1 Diagram of interior of camera unit oe MICROPHONE CONNECTOR DUAD AGE PROCESSOA JMHECTIONS TO CAMERA BOXMITRIGGER Figure 2 Diagram of interior of recording unit The basic system was developed over an eight month time period during which numerous equipment options were explored and tested Further changes e g an enlarged battery pack a trigger switch wired into the vehicle s fuse box the ad dition of an infrared light source for nighttime recording were made over the next several months as the system was piloted in various ve hicles Our challenge was to develop a system within the project s bud get constraints that was not only compact and unobtrusive but also relatively quick to install reliable and able to simultaneously record audio and video both inside and outside the vehicle for up to 10 hours of total driving time The approach taken in develop ing the methodology was iterative involving considerable trial and er ror For example we initially tried a wireless trigger mechanism for turning the recording unit on when the car engine was turned on but found that there was too much interference for it to work reliably Numerous camera brands and lenses were tried with the final choice based on a combination of performance cost and availability We also tried
67. e on the driver s side of the car A schematic of the overall system is contained in Figure 7 Along with developing the equipment itself we also developed guidelines for in stalling and operating the equipment in vehicles These guidelines are summarized in the User s Manual that has been reproduced in Appendix A Equipment installation involved the following steps e Attaching the camera box to the windshield just below the rearview mirror so that it blocked as little of the driver s field of view as possible e Attaching the camera cable to the camera box and running it along the driver side doors to the trunk area of the car e Connecting the cable to the recording unit box placed in the trunk 17 18 e Installing the trigger by locating the fuse box and re placing the accessories or similarly labeled fuse with an identical type fuse with a quick disconnect plug and routing this wire through to the recording unit box as well connecting to its matching trigger mount e Connecting the moni tor to the VCR unit for use in focusing the cameras and aligning the camera angles Also testing the audio on the system e Removing the monitor Making sure all equipment and cables are secure to the vehicle using non residue Gaffer s tape to secure the cam era cable to the vehicle The User s Manual also contains information on trunk access for various vehicle makes and a section on Figure 5 Photo of
68. ean StdDev StdErr MinDur MaxDur extnot 725 728212 1 97 65 1004 43 1771 14 65 78 0 0 10848 6 extdstr 659 17497 7 2 35 26 55 58 78 2 29 0 4 770 5 Total 1384 745709 8 100 00 538 81 1372 05 36 88 0 0 10848 6 Class intdstr Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur intnot 4153 708041 2 94 95 170 49 298 75 4 64 0 0 4351 3 intfall 11 intinsct 1 intpet 14 184 4 0 02 13 17 13 00 3 48 0 5 47 0 intother 481 10364 9 1 39 21 55 46 38 2 11 0 0 496 3 intreach 2246 17014 6 2 28 7 58 36 70 0 77 0 0 1351 0 intmvc 2095 10104 7 1 36 4 82 11 53 0 25 0 0 283 8 Total 9001 745709 8 100 00 82 96 219 79 2 32 0 0 4351 3 Class converse Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur convnot 1614 630359 9 84 53 390 56 1446 35 36 00 0 0 11098 6 converse 1558 115349 9 15 47 74 04 234 50 5 94 0 0 4827 0 Total 3172 745709 8 100 00 235 09 1056 48 18 76 0 0 11098 6 Class occdistr Behavioral Elements Freq TotDur TotDur Mean StdDev StdErr MinDur MaxDur intnotz 305 738722 1 99 06 2422 04 3859 24 220 98 0 9 11098 6 intbaby 114 2677 9 0 36 23 49 28 39 2 66 0 8 192 6 intchild 81 2086 2 0 28 25 76 124 72 13 86 0 7 1124 2 intadult 48 2223 6 0 30 46 32 108 49 15 66 1 1 608 8 Total 548 745709 8 100 00 1360 78 3114 00 133 02 0 7 11098 6 APPENDIX F CONTEXTUAL FACTORS AND DRIVING DISTRACTIONS Table F 1 Percentage of time engaged by a potential distraction within categories of light condition
69. ected outside the vehicle and appear to be focused on the driving task i e not looking at passing scenery other drivers etc Also include any mirror checks in this category Eyes are directed inside the vehicle and are not focused on the driving task Do not include mirror checks and quick checks of the dashboard e g at the speedometer in this category Code whenever the driver yawns No associated duration Code whenever clear signs of driver anger or aggression towards others on the roadway for example yelling at other drivers hand motions teeth gritting etc No associated duration Code whenever signs of drowsiness including eyes drooping or closing head jerking etc No associated duration 27 28 Actual coding was carried out by simultaneously monitoring the three video screens on the quad camera monitor display and entering the two letter codes shown in the table to record all changes in behavior see Figure 8 Generally at least two complete passes of the videotape were required the first pass to record eye direction outside or inside the vehicle and hand position both one or no hands on the steering wheel and the second pass to record all other behaviors For very active drivers a third pass was sometimes necessary The Video Pro system makes it easy to start and stop the tape move forward or back ward to a specific spot on the tape move in slow motion Figure 8 Photo of data
70. ection 13 the image processor only outputs the quad image to the VCR This means that even if the unit is bumped and acci dentally switched to a fullscreen sequencing mode the signal going to the VCR remains un changed The quad view allows constant moni toring of all three cameras in the in car recording system Do not connect the VCR to the VIDEO OUT connection on the image processor This connection does not protect the quad display PROGRAMMING THE IMAGE PROCESSOR Some brief programming needs to be done to ensure proper functioning of the image proces sor This will be done by TiansAnalytics before shipping but is provided herein in the event that such programming is somehow deleted from the units memory Begin by pressing the MENU button 9 twice Press the FWD REV buttons to move to the AUTO RESET entry Set the auto reset time to 001 by using the lt and gt keys This enables the image processor to instantly reset itself if an alarm occurs Press MENU again This is the SENSOR TYPE menu Use the lt and gt buttons to set all four channels to OFF in the last position Press MENU again This is the VIDEO LOSS ALARM menu Use the above method to set all four channels to OFF Press MENU again This is the DISPLAY menu Use the above method to turn all display functions to OFF Change cross line to BLACK Set BUZZER to OFF Press the MENU button twice to reach the SECURITY LOCK s
71. ed for their participation Data collection activities were initiated in late November 2000 and extended into the following November thus spanning a one year time period Refinement of Driver Distraction Taxonomy An initial taxonomy of driver distractions was developed at the conclusion of the Phase I project activity based on the CDS and NC narrative data analyses This tax onomy was further refined in Phase II to work with the Observer Video Pro software that would be used to code and analyze the data Table 2 shows our final taxonomy of driver distractions The taxonomy incorporates the major categories of driver distraction appearing on the CDS datafile with some further refinements For example the CDS category of other occupant was expanded to identify whether the distraction was caused by an infant a child or another adult and smoking was expanded to include categories of lighting or extinguishing a cigarette pipe etc In addition a category of internal dis traction was added to capture distractions from manipulating vehicle controls as well as reaching or looking for objects inside the vehicle falling objects pets and insects or bugs inside the vehicle Conversing with a passenger was also added as a potential other distraction as were reading writing and grooming activities The large CDS category of outside person object or event appears in the tax onomy as external distraction No attemp
72. ed some further indications of the representativeness of the study population and whether hav ing the equipment installed in their vehicles might have altered driving behavior As noted earlier the survey results are summarized in Appendix C In the pre driving survey 81 of the participants reported driving to and from work on a regular basis with no significant differences between the North Carolina and Philadelphia area par ticipants The average one way commute distance was 17 8 miles while the median distance was lower at 12 0 miles The lower figure is closer to the national average of 11 6 miles based on 1995 Nationwide Personal Transportation Survey NPTS data but this number has most assuredly increased in the intervening years U S DOT undated Participants were most likely to drive on local roads or streets 92 9 daily followed by rural two lane roadways 50 0 daily multi lane roadways between ur ban areas 34 3 daily and Interstate highways 17 7 daily The final two catego ries of roadway were higher for the North Carolina than the Philadelphia participants Overall participants reported driving 244 6 miles a week on average This translates to 12 719 miles per year which is close to the national average for all drivers based on the 1995 NPTS data An additional set of questions on the pre driving survey asked how often drivers engaged in various activities while driving such as wearing seat belts stopping for ye
73. eel p lt 05 but was not associated with higher rates of eyes looking inward or adverse vehicle events Figure 10 Eating drinking effects on driving performance 20 15 10 x Kk 5 kk 04 No Eating Drinking Preparing to Eat Drink Eating Drinking Reference No Hands m Eyes In Adverse Events Hour p lt 05 p lt 01 Eating drinking Preparing to eat or drink i e holding a cup unwrapping fast food etc was associated with higher levels of all three negative outcomes no hands on the steering wheel eyes directed inward and adverse vehicle events Eating and drinking was associated with higher levels of no hands on the steering wheel and eyes directed inward but not with adverse vehicle events Figure 11 Music audio effects on driving performance 25 20 15 10 5 o 4 No Music Reference Music On Manipulating Music Controls No Hands m Eyes In Adverse Events Hour p lt 05 p lt 01 Musiclaudio Simply having a radio CD or tape playing while driving was not associ ated with any of our three negative driving outcomes However manipulating the radio controls inserting a tape or CD etc was associated with significantly higher levels of eyes directed inward and somewhat elevated levels of no hands on the steering wheel Adverse vehicle events were not significantly elevate
74. es was carried out in two phases Phase I involved the analysis of crash data to determine the major sources of distraction contributing to crashes and to provide input to the devel opment of a taxonomy of driver distractions The results of this effort were documented in a final report published by the AAA Foundation for Traffic Safety Stutts et al 2001 The objective of Phase II of the project was to collect information on the occur rence of these distractions in the general driving population and their effects on driving performance This was accomplished by installing small video camera units in volun teer subjects vehicles to automatically record driving behavior over a one week period then coding and analyzing the resulting data Key findings from the Phase II effort are summarized below followed by a discus sion of the limitations and strengths of the study and its significance for future pro grammatic activities Summary of Study Findings The Phase II study was intended to provide insight into 1 how often drivers engage in various distracting or potentially distracting behaviors 2 whether there are age and sex differences in drivers exposure to various distractions 3 whether driving conditions alter the frequency and duration of distractions and 4 the consequences of distractions on selected measures of driving performance The data were analyzed de scriptively using the software provided with the Video Pro dat
75. essed on the video taped data e Whether the drivers hands were on the steering wheel no hands versus one or both hands e Whether the driver s eyes were directed inside or outside the vehicle and e Whether there were any adverse vehicle events defined as any wanderings in the travel lane encroachments into another travel lane or sudden brakings The first two of these outcomes like all of the distraction variables were coded as event occurrences over time Thus our analysis would tell us whether drivers spent a greater proportion of their driving time with no hands on the steering wheel or looking inside their vehicle when using a cell phone compared to not using a cell phone The last outcome measure adverse vehicle events had no associated duration hence it was measured in terms of an event rate or number of adverse vehicle events per hour of driving Table 13 shows the distribution of these three outcome measures overall and only when the vehicle was moving As expected both the proportion of time no hands were on the steering wheel and the proportion of time eyes were directed inside the vehicle dropped when the observations were restricted to periods when the vehicle was in mo tion Interestingly drivers were more likely to drive with only one hand on the steering wheel than both hands a result that was especially characteristic of male and younger or middle aged drivers The three categories of adverse vehicle e
76. for activities such as removing bottle caps unwrapping take out food etc As noted in the preceding two tables ap proximately three fourths of the subjects ate and or drank something at some point during their three hours of recorded driving Eating and drinking were about equally common activities and together with preparing to eat or drink were coded 4 6 of the total recorded driving time for the 70 subjects Musiclaudio Some form of audio either music or talk was playing in the vehicles 71 5 of the time most often the radio Only four individuals did not listen to the radio or music at all Subjects adjusted their audio controls a total of 1 539 times or an average of 7 4 times each per hour of driving 1 539 207 2 coded hours of driving If the twelve hours for the four individuals who did not record time with the radio on is subtracted from the total hours driving the average number of music control manipulations per hour of driving increases to 7 9 These manipulations averaged 5 5 seconds each Smoking As shown in Table 6 only five subjects smoked at all while driving These five subjects lit 38 cigarettes cigars etc and were recorded smoking on 45 occasions Note that since the data were coded in one half hour segments some of the coding sessions could include smoking but not lighting or extinguishing This averages to 2 5 lightings and 3 0 periods of smoking per hour of driving for the five subjects who smoked
77. g to eat or drink And like the CDS our exposure data revealed manipulating audio controls to be a fairly common driver behavior However smoking and use of a cell phone were about as likely based on their total times observed while distractions by other occupants in the vehicle were relatively less likely Altogether there appears to be only a very modest correlation between the CDS crash data and the exposure data gathered in the current study Part of this can be attributed to known and unknown differences in the two coding schemes Although the Phase II driver distraction taxonomy was originally based on the CDS groupings there were a number of additions and refinements and it is not known how these might map onto the CDS data At the same time the CDS data contained large percentages of other and unknown distractions which may or may not have been captured in 59 60 the Phase II taxonomy But more importantly is the fact that the Phase II taxonomy variable distributions are based entirely on event durations and do not take into ac count other important characteristics of potential driving distractions such as intensity or expectancy Consequently we feel it is inappropriate to try to combine the CDS crash and Phase II exposure results in any attempt to gauge the relative level of risk associated with the various distractions DISCUSSION This project entitled The Role of Driver Distraction in Traffic Crash
78. good weather and on other roadway types and when the vehicle was stopped e g when at a drive through bank Baby as well as child distractions were more common during daylight driving under light traffic conditions and on neighborhood streets Adult distractions showed less variation by contextual variables e Drivers primarily conversed when other adults were present in the vehicle and to a much lesser extent when children were present Again it should be emphasized that without appropriate statistical testing no conclusions can be drawn regarding the significance of these patterns and their applica bility to the overall driving population Results in Table E6 with respect to vehicle movement whether the vehicle was stopped or moving at the time of the distraction are of particular interest Vehicle movement is likely to affect not only a drivers willingness to engage in certain distract ing behaviors but also the severity of the consequences of that behavior As an example drivers may be more likely to place cell phone calls when their vehicle is stopped in traffic and the consequences of doing so taking their hands off the steering wheel and eyes off the roadway are much less important than if the vehicle is moving in traffic Or as another example if drivers only choose to read when their vehicle is stopped then reading is not likely to be associated with adverse driving events such as encroach ing into an
79. he actual view depended somewhat on the geometry inside the vehicle Both the driver and cabin cameras were made invisible to the driver by placing an infrared filter over the face of the camera unit The filter eliminated visible radiation and passed the NIR radiation provided by an LED array Near Infrared Source An array of 8 NIR 880 nm LEDs spaced 25 inch apart pro vided illumination of the cabin which permitted video recording under low light con ditions Angular coverage was 120 degrees due to the use of LEDs with flat tops Road Camera A micro miniature low cost CMOS complementary metal oxide semi conductor camera was used to provide a low resolution image of the road ahead of the vehicle This camera was configured with a 2 5 mm lens which provided a 150 degree HFOV The purpose of this camera was to provide information about the driving con text The CMOS camera was chosen because it appeared to be less susceptible rela tive to CCD to image artifacts caused by direct sunlight Microphone A high sensitivity microphone was mounted on the inside of the camera box A 3 mm diameter hole was drilled in the box to optimize audio gain The audio signal was transmitted through the camera cable and fed directly into the audio input on the VCR 69 70 RECORDING UNIT Monochrome Quad Processor The video signals were fed into separate quadrants and combined into one video signal at the output of the processor Sanyo S
80. hing for something inside the vehicle and ma nipulating vehicle controls are two internal distractions that affected all drivers in our study sample In addition all but four drivers were observed manipulating their vehicle s music or audio controls External distractions were also quite common coded for 90 of the drivers as were conversations with other occupants in the vehicle coded for 80 of the drivers Conversing was coded regardless of whether it was identified as distract ing or not and likely closely mimics the presence of passengers in the vehicle Eating and drinking 73 and preparing to eat or drink 61 were also engaged in by a majority of participants All of these percentages declined only slightly when based solely on times when subjects vehicles were moving Almost two thirds of the study participants were observed reading and or writing at some time during their three hours of recorded driving however this number was reduced to 40 when based only on times when subjects vehicles were moving Groom ing behavior also declined when only moving vehicles were considered from 57 to 46 of participants One third 33 of the study participants talked on a cell phone at some time while driving while 36 dialed a cell phone and 16 received a call on their phone Overall 28 of the 70 participants or 40 used a cell phone in some way during their three hours of observed driving either dialing answering and
81. his important research area along with the reduction of multi stream in vehicle video data that might be applied to future research efforts This study is one of only a few to examine the full range of distractions contribut ing to crashes Other similar studies include an earlier analysis of the CDS data Wang Knipling amp Goodman 1996 a recent analysis of Pennsylvania crash data Pennsylva nia Joint State Government Commission 2001 and a study of fatal crashes occurring in Great Britain Stevens and Minton 2002 The current study is also believed to be the first to collect real world driving data on the frequency and duration of a full range of driving distractions and measures of their effects on driving performance The expo sure data not only provides a complement to the crash data but can serve as a compari son for data collected in laboratory and other more controlled settings as well as a baseline for future studies of driving behavior Recent research into driver distraction has been spurred by the tremendous growth in new in vehicle technologies that may adversely affect driving performance The Driver Distraction Internet Forum hosted by NHTSA in the summer of 2000 explored cell phones in vehicle navigation systems night vision systems wireless Internet and in formation and entertainment systems Over 9 500 persons visited the site on over 23 000 occasions Llaneras 2000 The National Conference of State Legislatures
82. hone with a built in pre amp which records audio within the cabin of the host vehicle and to a lesser degree outside the vehicle THE RECORD BOX The RECORD BOX is a ten 10 gallon du rable plastic box with a lockable lid which houses a videocassette recording system real time image processor and power supply Once properly interfaced with the host vehicle the RECORD BOX is controlled by the vehicle ignition system The RECORD BOX activates and begins recording once the host vehicle is turned on stops recording and deactivates when the host vehicle is turned off This process will be explained further under TRIGGER CABLE By allowing the system to turn itself off when the host vehicle is not in use THE RECORD BOX benefits from a much smaller power supply The RECORD BOX uses ten 10 hour VHS videotapes so the maximum recording time for the RECORD BOX is 10 hours The VCR is preprogrammed to automatically eject the tape at the end of the 10 hours so that it cannot record over itself The RECORD BOX has a lockable lid to pre vent tampering and will maintain integrity in temperatures up to and possibly beyond 190 Fahrenheit THE CAMERA CABLE The CAMERA CABLE connects the CAM ERA BOX to the RECORD BOX It is made of high quality Belden wire in a thick durable gray plastic casing This cable delivers audio and video signals from the cameras and microphones to the image processor and VCR in the RECORD
83. ile method is easy to generate it has no complex analytical formulas to estimate nor any tabulated chi square values to look up It is also very intuitive does not rely on distributional assumptions and is the mostly widely used bootstrap technique The bootstrap percentile method allows the bootstrap estimate of the distribution of the statistic to conform to any shape that the data suggest This allows confidence 93 94 intervals to be asymmetrical around the expected value of the estimate The two draw backs of the percentile method are 1 an adequate sample size n gt 30 is required and 2 a lot of bootstrap samples are required For a sample size of 70 we used 20 000 bootstrap re samples This should permit us to estimate the tails of the sample distribu tion of the statistics permitting appropriate estimates of confidence intervals Another potential bootstrap method that might have been chosen is bias corrected bootstrap techniques However these make a limited parametric assumption which may or may not be true for our data Also these techniques assume that the estimator in the original data set is an unbiased estimator of the population parameter That is to say that if there is selection bias then the estimator in the original data set is mot an unbiased estimator of the population parameter Thus the bias corrected bootstrap techniques should not be used in the presence of selection bias Whether the original sample di
84. ime one state and one state only can be active As an example one of the 16 channels was designated for recording cell phone use its four mutually exclusive and exhaustive states were 1 phone not in use 2 phone in use talking or listening 3 dialing phone and 4 ringing phone Since each distraction category was recorded in a separate channel multiple distractions could be coded simultaneously e g a person could be both talking on the cell phone and ma nipulating the radio controls or eating and drinking Separate channels were also used to record the various contextual circumstances lanes traffic level light conditions etc and to track the three outcome measures hand position direction of eyes and vehicle movement Events could be coded independently of states and had no asso ciated duration For example passing through an intersection was an event that would only be coded while the vehicle was moving a state Table 3 shows the overall data coding scheme developed for use with the Video Pro software Modifiers to a state were always coded whenever the state was activated and the comment fields were used for providing additional descriptive information such as the nature of an external distraction or what was being read Operational defi nitions used in the coding are provided in Table 4 Noldus Information Technology Inc Sterling Virginia For information on the Observer Video P
85. imitations in 13 659 119 13 504 268 ae 7 6 7 5 mind These include the limited sample size and recorded hours of 17 593 312 18 101 568 ee 9 7 10 0 driving especially with regard to Tatai pions 40 mead a6 any parsed data and the noted dif 50 3 49 7 ficulties in objectively coding the 1 Percent of all licensed drivers age 18 N 180 924 076 video data and resulting low inter rater reliabilities The results have not been weighted by driver age and sex since the stratified sample identified for the study very closely reflects the overall age and sex distribution of li censed drivers ages 18 and older in the nation see Table 5 With 35 male and 35 female subjects identified in five age categories each age sex combination contributed 10 percent of the total subjects and approximate driving time for our study Overall in 31 32 the U S 50 3 of licensed drivers ages 18 and older are male and 49 7 are female U S DOT 2001 The only age group that was overrepresented in our study were 50 59 year olds who only constitute 7 5 of males and 7 6 of females in the licensed driver population Additional characteristics of the study population drawn from the pre and post survey results are reported below Characteristics of the Study Population The two brief surveys that subjects were asked to complete while the video logging equipment was being installed and later removed from their vehicles provid
86. internal distraction Manipulating vehicle controls Falling object Insect distracting Event Pet distracting Event Reach lean look for etc Other internal distraction External distraction No external distraction External distraction Total Duration min 12246 0 26 1 2 0 154 4 11856 4 385 8 95 8 89 6 0 8 12228 0 2 6 195 8 2 1 12290 3 93 1 45 1 12312 0 44 6 34 8 37 1 10506 0 1922 5 11800 7 168 4 3 1 283 6 172 7 12136 9 291 6 Total Duration Mean Duration sec Minimum Duration sec Maximum Duration sec 11027 7 65 7 19 7 1264 2 11027 7 755 5 350 0 104 9 17 6 11027 7 10601 6 2209 4 2723 4 3412 9 80 3 11098 6 10 2 1043 2 23 7 11027 7 282 4 340 0 11098 6 192 6 1124 2 608 8 11098 6 4827 0 10848 6 770 5 Distractions by other occupants in the vehicle were less common among study par ticipants as was smoking Both were only slightly affected by whether the vehicle was stopped or moving at the time Table 7 presents the results for moving vehicles by the age of the driver and Table 8 by the driver s sex In general older drivers and especially those age 60 appeared less likely to eat or drink while driving and less likely to manipulate the vehicle s music or audio controls Younger drivers on the other hand appeared more likely to have been distracted by other occupants riding in their vehicle These results could not be
87. ither new nor technological in nature Rather they are aspects of everyday driving that people are likely to seldom think about A challenge for the highway safety community is to develop effective strategies for modifying people s driv ing behavior so that they do not engage in these potentially dangerous activities at inappropriate times while driving The human element is and always has been the most difficult to influence in the quest for increased safety on our roadways INTRODUCTION This report presents the results of Phase II of a project funded by the AAA Foun dation for Traffic Safety AAAFTS entitled The Role of Driver Distraction in Traffic Crashes The Phase I report for the project contained the results of an analysis of five years of national Crashworthiness Data System CDS data along with crash narrative data from both the CDS and North Carolina crash reports Stutts Reinfurt Staplin and Rodgman 2001 The CDS a part of NHTSA s National Automotive Sampling System is based on a national probability sample of approximately 5 000 police re ported traffic crashes involving at least one passenger vehicle towed from the crash scene The goal of the initial phase of the project was to use the available data to identify the major sources of distraction contributing to crashes and to develop a taxonomy of driver distractions that could be used to guide the Phase II efforts The second phase of the project called for de
88. l low traffic lights or checking the rear view mirror While these questions were prima rily asked to emphasize the general nature of the study they also provided some early indication of the levels of certain behaviors we might expect to observe on the video tape For example 32 9 of the respondents reported often always or occasionally talking on the phone while driving the remaining 65 2 falling into the never rarely category As might be expected drivers appeared to underreport less desirable behav iors such as cell phone use and smoking and overreport more desirable behaviors such as wearing seat belts and stopping for yellow lights While these results may be of interest they offer no real basis for judging the representativeness of the study sample The post driving survey provided an opportunity for questioning the respondents more directly about how having the video logging equipment installed in their vehicle might have influenced their driving behavior Ten of the 70 study participants 14 5 said that having the equipment installed in their vehicle caused them some problems Three drivers noted specific problems with the equipment itself suction cups coming loose loose wires or tape and two noted potential electrical or battery problems nei ther serious Two other participants commented on the slight obstruction of view caused by the equipment One noted that he was continuously asked to explain the e
89. le dial a cell phone carry on a cell phone conversation and perform grooming activities when their vehicle was stopped This suggests that at least to some degree drivers are choosing to engage in these activities at safer times on the roadway At the same time eating and drinking manipulating music controls smoking baby and child distrac tions and conversing were no more likely to occur when the vehicle was stopped than when it was moving There were also indications that distractions can negatively affect driving perfor mance as measured by higher levels of drivers having no hands on the steering wheel their eyes directed inside rather than outside the vehicle and their vehicles wandering in the travel lane or crossing into another travel lane Combining the exposure results from the Phase II analysis with the crash results from the Phase I analysis produced results that were inconclusive at least with respect to assessing relative safety risks for the identified distractions This was attributed to limitations in each of the data sources to known as well as unknown differences in identifying and recording driver distractions and to the definition of a driver distrac tion solely in terms of its duration There are a number of important limitations to this study The relatively small sample size 70 drivers and relatively small number of hours analyzed 3 out of 10 hours observed could limit generalizability Difficulty in
90. le to identify an initial taxonomy of driver distrac tions that had been reported as contributing to crashes In Phase II of the study we used miniature video cameras unobtrusively installed in subject vehicles to collect exposure data on these distractions along with observational data of their effects on some aspects of driving performance This section reports on the results of the analysis of this video taped data The primary research questions that we sought to address in our analyses were e How often do drivers engage in distracting behaviors e Are there age and sex differences in drivers engagement in distracting behaviors e Under what conditions do drivers engage in distracting behaviors e What are some of the consequences of the distractions on driving performance These research questions were ex Table 5 Licensed U S drivers age 18 and above by age group and sex ee i Source Federal Highway Administration descriptive tables and calculation of http www fhwa dot gov ohim hs99 dlpage htm amined through the generation of confidence intervals for proportions of interest using the bootstrap per centile method Both Video Pro software and SAS were used in the Age Group Females 20 025 913 19 095 528 11 1 10 6 analyses and output from both are 20 374 005 19 942 998 11 3 11 0 presented in this section 19 357 391 19 269 974 n 10 7 10 7 All results should be viewed with important study l
91. lled or dropped and stops when picked up Starts when drink is spilled or dropped and stops when picked up Audio is not on no sound is heard from radio CD etc default setting Audio is heard music talk or other audio can be heard from radio CD etc Starts when driver reaches to manipulate audio controls on dashboard and stops when hand is removed from controls Includes inserting removing CDs and tapes None of the below default setting Starts when reaches for lighter and stops when cigarette is lit and lighter is no longer being held in hand Starts when first begin to extinguish cigarette and stops once cigarette is extinguished or disposed of Coded for duration of time cigarette cigar etc is lit None of the below default setting Starts when first looks at reading material and stops when eyes look away from the material or whenever subject is actively writing not just holding pen in hand Starts when grooming activity is initiated and stops when grooming activity ends Includes brushing hair cleaning teeth putting on makeup looking in mirror etc Code if reading or writing and grooming simultaneously None of the below default setting For all levels starts when distraction first occurs and stops when it ends Distractions might include screaming or other loud noises sudden movements crying whining or complaining kicking the seat grabbing onto the driver emotional conversations or outbur
92. m the trunk Confirm that the camera box is securely attached to the windshield Confirm that the camera cable and trigger cable are securely routed through the interior of the vehicle e Remove any trash tools excess equipment from the interior of the vehicle CHARGING THE BATTERY PACK A sealed gel cel battery pack located underneath the VCR and image processor in the RECORD BOX provides power to the cameras micro phone VCR and image processor This battery pack is designed to power the incar recording sys tem for 10 15 hours and is rechargeable Disconnect the battery from the rest of the recording system and connect it to the provided battery charger Plug the charger into any avail able power outlet The green LED on the charger will light up When the green LED begins to blink the battery is fully charged The charger will not overcharge the battery so leave the bat tery on the charger whenever it is not in use Allow 48 hours for a complete charge Charge after each installation A 78 THE BATTERY CHARGER The in car recording system is designed to func tion optimally with the included BatteryMINDer 12 Volt Battery Charger Maintainer Conditioner This battery charger is polarity protected to protect the battery and can operate indefinitely on trickle charge keeping the battery at full power for extended periods of storage The BatteryMINDer generates a 3 26mHz resonant frequency pul
93. more likely to smoke when no passengers either adults or children especially were present in the vehicle but note that our smokers may or may not have carried passengers in their vehicle Subjects were only slightly less likely to read when light conditions were dark and traffic and weather conditions appear to have had little impact Reading was more frequent when driving on neighborhood and other streets which might include for example parking lots or driveways when other adults were not present in the vehicle and not surprisingly when the vehicle was stopped Grooming activities were more frequent when driving in bad weather conditions when other adults were not present in the vehicle as well as when the vehicle was stopped Reaching distractions were more likely to occur when light conditions were gray or dark when traffic was light when weather conditions were bad when no occu pants were present in the vehicle when traveling on other streets parking lots driveways etc and when the vehicle was stopped Similarly manipulating ve hicle controls another source of internal distraction occurred more often when traffic conditions were light on other roadway types and when the vehicle was stopped In contrast other internal distractions were more prominent under bad weather conditions as well as when the vehicle was stopped External distractions were more likely under conditions of daylight driving
94. n and Infrastructure U S House of Representatives May 9 2001 Available http www nhtsa dot gov nhtsa announce testimony distractiontestimony html Stevens A and Minton R 2001 In vehicle distraction and fatal accidents in En gland and Wales Accident Analysis and Prevention 33 539 545 Strayer D L Drews FA amp Johnston W A 2003 Cell phone induced failures of visual awareness during simulated driving Journal of Experimental Psychology Applied 9 22 23 Stutts J C Reinfurt D W Staplin L amp Rodgman E A 2001 The Role of Driver Distraction in Traffic Crashes Washington D C AAA Foundation for Traffic Safety Sundeen M 2002 Along for the Ride Reducing Driver Distractions Final Report of the Driver Focus and Technology Forum Denver CO National Conference of State Legislatures Treat J R Tumbas N S McDonald S T Shinar D Hume R D Mayer R E Stansifer R L amp Castellan N J 1979 Tii Level Study of the Causes of Traffic Acci dents Final Report Bloomington IN Institute for Research in Public Safety Report No DOT HS 034 3 535 79 TAC U S Department of Transportation Bureau of Transportation Statistics undated Our 67 68 Nation s Travel 1995 NPTS Early Results Report Washington DC Bureau of Trans portation Statistics Available through link at http www cta ornl gov npts 1995 Doc index shtml Wang J S Knipling R R amp Goodman M J
95. n drivers came in to have the equip ment installed in their vehicle In summary these results neither provide firm evidence of the representativeness of the study sample nor do they raise red flags about possible unrepresentativeness 33 34 Rather they suggest that participants engaged in generally normal driving activity and that they were not unduly influenced by the equipment installed in their vehicles Each of the research questions identified at the outset of this chapter is addressed in the report sections that follow A final section interprets these Phase II exposure data results in light of the Phase I crash data results Frequency and Duration of Potential Driving Distractions NUMBER OF DRIVERS OVERALL AND BY AGE AND SEX Table 6 provides information on the number and percent of subjects engaging in potentially distracting activities while driving Results are presented both overall re gardless of whether the vehicle was moving or stopped and only when the vehicle was moving They reflect any recorded incidence of a behavior during the approximately three hours of driving time coded for each of the 70 subjects without considering the actual number of occurrences or the durations of these behaviors Thus a subject who tried one time to place a call while driving would be coded as using a cell phone the same as a subject who placed calls and talked for most of the three hours Examining the overall results reac
96. nal cameras were also directed in towards the vehicle interior was not detected Almost a third of the drivers reported having had a close call while driving dur ing the past week although there were no reported crashes Several of the drivers re ported having to hit the brakes but in general it was not possible to tell from the drivers brief descriptions which of these events one might expect to detect on the video recordings assuming they were part of the three hours of coded data Only one of the close calls was specifically linked to driver distraction a driver reported witnessing a woman driving through a red light while using a cell phone just missing striking the car ahead Finally it should be noted that although one in four participants said that they chose to participate in the study primarily because they thought the topic was interest ing and worthwhile two thirds of the participants indicated that the 100 incentive played a role in their decision Thus it might be argued that the study attracted less well off drivers who were in need of the additional cash On the other hand the requirement that subjects drive a vehicle with trunk access through the back seat or a van or SUV likely favors more affluent persons driving more recent model vehicles In practice the restriction to vehicles with rear seat trunk access excluded few potential study partici pants either as part of the screening process or whe
97. ne Dialing cell phone Eating or drinking Preparing to eat or drink Manipulating music audio 33 controls radio CD etc 94 3 Smoking 3 2 8 6 5 7 7 Reading or writing 14 10 40 0 Grooming 28 6 Other vehicle occupant Baby distracting Child distracting Adult distracting Conversing Internal distraction Reaching Manipulating vehicle controls All other internal External distraction Percent of all subjects in sex group n 35 Percent of total subjects n 70 p lt 05 and p lt 01 based on chi square test of association with sex 2 37 38 Table 9 Frequency and duration of distracting events includes when vehicle stopped Potential Distracting Event Frequency Cell phone pager Phone not in use Dialing phone Answering ringing phone Talking Listening Eating or drinking Not eating or drinking Preparing to eat drink Eating bringing hand to mouth Drinking hand to mouth Spilled dropped food or drink Music audio Music audio not on Radio on Cassette tape on CD on Unknown music audio on Manipulating audio controls Smoking Not smoking Lighting cigarette pipe etc Smoking Finishing smoking Reading writing or grooming Not reading or grooming Reading writing Grooming Occupant distraction No occupant distraction Distracted by baby Distracted by child Distracted by adult Conversing Not conversing Conversing Internal distraction No
98. ng proportion of drivers engaging in that activity Also represents percentage of total time assuming all drivers engaged in the activity 3All categories except for falling object and insect etc in vehicle which were recorded as events without an associated duration to someone outside the vehicle looking at houses or pretty scenery toll booths from the Philadelphia area drivers drive through windows at banks or fast food restaurants work zone activity simply looking out the side window at something and bright sun glare We recorded few instances of being distracted by pedestrians children or ani mals outside the vehicle Whereas most of the coded distractions were fairly short dura tion events sun glare often lasted much longer increasing the average duration time for an external distraction At least one external distraction was coded for 90 of the par ticipants from Table 6 The overall count of 659 external distractions corresponds to an average of 3 2 external distractions per hour per driver based on the full sample of 207 2 coded hours ADJUSTED DURATIONS FOR MOVING VEHICLES The only information from Table 9 that can be replicated for moving vehicles only is that on the duration and percent of total duration of each activity This latter infor mation is presented in Table 10 Thus the first column in Table 10 shows that a cell phone was in use 1 3 of the time that the vehicle was moving based on the recorded drivi
99. ng descriptive results using the Video Pro software and then duplicate these with the SAS database Once the data files were finalized the analysis approach primarily involved using both the Video Pro and SAS software to produce descriptive tabulations and cross tabulations of the data Given that the longitudinal nature of the data did not meet the assumptions for classic statistical analysis methods confidence intervals for proportions and linear combinations of proportions e g differences in the likelihood of eyes di rected inward when dialing a cell phone versus not dialing a cell phone were con structed using the bootstrap percentile method Details of this procedure are contained in Appendix D Due to the rigorous nature of the procedure it was only applied to testing with respect to the consequences of distractions on driving performance In addition to the video data the results of the pre and post surveys completed by the subjects while the equipment was being installed and removed from their vehicle were entered into a separate Microsoft Excel database and analyzed descriptively using SAS These results are included with Appendix C RESULTS The primary goals of this project were to identify the major sources of driver dis traction and to examine their occurrences in everyday driving and their potential con sequences for driving safety From the Phase I analysis of NHTSA s Crashworthiness Data System CDS data we were ab
100. ng of all 70 subjects This is only slightly less than the 1 4 of total time regard less of whether the vehicle was stopped or moving shown in Table 9 These percent ages for moving vehicles were combined with information on the percentage of subjects engaging in each of the potentially distracting activities from Table 6 to produce the adjusted percentages shown in the last column of the table These adjusted results require careful interpretation The unadjusted percentage of total time column reflects what was actually observed in our sample of 70 drivers Thus for this population of 70 drivers a cell phone was in use 1 3 of the total time that their vehicles were moving However for those n 24 drivers who used a cell phone at all in their moving vehicle the cell phone was in use 3 8 of the time they were driv ing The first percentage reflects the observed population level of exposure to cell phone distractions while the second reflects the individual cell phone users exposure to cell phone distractions The latter also reflects the percent of total time exposed if all drivers had used cell phones at the same level as did our 24 subjects Table 10 only presents information on the total durations of various distracting or potentially distracting activities Some of these activities are by their nature more time consuming than others For example cell phone conversations and smoking a cigarette generally take longer than manipula
101. ng pills putting on gloves and a variety of other such activities These tended to be of shorter duration averaging 11 8 seconds Other occupant distractions There were a total of 243 recorded instances of drivers being distracted by other occupants in the vehicle most often by babies n 114 or children n 81 but also by other adults in the vehicle n 48 This information is best interpreted in light of the percentage of time passengers in these various age categories were being carried in the vehicles This information recorded as a context variable is available from the occu pant class listed in Appendix E Combining the various categories of occupants babies were carried in drivers vehicles a total of 13 6 hours children 18 1 hours and other adults 43 9 hours Thus the hourly rate of driver distractions for infants was 8 4 for children 4 5 and for other adults 1 1 Baby and child distractions were generally brief events lasting about 25 seconds on average while adult distractions lasted somewhat longer over 45 seconds However occupant distractions could also last for longer time periods as evidenced by the nearly 19 minute maximum length listed for a child dis traction Conversing Talking or carrying on a conversation with another occupant in the vehicle was coded whenever it occurred without attempting to judge whether it was distracting to the driver Conversations were recorded 15 5 of the time
102. nlighted roadways Headlights should be on Early morning or evening driving conditions lighted or unlighted roadways May be a mix of headlights on and off Use this code only for segments of tape that cannot be coded accurately due to poor lighting conditions Table 4 continued Neighborhood street 2 lane road 3 lane road Divided multilane road Interstate Other unknown road type Traffic light Traffic moderate Traffic heavy Vehicle moving Vehicle stopped Vehicle passing straight through intersection Vehicle turning at intersection Generally low speed low traffic streets with no center or edge lines 2 lane roadways typically with center line and edge lines 3 or more lane roadways undivided 3 or more lane divided roadways but not an Interstate roadway Interstate highways Other roadways that do not fit categories above including one way streets private and commercial driveways parking lots parking decks etc Specify in comments Consider same direction traffic level oncoming traffic level and amount of stop and go traffic Code light when no more than a few other cars on the road Code moderate when there is a moderate number of cars in both directions but traffic flows smoothly Code heavy when there is a line of cars in the travel lane and some stop and go driving similar to rush hour traffic Vehicle is in motion Code vehicle moving as long as can detect any moti
103. no If yes please describe See comments attached 5 How did you learn about this study Coded responses appear below 69 1 Newspaper or flyer 30 9 Word of mouth 87 88 6 Why did you choose to participate in the study Coded responses appear below 23 2 Thought interesting worthwhile etc 33 3 For the money 34 8 Both of the above 8 7 Other 7 Do you have any suggestions for improving data collection with future participants 23 4 yes 76 6 no If yes please tell us See comments attached 8 If we were interested in using some of your data for a public presentation could we contact you for permission to do this We would not use any individual data with out permission and would never reveal the names of participants 98 5 yes 1 5 no See comments attached 9 Please indicate how often you do each of the following while driving Never Rarely Occasionally Often Always a Wear seat belts 4 6 4 6 90 9 b Stop for yellow traffic lights 3 0 47 0 50 0 c Turn on headlights during daytime 48 5 28 8 22 7 when it is not raining d Check rear view mirror 0 0 13 6 86 4 e Check side mirrors 1 5 24 2 74 2 f Use horn 47 0 42 4 10 6 g Listen to radio 6 1 12 1 81 8 h Play CDs or tapes 39 4 28 8 31 8 i Eat or drink 21 2 53 0 25 8 j Smoke 93 9 3 0 3 0 k Talk on the phone 65 2 31 8 3 0 l Read other than check road maps 89 4 10 6 0 0 TH
104. nsive nonparametric technique for con structing confidence intervals It is not a statistic in the sense that the confidence inter vals are nota function of the data Bootstrapping differs from the traditional parametric approach to inference in that it employs large numbers of repetitive computations to estimate the shape of a statistics sampling distribution rather than strong distributional assumptions and analytic formulas Bootstrapping uses the computer to re sample an original sample extensively inductively arriving at an estimate of a statistic s sampling distribution The basic bootstrap approach is to treat the sample as if it is the popula tion and apply Monte Carlo sampling to generate an empirical estimate of the statistic s sampling distribution The sampling distribution of the estimate of the statistic can be thought of as the distribution of values of that statistic calculated from an infinite num ber of random samples from a population In this study from the sample of 70 drivers we drew 20 000 re samples Each re sample of 70 was drawn with replacement with each driver having an equal probability of selection The desired proportions were then calculated for each bootstrap sample The distribution of the 20 000 re samples is the basis for creating the reported confi dence intervals The percentile method was chosen because the bootstrap distribution of the conditional proportions is not normally distributed The percent
105. o save and exit TO ENGAGE THE SECURITY LOCK The VCR is equipped with a Security Lock function which disables all button commands to prevent any accidental changes to program ming or recording or to prevent tampering To engage the Security Lock while holding down button 15 press button 14 An image of a key will appear on the display panel Use the same procedure to disengage the Security Lock MAINTENANCE When the unit is first powered up moved from a cold area to a warm area or exposed to high humidity condensation may form inside the VCR If this happens the DEW INDICATOR dddd will blink The VCR can be damaged if operated during this time Allow the unit to sit plugged in for 1 2 hours until the DEW INDI CATOR shuts off The VCR is equipped with an internal auto matic head cleaning system Additional clean ing is not necessary Always inspect the videotape before installing the in car recording system again Play the video back and check for image quality and stability If problems arise contact TransAnalytics LLC Michael Mercadante 215 855 5380 mercadante transanalytics com THE COUNTERS POWER FAILURES DEW FAILURES AND USED TIME Press the MENU button until the POWER FAILURE DEW or USED TIME menu is dis played These menus display date time stamps for failures and a record of accrued record time Each time the in car recording system powers down with the host vehicle will count a
106. objectively defining the vari ous driver distraction and contextual outcome variables also made it hard to achieve high levels of inter rater reliability when coding the data Some potentially important variables could not be coded at all In particular we were unable to capture any mea sure of cognitive distraction which the literature suggests may pose the greatest risk to driving safety Consequently our study is not able to provide a definitive answer as to which activities or which driver distractions carry the greatest risks of crash involve ment 10 The National Highway Traffic Safety Administration NHTSA has estimated that driver inattention or distraction is responsible for 25 to 30 of police reported traf fic crashes or an estimated 1 2 million crashes per year Sundeen 2002 Shelton 2001 This study performed for the AAA Foundation for Traffic Safety is one of only a few studies identified in the literature to examine the full range of distractions contributing to crashes It is also believed to be the first to collect real world driving data on the frequency and duration of these distractions and measures of their effects on driving performance Although recent research has primarily been focused on the safety implications of wireless communications and other in vehicle technologies the results of both the Phase I crash data analysis and the Phase II field data study have demonstrated that many driver distractions are ne
107. of adverse vehicle events Notable exceptions to this trend were a lower proportion of eyes directed in when talking or listening on a cell phone and lower proportions of both eyes looking in and no hands on the steering wheel when smoking Smoking was also associated with a significantly lower rate of adverse vehicle events lane wanderings encroachments sudden brakings e Efforts to combine the exposure results from the Phase I analysis with the crash results from the Phase I analysis produced inconclusive findings with respect to relative crash risks associated with each of the identified driving distractions This outcome was attributed to limitations in both data sources to known as well as unknown differences in identifying and recording driver distractions and to the limited definition of a driving distraction solely in terms of its duration Limitations There were a number of important limitations to this study Foremost were prob lems in objectively defining all categories of driver distraction as well as context and outcome variables This made it difficult to achieve high levels of inter rater reliability when coding the data We were also not able to distinguish between different levels of intensity of a distraction since this would introduce additional subjectivity into the coding Some potentially important variables such as vehicle speed vehicle decelera tion and following distances could not be objectively coded from the da
108. ogate measure for driver at tention as are lane exceedances Sudden brakings were identified as a measurable alter native to near misses which along with actual crash occurrence is probably the most direct measure of driving performance While hand position is not in itself related to driving performance it can signify a potential loss of vehicle control in an emergency situation Other measures of driver attention such as following distance and vehicle deceleration rates would have been desirable but these could not be captured from the video data What is arguably the cleanest measure of driver attention driver workload as measured by performance on a peripheral detection task can only be captured in more controlled research settings Data Coding and Reduction The system selected for use in transforming the video data into a database for analysis was the Observer Video Pro developed by Noldus Information Technology Inc The software was specifically developed for annotating coding editing and ana lyzing behavioral processes from video The complete system includes both software and hardware components for processing either analog or digital video tape For our purposes we used the analog coding system The Video Pro system allows up to 16 channels of data to be coded simulta neously Within each channel coding options called states are required to be mutu ally exclusive and exhaustive i e at any one t
109. on Overall Moving Stopped Talking on cell phone 1 1 2 1 1 3 Dialing answering cell phone 0 2 0 4 0 2 Eating drinking 1 5 1 8 1 9 Preparing to eat or drink 3 2 2 8 3 1 Music audio on 71 5 71 0 71 4 Manipulating audio etc 1 1 1 1 1 1 Smoking 1 5 1 3 1 5 Lighting extinguishing 0 04 0 04 0 04 Reading 0 3 3 4 0 8 Grooming 0 3 0 8 0 4 Internal distraction Reaching 1 7 5 5 2 3 Manipulating controls 0 9 3 8 1 4 Other internal distraction 1 1 3 2 1 4 External distraction 1 6 6 4 2 3 Other vehicle occupant Baby 0 4 0 3 0 4 Child 0 3 0 2 0 3 Adult 0 3 0 4 0 3 Conversing 15 3 16 3 15 5 Overall 84 7 15 3 100 104 Funded through the generosity of and their Wy members AAA FOUNDATION FOR TRAFFIC SAFETY 607 14th Street NW Suite 201 Washington DC 20005 2003 202 638 5944 800 993 7222 e www aaafoundation org Printed in USA
110. on forward or backward Vehicle is stopped Code whenever vehicle passes straight through an intersection of roadways Do not code for private or commercial driveway intersections No associated duration Code whenever vehicle makes a turn at a roadway intersection No associated duration Outcome Measures Vehicle wandering in lane not across line Vehicle encroaching across lane line Vehicle braking suddenly sharply Both hands on steering wheel One hand on steering wheel Both hands off wheel Eyes outside vehicle Eyes inside vehicle Yawning Anger aggressiveness Drowsiness Code whenever vehicle strays from its normal path and approaches but does not cross one of the lane lines or what would be a lane line on an unmarked road Code whenever vehicle strays from its normal path and crosses over one of the lane lines or what would be a lane line on an unmarked road Code whenever the driver brakes hard enough that the vehicle stops or slows suddenly Listen for braking sounds and watch for bracing or rebounding of the body for clues Code whenever both hands are on the steering wheel Wrists or arms draped over or through steering wheel should be counted as hands on but noted in comments Code if only one hand is on steering wheel using definitions above Code if no hands are on steering wheel using definitions above Driving with legs knees is considered to be no hands on wheel Eyes are dir
111. ongitudinal nature of the data and its failure to meet the underly ing assumptions required of classic statistical methods testing the significance of each contextual variable crosstabulated by each distraction variable would have required 60 independent bootstrap analyses which was beyond the resources of the project What follows are some descriptive highlights reflective of the coded data as it appears in the Appendix F tables Due to the absence of statistical testing and relatively small sample sizes in many of the cells these results should be taken as suggestive only e Our 28 subjects who used cell phones 24 while their vehicles were moving were more likely to use them when light conditions were gray or dark when weather conditions were bad and when driving on Interstate roadways They were also more likely to use the phone when their vehicle was stopped and less likely to use it when other adults or children were present in the vehicle Table 11 Distribution of total recorded driving time by context variables from Video Pro Elementary Statistics Total Context Variable Duration min Total Duration Vehicle Movement Vehicle moving 10480 1 Vehicle stopped 1948 4 Light condition Light Dark Gray dim light Weather Condition Good weather 11861 8 Bad weather 566 7 Travel lanes Neighborhood street 2 lane roadway 3 lane undivided roadway Divided multilane roadway Interstate Other unknown roadway type Tr
112. or moving Outcome or driver performance variables included whether both the driver s hands only one hand or no hands were on the steering wheel whether the drivers eyes were directed outside or inside the vehicle and whether the vehicle itself wandered in the travel lane encroached across the lane line or braked suddenly A total of three hours of driving data was coded per subject Since the logging system was designed to capture up to ten hours of total driving time this typically involved sampling a series of one half hour segments of data for coding Altogether the 70 subjects produced 207 2 hours of coded data The data were coded as a stream of events in one tenth second intervals At any point in time each of the various distrac tion contextual and outcome variables was either on or off The data were analyzed descriptively using the Video Pro analysis software and were also converted into SAS data files for further analysis Given that the longitudinal nature of the data did not meet the assumptions for classic statistical analysis methods confidence intervals for proportions and linear combinations of proportions were con structed using the bootstrap percentile method Results revealed that distractions are a common component of everyday driving see table During their three hours of coded driving time nearly all subjects were observed manipulating vehicle controls such as air conditioning or window controls
113. orm A copy of the form is included in Appendix B While the equipment was being installed they were also asked to com plete a brief survey form see Appendix C Although the survey included some ques 19 20 tions relevant to the study e g on what type of roadways do you typically drive and do you carry children or other passengers in your car its primary purpose was to occupy the subjects time while the equipment was being installed in their vehicle No ques tions regarding driver distractions were included so as not to tip off the participant to the specific purpose of the study Participants were asked to wait inside the research center offices while the equipment was being installed This generally was accomplished within half an hour Before leaving subjects were scheduled for a time to return about a week later to have the equipment removed and were given a copy of the consent form with our contact information in case they experienced any difficulties with the equip ment In general subjects were told just to drive as you normally would drive When subjects returned to have the equipment removed from their vehicles they were given a second brief survey also in Appendix C to complete that included more detailed questions about their driving habits and their reactions to having the equip ment in their vehicle Removal of the equipment generally only took about 15 minutes after which subjects were paid and thank
114. other travel lane or sudden braking Certainly not all potential distractions are under the driver s control e g a baby crying a package dropping to the floor an emergency vehicle flashing its lights But to the extent drivers limit their engagement in potentially distracting activities to times when their vehicle is stopped any negative impact of these activities on driving performance and safety should be lessened Table 12 provides information on the percentage of time while engaged in a po tentially distracting activity that the vehicle was stopped Overall the vehicle was stopped 15 3 percent of the total recorded driving time The following distractions were associ ated with percentages stopped of 20 percent or more suggesting that at least to some extent drivers were choosing to engage in them at safer times while driving Reading writing 69 5 Manipulating vehicle controls 43 3 External distraction 41 4 Reaching leaning etc 36 6 Other internal distraction 34 8 Grooming 34 1 Talking listening on cell phone 25 7 Dialing cell phone 25 0 Distracted by adult 22 2 Not appearing on this list are distractions related to answering a cell phone eating and drinking listening to music audio manipulating music audio controls smoking con versing and distractions associated with babies and children traveling in the vehicle When engaging in these latter behaviors drivers vehicles were no more likely than normal to be stopped
115. overall From the occu pant context variable see Appendix E we know that occupants were present in ve hicles 30 3 of the time which would indicate active conversations about half the time another occupant was present in the vehicle Internal distractions An attempt was made to identify other potential sources or forms of distraction inside the vehicle for coding purposes The most frequently cited internal distraction was reaching leaning looking for picking up etc something inside the vehicle purse sunglasses sun visor glove compartment tissue garage door opener change for the toll booth etc This behavior was noted on 2 246 occasions or an average of 10 8 times 2 246 207 2 hours of coded data per hour of driving per subject Almost as frequent was manipulating vehicle controls other than the radio or music controls These might include heat and air conditioning controls window controls cruise con trol etc but not turn signals horn or other controls integral to the operation of the vehicle Manipulating vehicle controls was recorded a total of 2 095 instances or 10 0 times per hour per subject Reaching events lasted an average of 7 6 seconds while manipulating vehicle control events lasted an average of 4 8 seconds 4 42 Distractions by pets falling objects in the vehicle and insects or bugs were quite infrequent although it should be noted that only a few drivers carried pets in their vehicle
116. overrepresented in terms of their durations when the vehicle was stopped were in decreasing order from most to least overrepresented reading writing manipulating vehicle con trols external distraction reaching leaning etc other internal distractions groom ing dialing or talking on a cell phone and distractions by another adult in the vehicle This suggests that at least to some degree drivers are choosing to engage in these activities at safer times when their vehicles are stopped in traffic In contrast eating and drinking manipulating music controls smoking baby and child distractions and conversing were no more likely to occur when the vehicle was stopped than when it was moving e The effects of the various distractions on driving performance were measured in terms of whether they led to higher proportions of time spent with no hands on the steering wheel or eyes looking inside rather than outside the vehicle A third outcome measure was the number of adverse driving events primarily lane wan derings or encroachments into adjacent travel lanes but also sudden brakings per hour of driving time The bootstrap percentile method was used to calculate con fidence intervals around estimated proportions for each level of each distraction In general the various distractions were found to be associated with higher levels of no hands on the steering wheel eyes directed inside the vehicle and less fre quently higher rates
117. pant s only Rear seat occupant s only Both front and rear seat occupants Light Dark Gray dim light Unable to code due to tape conditions Neighborhood street 2 lane road 3 lane road Divided multilane road non Interstate Interstate Other unknown road type Modifiers Comments Handheld Handsfree Handheld Handsfree Handheld Handsfree Music type CD Tape Radio Unknown Specify in comments Specify in comments Specify in comments May specify in comments May specify in comments May specify in comments Specify in comments Specify in comments Specify in comments Specify in comments Specify in comments Specify in comments Specify in comments Number 1 9 Age Baby only Child only Adult only Baby child Baby adult Childt adult All ages Weather Good Bad Weather Good Bad Weather Good Bad 23 24 Table 3 continued Traffic level Vehicle movement Outcome Position Measures continued Eyes head drowsy aggressive Traffic light Traffic moderate Traffic heavy Vehicle moving Vehicle stopped Vehicle passing straight through inters Vehicle turning at intersection Vehicle wandering in lane Vehicle encroaching across lane line Vehicle braking suddenly sharply Both hands on steering wheel One hand on steering wheel Both hands off wheel Eyes outside vehicle Eyes inside vehicle Yawning Clear anger aggressiveness Clea
118. pe etc Smoking Conversing with another occupant in vehicle Reading or writing Grooming In addition to the categories of driver distraction the coding taxonomy also incor porates a variety of contextual variables to describe other occupants present in the vehicle weather conditions light conditions number of travel lanes level of traffic vehicle movement and whether the vehicle was passing through an intersection or turning at an intersection It was felt that these variables were important for describing the context or conditions under which drivers engage in various distracting activities information that could be key to making inferences about the extent to which drivers self regulate such behaviors according to real or perceived risk in the driving situation Finally the coding taxonomy incorporates three separate outcome measures 1 whether one hand two hands or neither hand is on the steering wheel 2 whether the 21 22 driver s eyes are directed outside the vehicle i e on the driving task or inside the vehicle and 3 whether the vehicle is swerving or wandering within the travel lane crossing into another travel lane or stopping from sudden braking These driver per formance measures were selected based on what has been suggested in the driver dis traction literature and what was feasible for us to code from our naturalistic data Llaneras 2000 Eye glance behavior is a commonly employed surr
119. phone Eating or drinking Preparing to eat or drink Manipulating music audio 14 controls radio CD etc 100 0 Number Subjects n 14 in each age group 30 39 14 100 0 40 49 14 100 0 50 59 12 85 7 Overall 50 71 4 41 58 6 10 64 71 4 91 4 pr O fe oa do ein dof e 7 1 28 e 6 a 7 a 9 a 6 40 0 32 ar i ri 0 a 9 ae 7 45 7 Reading or writing re ner Other vehicle occupant Baby distracting Child distracting Adult distracting Internal distraction Reaching 14 100 0 Manipulating vehicle 14 controls 100 0 All other internal 8 57 1 External distraction 12 85 7 Percent of all subjects in age group n 14 28 6 5 35 7 3 21 4 11 78 6 14 100 0 14 100 0 9 64 3 13 92 9 0 0 2 14 3 2 14 3 9 64 3 13 92 9 14 100 0 12 85 7 11 78 6 0 0 0 0 0 3 21 4 10 71 4 14 100 0 14 100 0 9 64 3 12 85 7 13 92 9 13 92 9 14 100 0 9 64 3 12 85 7 Percent of total subjects n 70 Table 8 Number and percent of subjects by sex who engaged in potentially distracting activities at any level during the three hours of recorded driving while their vehicle was moving Number Subjects Potential Distraction n 35 males and 35 females Male Female Talking on cell phone Answering cell pho
120. quipment to others and another that he was denied entrance to a military base because the locked equipment box could not be properly checked Although 15 respondents reported that their driving in the past week with the equipment installed was not typical of their usual driving the comments recorded proved unenlightening see Appendix C Of greater relevance were responses to the question of whether having the equipment in their car changed their driving in any way Fifteen participants 21 7 said that it did Of these eight indicated that it made them more safety conscious or more aware of their driving Interestingly the comments of five of the participants suggest that they were aware of the audio but not necessarily the video recording Made me more conscious of cussing I felt like I had to watch what I said This also came out in response to a later question about how to improve the study While the greatest number of comments pertained to the size of the boxes both the camera unit mounted on the windshield and more often the large locked box in the trunk housing the VCR processor and battery several participants sug gested that the study should study people and their driving as well and mount a camera facing out the back window Apparently a hole cut into the outward facing side of the camera box but covered with an infrared filter was detected as an opening for a camera but the fact that additio
121. r The rest of the CAMERA BOX should be hidden behind the mirror as best as possible Lick the suction cups and press them against the windshield in the proper location Once se cured to the windshield the CAMERA BOX can be angled by loosening the wing nuts on the sides Select the best angle then retighten the wing nuts This will likely have to be redone during ALIGNMENT so an exact angle is not neces sary at this time If at any time during installation the suction cups need to be removed from the windshield thoroughly re clean all surfaces before reinstall ing STEP 4 Route the Camera Cable Now the CAMERA CABLE will travel back to the trunk of the car where it connects to the Record Box For now simply drape the cable over the seats It will be secured nearer the end of in stallation Once the CAMERA CABLE is at the rear of the cabin locate the access point into the trunk see APPENDIX for more information Feed the cable into the trunk through this point All the excess CAMERA CABLE should be fed into the trunk so that the cable through the cabin is mod erately taught STEP 5 Install the Record Box in the Vehicles Trunk Cargo Area Open the trunk of the host vehicle Determine the best location for the RECORD BOX and place it into the trunk with the lid UP the lock facing the rear bumper and the hinge facing the cabin Try to rest it as low and as flat in the trunk as you can Locate the Cable conne
122. r drowsiness head jerk eyes drooping closed 1 Initial value for each variable is the default setting if a default is set E signifies an event having no time duration as opposed to a state with an associated time duration Table 4 Operational definitions used in coding the video data Distraction Variables Phone not in use Dialing phone Answering phone Talking Listening Not eating drinking Preparing to eat drink Eating Drinking Spilled dropped food Spilled dropped drink Audio not on Audio on Manipulating audio controls Not smoking Lighting Finishing smoking Smoking No read write groom Reading writing Grooming Read write groom No occupant distraction Baby distracting Child distracting Adult distracting Variable Level Definitions None of the below default setting Starts when first button on phone is punched and stops when last button is punched Starts when first hear phone ringing on tape and stops when person says hello Starts when person says hello and stops when person turns phone off or terminates call None of the below default setting Holding food or drink in hand unwrapping food opening food packages etc Starts when food is brought to mouth or mouth to food and stops when food or hand is removed from mouth Starts when drink is brought to mouth or mouth to drink and stops when drink is removed from mouth Starts when food is spi
123. ra facing out the back window Because of terrorists etc it might be nice if we as participants did not have to have complete trust and know nothing about the research Even being told the title of this survey would have been enough Somehow making the box smaller The size of the black box was a little distracting If it could be smaller it might be easier to forget APPENDIX D DESCRIPTION OF BOOTSTRAP PERCENTILE METHOD BOOTSTRAPPING THE CONFIDENCE INTERVAL FOR PROPORTIONS AND LINEAR COMBINATIONS OF PROPORTIONS Classic statistical methods for calculating confidence intervals in this study were not used because the underlying assumptions were not valid Classic statistical methods assume that the observations are statistically independent draws from a binomial distri bution This assumption is precluded by the longitudinal measures nature of the data Thus confidence intervals were computed using the bootstrap percentile method 20 000 bootstrap samples were drawn with replacement from the sample of 70 sub jects The percentile method was then used to construct the confidence intervals The percentile method was chosen because the bootstrap estimate of the sampling distribu tion was in general not distributed normally 20 000 re samples were drawn because the percentile method requires a large number of re samples in order to accurately estimate the tails of the sampling distribution Bootstrapping is a computationally inte
124. raction event when vehicle was moving Results of bootstrap analyses for adverse ve hicle event as a function of each distraction event when vehicle was moving LIST OF FIGURES Fig Pg 1 16 2 16 3 17 4 18 5 18 6 19 7 19 3 28 9 54 10 54 11 55 12 55 18 56 14 56 15 57 16 57 17 58 18 58 Diagram of interior of camera unit Diagram of interior of recording unit Photo of exterior of camera unit Photo of interior of camera unit Photo of recording unit Photo of recording equipment installed in vehicle Schematic of recording system Photo of data coding station Cell phone effects on driving performance Eating drinking effects on driving performance Music audio effects on driving performance Smoking effects on driving performance Reading writing effects on driving performance Grooming effects on driving performance Other occupant effects on driving performance Conversing effects on driving performance Internal distraction effects on driving performance External distraction effects on driving performance ACKNOWLEDGMENTS The authors express sincere appreciation to Ms Carol Martell who provided invaluable assistance in assimilating the necessary video recording and computer hardware and software for the video data coding and to Mr Brad Keady who assisted with the coding Their contributions to the project are greatly appreciated We are also indebted to the four anonymous reviewer
125. recording unit troubleshooting the system Data Collection Protocol The data collection protocol called for installing the video recording equipment in the vehicles of volunteer subjects for a period of approximately one week The goal was to capture at least six hours of usable data for each subject Half of the data was to be collected in the Durham Raleigh Chapel Hill area in the central Piedmont region of North Carolina and the other half in the Philadelphia area and its suburbs The first location could be expected to provide more rural and Interstate driving and the second more urban and suburban driving Although we had originally hoped to gather data on a much larger sample of sub jects N 144 due to the time requirement for coding the collected data the total number of participants was reduced to 70 35 from each of the two sites This number included 14 participants in each of five age groups 18 29 30 39 40 49 50 59 and 60 There were equal numbers of males and females in each age group Subjects were recruited prima rily through ads placed in local papers In addition flyers were posted to recruit specific subpopu lations of participants e g at a local senior center to recruit older adults and the Philadelphia re cruitment included a limited direct mailing to potential study partici pants In recruiting participants we did not identify the true purpose of the study but only that it was to lea
126. red filter on the dashboard Sit back comfortably in the seat and examine the camera angles carefully in the monitor De termine which cameras need to be repositioned and how ol 03 16 00 12 05 54 Sample in car recording image STEP 3 Focusing the Cameras Focusing the cameras is performed by rotating the lens clockwise or counterclockwise until the image comes into focus The two cabin cameras have set screws hold ing the lens in place This will need to be removed in order to change the focus Replace it afterward to hold the focus The road view camera will need to be pulled out to be refocused Pull firmly on the body of the road camera and slide it out of its plastic mount The lens is easily reachable now Adjust the focus using the same method as the other cameras Press the road camera back into place when focusing is complete STEP 4 Aligning Camera Angles The two cabin cameras are mounted on swivel ing bases To adjust the HORIZONTAL angle of these cameras loosen the screw that connects the camera to the bottom of the CAMERA BOX Swivel the camera into the appropriate position and then retighten the screw to hold the camera in place To adjust the VERTICAL angle of these cameras loosen the small set screw connecting the mounting hardware to the camera There is one screw on each side of the camera but loos ening one of these will allow the camera to be repositioned Correct the camera
127. ring Talking listening Eating or drinking Not eating or drinking Ref Preparing to eat drink Eating drinking spilling Music audio Music audio not on Ref Music audio on Manipulating audio controls Smoking Not smoking Ref Lighting or extinguishing Smoking Reading writing Not reading writing Ref Reading writing Grooming Not grooming Ref Grooming Occupant distraction No occupant distraction Ref Distracted by baby Distracted by child Distracted by adult Conversing Not conversing Ref Conversing Internal distraction No internal distraction Ref Manipulating vehicle controls Reach lean look for etc Other internal distraction External distraction No external distraction Ref External distraction Each variable level compared to reference Ref level e g talking listening on cell phone compared to phone not in use dialing answering cell phone compared to phone not in use etc N S non significant Table 16 Results of bootstrap analyses for adverse vehicle event as a function of each distraction event when vehicle was moving Vehicle 95 Potential Distracting Event Events per Confidence Significance Hour Intervals Level Cell phone pager Phone not in use Ref Dialing answering Talking listening Eating or drinking Not eating or drinking Ref Preparing to eat drink Eating drinking spilling Music audio Music audio not on Ref Music audio on Manipulating audio
128. rn how traffic and roadway conditions affect driving behavior Figure 6 Recording equipment installed in vehicle In a few cases subjects heard about the study from their coworkers or friends lt 150 degrees Road Camera View gt _ Microphone i a Cl Accessory Fuse Ground Connector 150 degrees 30 degrees Trigger Cable d BO a IR Source 4 Cabin Camera View Xoya Driver Camera View Camera Box Cable Camera Box NIR pass filter Quad Processor Sanyo VCR Battery Pack Figure 7 Schematic of recording system Potential study participants were screened via a brief telephone interview The primary require ments for participa tion were that they have a valid driver s li cense drive at least six hours per week and drive a vehicle that had rear seat access into the trunk through which the camera and trigger cables could pass They also had to be willing to come to the research offices to have the equipment installed and return a week later to have it removed The only other requirement for participation was that the subject s age and gender quota had not already been met All participants were compensated 100 paid in cash when they returned to have the re cording equipment removed When subjects came in to have the video equipment installed in their vehicle they were first required to read and sign a consent f
129. ro see their website at http www noldus com Table 3 Data coding scheme for use with Observer Video Pro software Class Driver Variable Phone Distraction pager Context Eating drinking Music Audio Smoking Reading writing or grooming Occupant distraction Conversing Internal distraction External distraction Occupants in vehicle Light weather conditions Travel lanes Code CX CC IX IM IF E Il E IP IR IO EX EE OX OF OR OB WL WD WG LN Definition Phone not in use Dialing phone Answering ringing phone Talking Listening Not eating or drinking Preparing to eat drink Eating Drinking Spilled dropped food Spilled dropped drink Music radio etc not on Music radio etc on Manipulating music controls Not smoking Lighting cigarette pipe etc Finishing smoking Smoking Not reading writing or grooming Reading or writing Grooming Reading writing and grooming No distraction from other occupants Baby distracting Child distracting Adult distracting Not conversing Conversing No internal event distracting driver Manipulating vehicle controls other than radio or other audio Falling object not food or drink Insect distracting Pet distracting Reaching leaning looking for picking up Other internal distraction No external event distracting driver External event distracting driver No other occupants Front seat occu
130. s a POWER FAILURE APPENDIX THE IMAGE PROCESSOR BF ae Sees ala 48 atte wattle iret io Fl HH se a H F iH m aij ye SO aa i tof OHH l Fiag i fs a i ey zH a Cial O n ny E 8 Wo gi Ma of E z wf x DESCRIPTIONS OF CONTROL BUTTONS 1 LED light on indicates unit has power 2 LED light indicates unit is in AUTO SEQUENTIAL SWITCHING MODE This light should NOT be lit 3 LED light indicates unit is in QUAD MODE This light should ALWAYS be lit 4 LED light indicates the small screen is frozen This light should NOT be lit 5 LED light indicates MENU will be displayed onscreen This light should NOT be lit once initial programming is performed 6 AUTO SWITCHING MODE Do not use 7 QUAD FULL SCREEN MODE Keep the unit in QUAD MODE at all times 8 MULTI FUNCTION BUTTON Do not use 9 MENU BUTTON Accesses various menus 10 CAMERA INPUTS 11 LOOP OUTPUTS Not used 12 VIDEO OUT Not used 13 TO VCR QUAD Connects to VCR 14 ALARM CONNECTOR Not used 15 POWER IN power from battery pack 16 GAIN CONTROLS to adjust contrast for each camera input Use with caution The image processor is a monochrome real time digital signal processor which can show up to four cameras independently or together in a quad display mode where each camera is shown in one quarter of the monitor image By connect ing the VCR directly to the TO VCR QUAD OUT conn
131. s of adverse vehicle events may be due to drivers choosing to read or write on less busy roadways neighborhood streets and other roadways including driveways and parking lots Figure 14 Grooming effects on driving performance 40 35 30 25 20 15 10 5 Not Grooming Reference Grooming Em No Hands E Eyes In Adverse Events Hour p lt 05 p lt 01 Grooming Grooming activities led to higher levels of both no hands on the steering wheel and eyes looking inward Adverse vehicle events also more than doubled but this increase was not significant statistically Figure 15 Other occupant effects on driving performance 30 25 20 15 10 5 o No Occupant Baby Distraction Child Distraction Adult Distraction Distraction Reference m No Hands m Eyes In Adverse Events Hour Other occupant distractions Distractions caused by babies children or other adults riding as passengers in the vehicle were all associated with higher levels of both eyes looking inward and adverse vehicle events although none of these results was signifi cant statistically In general these forms of distraction did not involve drivers having to take their hands off the steering wheel Figure 16 Conversing effects on driving performance
132. s of the draft report Their thoughtful and constructive feedback contributed substantially to the final product EXECUTIVE SUMMARY This report presents the results of Phase II of a project funded by the AAA Foun dation for Traffic Safety entitled The Role of Driver Distraction in Traffic Crashes The Phase I report for the project contained the results of an analysis of five years of National Automotive Sampling System Crashworthiness Data System CDS data along with crash narrative data from both the CDS and North Carolina crash reports Stutts Reinfurt Staplin and Rodgman 2001 The goal of this initial phase of the project was to identify the major sources of distraction contributing to crashes and to develop a taxonomy of driver distractions that could be used to guide the Phase II efforts The second phase of the project called for developing and validating a driving log method ology to determine the occurrence in the U S driving population of the various driver distractions identified in Phase I and to examine the potential consequences of these distractions on driving performance The methodology developed for the field data collection activities entailed a cam era unit containing three miniature video cameras a recording unit containing a VCR and battery packaged in a closed container cables for connecting the two units and a trigger device that automatically powered the units whenever the vehicle was turned on The camer
133. sation As a result of our inability to directly measure drivers level of cognitive attention our study is not able to provide a definitive answer as to which activities or which driver distractions carry the greatest risks of crash involvement At the same time it must be recognized that at least for the time being these sorts of measures can only be captured in more controlled research settings that lack the contextual and ecological validity needed to support generalizations about allocation of attention behaviors to the majority of real world driving situations Significance and Implications for Reducing Crashes A major strength of the current study is that it was carried out in the real world driving environment providing detailed data from a reasonably large sample of drivers about the activities that people engage in while driving These activities were shown to affect some aspects of driving performance that could also increase their risk or crash ing The project also had other strengths Foremost was the development of the video logging methodology itself and the demonstration of its feasibility and practicality for unobtrusively collecting real world driving data Another was the continued refine ment of a driver distraction taxonomy including more detailed levels of several distrac tions and identification of important contextual variables Finally much was learned about the practicalities of naturalistic data collection in t
134. se which is designed to break down lead sulfate crystals which form in side the battery pack This allows the battery pack to accept a full charge every time It is important to note the following safety pre cautions Do not expose the charger to rain or snow It is designed to operate INDOORS ONLY Discontinue use of the charger if it receives a sharp blow is dropped or in any other way damaged NEVER charge a frozen battery or one at a tem perature above 123 F Always allow the battery pack to achieve room temperature before charg ing The battery pack will emit small amounts of va pors Make sure the charging area is well venti lated Place charger as far away from battery pack as cables permit NEVER place the charger in di rect contact with the battery Keep the BatteryMINDer plugged in and con nected to the battery pack at all times the battery is not in use This will dramatically increase the functional lifetime of the battery packs If the green LED indicator does not start blink ing within 48 hours of being attached to a bat tery pack the battery pack may have a shorted cell or is too highly sulfated to accept a proper charge Remove the battery pack from use APPENDIX AUTOMOTIVE FUSES GLASS FUSES are found mainly in older cars pre 1980 Amp ratings are imprinted on the metal ends No glass fuses have been included Contact Technical Support if needed BLADE FUSES are found on most
135. sion 2001 In Great Britain a database abstracted from detailed reports on fatal crashes occur ring between 1985 and 1995 was examined for information on the role that in vehicle distractions play in fatal crashes Stevens and Minton 2002 In vehicle distractions were identified in 101 of the 5 740 cases available for analysis or 1 8 percent Leading the list were interactions with other passengers animated conversations horseplay and interaction with children or animals followed by entertainment devices prima rily radios cassette players and CDs and consumption of food drink and cigarettes Distractions by other controls within the vehicle old technology information devices such as vehicle instruments and maps and mobile telephones rounded out the list Given the age of the data 91 percent of the cases occurred prior to 1994 the authors note that new technology information devices were precluded from study however they hope to continue expanding the database for use in future such studies Stevens and Minton 2002 Many more studies have been carried out focusing on individual sources of driver distraction and in particular cellular telephones vehicle navigation systems and other in vehicle technologies The Internet Forum Llaneras 2000 described earlier provides a good summary of much of this research as does the recent report by the National Conference of State Legislatures Sundeen 2002 Most of these s
136. state crash report forms Although the proliferation of new in vehicle technologies certainly merits con cern the results of the Phase I report revealed many things distracting drivers and con tributing to crashes eating and drinking adjusting the radio tending to young children talking with a passenger smoking reading etc The focus of the current study is on the full range of events and activities that can draw a driver s attention away from the task at hand delaying recognition of safety threats and impairing effective control of the ve hicle Other than some earlier analyses of the CDS data Wang Knipling and Goodman 1996 the only other U S study to examine all of the various distracting events con tributing to traffic crashes is a recent report by the Joint State Government Commis sion of the Pennsylvania General Assembly 2001 For this report two years of Pennsylvania crash data 1999 and 2000 were examined and the driver distractions identified were categorized into a taxonomy similar to that employed in the CDS data and used in the Phase I results of the current study Stutts et al 2001 These results are summarized in Table 1 Table 1 Percentage distribution of specific driver distractions linked to crashes based on national Crashworthiness Data System CDS and Pennsylvania crash data Source Pennsylvania Joint State Government Commission 2001 United States Pennsylvania Driver Distraction CDS Data Da
137. stribution is a biased or unbiased estimate of the population distribution is unknown While the bias corrected bootstrap techniques will indeed correct for an original sample which is by chance a biased estimate of the population distribution this assumes that the estimator in the original sample is an unbiased estimate of the population parameter If there is selection bias or other biases in the study then the estimator in the original sample is not an unbiased estimator of the population parameter Thus if this assumption is not true then the bias corrected bootstrap techniques should not be used Is there selection bias All the subjects in this study were volunteers the potential is certainly there Thus there is potential for selec tion bias In contrast both of the shortcomings of the percentile bootstrapping technique can be reduced but not eliminated by using a lot of bootstrapping resamples This study employed 20 000 resamples In conclusion there is no best bootstrapping tech nique which one should be used depends on which assumptions the analyst is willing to believe References Mooney C Z amp Duval R D 1993 Bootstrapping A Nonparametric Approach to Sta tistical Inference Sage University Paper Series on Quantitative Applications in the So cial Sciences 07 095 Newbury Park CA Sage The JackBoot Macro 2000 Cary NC SAS Institute Inc Available http ftp sas com techsup download stat jackboot html
138. sts asking the driver to look at something etc Normal conversation is not considered to be a distraction Babies are children under age 2 children are between the ages of 2 and 13 adults are ages 13 and above based on estimated age since age information for passengers not available 25 Table 4 continued Not conversing Conversing No internal distraction Manipulating controls Falling object Insect Pet distracting Reaching leaning looking for picking up Other internal distraction No external distraction External distraction Context Variables No other occupants Front seat occupant s Rear seat occupant s Front and rear seat occupants Light Dark Gray dim light Unable to code Driver not engaged in conversation default setting Driver engaged in conversation Conversation is active as long as someone is responding within 10 seconds Singing is not considered conversation None of the below default setting Manipulating vehicle controls other than turn signal and gear shift which are considered integral to vehicle operation Includes temperature and air controls on dashboard window controls sun visor mirrors cruise emergency brake etc Starts when hand reaches for controls and stops when hand pulls back from controls Code for any falling object other than food or drink including packages that fall to floor dropped CDs dropped change things sliding off dashboar
139. t was made to further define this category since it encompasses such a wide range of possible distractions pedestrians animals road construction sun glare other vehicles on the roadway traffic signs emergency vehicles etc However comment fields were included for documenting the specific nature of this and other driving distractions Table 2 Taxonomy of driver distractions Original CDS Data Variables Outside person object event Adjusting radio cassette CD Other occupant Moving object in vehicle Other device or object Vehicle or climate controls Eating drinking Using dialing cell phone Smoking related Other distraction Revised Taxonomy for AAAFTS Study External distraction Nature of distraction specified in comment field Music radio etc on Manipulating audio controls inserting tape or CD etc Distracted by baby Distracted by child Distracted by adult Internal distraction Manipulating vehicle dashboard controls not gearshift Falling object not food or drink Insect distracting Pet distracting Reaching leaning looking for picking up something includes glove compartment Other internal distraction Preparing to eat or drink Eating bringing hand to mouth Drinking bringing hand to mouth Spilled or dropped food Spilled or dropped drink Dialing cell phone Answering cell phone Talking listening on cell phone Lighting cigarette pipe etc Extinguishing cigarette pi
140. ta 1995 19991 1999 20007 Outside object person or event Adjusting radio cassette CD Other occupant Moving object in vehicle Using other device object brought into vehicle Adjusting vehicle climate controls Eating and or drinking Using dialing cell phone Smoking related Other distraction Unknown distraction N ONOAANN A N o AHRONNAOW gt MSO Oo o OCONNSNN Total O O 1 Based on the 8 3 of drivers identified as distracted in crashes Based on the 3 5 of crashes involving a distracted driver Although only 3 5 percent of Pennsylvania crashes were identified as involving a distracted driver compared to 8 3 of drivers in crashes on the CDS datafile the hierarchy of distracting events was very similar with events outside the vehicle leading the list followed by adjusting the radio or audio system controls and other occupants in the vehicle Even in the more recent Pennsylvania data cell phone use was identified in just 5 2 percent of the distracted driving crashes The Pennsylvania data also in cluded information on whether the identified distraction was a primary or non pri mary contributory factor information that is not available on the CDS data Based partly on these findings the Pennsylvania Commission concluded that A statutory or regulatory restriction on specific driver distractions does not yet appear to be warranted based upon available data Pennsylvania Joint State Government Commis
141. ta at all There was also the substantial time and effort entailed by the coding that ulti mately led to reducing the sample size for the study from 144 to 70 subjects and to 63 64 only coding three hours of data per subject This contributed to increased variability in the data and the inability to perform some statistical tests There were additional challenges in analyzing the event time data so that the full richness of the data has yet to be mined For example we did not attempt to examine time lag events e g when Event A occurs what happens one half second two sec onds etc downstream of Event A or to define a visual outcome in terms of number and duration of glances inward which is arguably one of the better indicators of driver distraction Dingus 2002 We also did not compare one distraction e g talking on a cell phone directly to another e g conversing with a passenger Instead our analysis was focused on the overall frequencies and durations of the identified events Another important limitation of the study is that the measures of driving perfor mance we were able to code and analyze hands on steering wheel direction of eye focus and vehicle wanderings or encroachments across travel lanes have not been directly linked to crash risk While we may intuitively feel that drivers who engage in activities that require them to take their hands off the steering wheel or their eyes off the road for short
142. th regard to driver sex females were signifi cantly more likely than males to be observed grooming or attending to something outside their vehicle e Additional descriptive analyses of the data provided insight into the frequency and duration of potentially distracting behaviors while driving Some behaviors like manipulating audio controls were quite frequent but of short duration Others like smoking were less frequent but of much longer duration The average time to place a cell phone call was just under 13 seconds while the average time to answer a ringing phone was just under eight seconds Drivers averaged less than eight seconds reaching leaning looking for or picking up objects in their vehicles but did so on many occasions They also spent an average of nearly five seconds ma nipulating vehicle controls on nearly as many occasions e Altogether excluding any time spent conversing with other passengers in the ve hicle drivers were engaged in some form of potentially distracting activity up to 16 1 percent of the total time that their vehicles were moving assuming no over lap among the various activities Eating and drinking including preparing to eat or drink and holding food in one s hands headed the list at 4 6 of the total time vehicles were in motion This was followed by internal distractions at 3 8 and external distractions and smoking each at 1 6 Completing the list were ma nipulating audio controls 1 4
143. ting audio controls or taking a bite of a sandwich These latter activities however might occur more frequently thus increasing their total time Overall the most time consuming activities were conversing with other passen gers in the vehicle eating or drinking including holding food in preparation for eating or drinking and smoking It is interesting to note that even without including conversing the combined percentages shown in Table 10 for subjects engaging in the various potentially distract ing activities was 16 of the total driving time 31 including conversing While 43 some of these activities may have occurred simultaneously e g eating and at the same time noting something outside the vehicle this total nevertheless represents a signifi cant portion of overall driving time The adjusted totals present a less realistic view since they assume that a drivers are exposed to all potential distractions while driving i e they all smoke carry cell phones carry passengers in the various age groups etc The results presented in the following section examine the influence of various contex tual variables on driver distraction Context for Driving Distractions From the original data coding scheme shown in Table 3 seven contextual variables were identified In addition to whether the vehicle was stopped or moving these in cluded light condition weather condition travel lanes or road type traffic level
144. tion and internal distraction were probable causes in 93 of crashes environmental factors in 34 and vehicular factors in 13 Treat Tumbas et al 1979 The NHTSA has estimated that driver inattention or distraction is re sponsible for 25 to 30 of police reported crashes or an estimated 1 2 million crashes per year Sundeen 2002 Shelton 2001 A challenge for the highway safety community is to develop effective strategies for modifying people s driving behavior so that they are less likely to engage in potentially distracting activities at inappropriate times while driving While some distracting events are outside the driver s control e g the actions of another vehicle or a child s sudden cries most can be avoided by some simple precautionary measures children can be properly restrained in child seats CDs can be preselected cell phones can be switched off packages can be safely positioned on the floor of the car hot drinks and messy foods can be saved until the vehicle is safely stopped and driving directions can be reviewed ahead of time and perhaps written out in large print and taped to the dashboard Persuading drivers to change their behavior may be a more daunting task than improving vehicle design to more safely accommodate emerging technologies or build ing a more forgiving roadway In some cases legislation might provide the necessary push such as the recently enacted New York State law prohibiting the
145. tion event when vehicle was moving Percent No 95 Potential Distracting Event Hands on Confidence Intervals Significance Level Cell phone pager Phone not in use Ref Dialing answering Talking listening Eating or drinking Not eating or drinking Ref Preparing to eat drink Eating drinking spilling Music audio Music audio not on Ref Music audio on Manipulating audio controls Smoking Not smoking Ref Lighting or extinguishing Smoking Reading writing Not reading writing Ref Reading writing Grooming Not grooming Ref Grooming Occupant distraction No occupant distraction Ref Distracted by baby Distracted by child Distracted by adult Conversing Not conversing Ref Conversing Internal distraction No internal distraction Ref Manipulating vehicle controls Reach lean look for etc Other internal distraction External distraction No external distraction Ref External distraction Each variable level compared to reference Ref level e g talking listening on cell phone compared to phone not in use dialing answering cell phone compared to phone not in use etc 2N S non significant 51 52 Table 15 Results of bootstrap analyses for eye direction as a function of each distraction event when vehicle was moving Percent 95 Potential Distracting Event Eyes Confidence Looking In Intervals Significance Level Cell phone pager Phone not in use Ref Dialing answe
146. traffic lanes had checked rear view mirror to determine safety and it was OK and then a car that must have been speeding came up suddenly The only difference being I typically drive to Raleigh a couple of times a week for school More shopping trips Less of typical trips kid s art class I usually drive to and from work This past week I was on vacation and drove to Lancaster Philadelphia King of Prussia On the weekends Brian my fiance usually drives I did this weekend 89 90 I started classes I did more turnpike or expressway driving long trips to Atlantic City NJ and Scranton PA No we went to the shore We did no real interstate driving this past week which is our usual custom at least once per week i e NJ NY or Conn Drove father to emergency room Question 3 Did having the equipment in your car change your driving in any way Comments for yes responses I was more conscious of safety such as looking before changing lanes giving turn sig nals checking rear view mirror etc Made me more conscious of cussing It made me more cautious of how I was driving I felt like I had to watch what I said I couldn t remember if video was also attached so I tried to not pick my nose More cautious Probably drove more carefully Partially at the beginning one is aware of gadget on your windshield possibly staring at you Paranoia that my conversations were recorded For the most part
147. ttaches to the windshield of the host vehicle with two industrial grade rub ber suction cups Both suction cups and the wind shield will need to be thoroughly cleaned in order to insure a strong vacuum bond Rubbing Isopropyl alcohol works best as a cleaner Wet a paper towel with the alcohol or use alcohol cleaning pads and clean the insides of both suction cups Repeat using a clean pad or paper towel on the windshield This will remove any dirt grime and impurities on either surface If all surfaces are not thoroughly cleaned the suction cups may lose suction after installation allowing the CAMERA BOX to fall out of place STEP 2 Attach the Camera Cable The CAMERA CABLE has identical female screw type connectors with six pinholes at each end This cable transports audio video and power signals between the CAMERA BOX and the RECORD BOX Examine the CAMERA BOX There is a matching male screw type connector with six pins on the top of the CAMERA BOX The CAMERA CABLE attaches at this point There is a small indentation on the side of each connec tor Match these to one another and they should slide together Screw the two ends together se curely STEP 3 Attach the Camera Box to the Windshield Position the CAMERA BOX behind and below the rearview mirror as far as it will go The cabin cameras need only the lower inch of the box from which to see so only an inch needs to be exposed below the mirro
148. tudies have been carried out in controlled settings in laboratories on test tracks or using driving simula tors As a group they offer strong evidence that the new in vehicle technologies can negatively affect some aspects of driving performance however as pointed out by Dingus 1995 as cited in Stevens and Minton 2001 this may or may not translate into in creased crashes on the roadway What remains missing from the literature is data on drivers exposure to various potentially distracting events while engaged in everyday driving Without information on the frequencies with which drivers engage in these various behaviors and the cir cumstances of this engagement it is difficult to more precisely estimate their potential impact on driving safety Phase II of the current AAAFTS funded study was intended to address the need for real world data on driving distractions The study involved collecting unobtrusive video data from 70 volunteer subjects driving their own ve hicles over a period of a week The video data were coded using special software that 13 14 produced an event log for each 1 10 second of driving Analysis of the data provides insights into the range of events distracting drivers and their potential safety conse quences This report presents the results of the Phase II study The Methods section that follows provides more detailed information on the study methodology and on the for mation of the data analysis files
149. using a cell phone 1 3 other occupant dis tractions 0 9 reading or writing 0 7 and grooming 0 3 Again both event frequency and duration factored into these overall results e Some distracting events such as the various internal and external distractions are available to drivers any time they are behind the wheel of their car while others are contingent upon some other event For example drivers cannot be distracted by a cell phone unless a phone is with them in the vehicle If all of our drivers had had cell phones and used them at the same level as the participants in our study the adjusted percent of time spent using the phone including dialing and answer ing while in a moving vehicle would have increased to an estimated 3 8 percent e Along a similar line taking into account the smaller amounts of time that children and babies were carried in vehicles children were about four times and infants almost eight times more likely than adults to be a source of distraction to the driver based on number of distraction incidences e The occurrence of driver distractions also varied according to a number of contex tual variables such as whether the vehicle was stopped or moving at the time travel lanes or road type traffic level light conditions and weather conditions as well as the presence of passengers By far the most influential of these variables was whether the vehicle was stopped or moving Behaviors that were
150. veloping and validating a driving log methodology to determine the occurrence in the U S driving population of the various driver distractions identified in Phase I Information on how often and under what circumstances drivers engage in potentially distracting behaviors and how these behav iors affect driving performance can help the highway safety community in its efforts to reduce the number of crashes caused by distracted or inattentive drivers Driver distraction and its implicit effects on hazard recognition and vehicle con trol has been a prominent topic on highway safety agendas as well as for the U S Congress state legislatures the media and the public at large Much of this attention stems from the enormous increase in cellular telephone use by drivers and the prospect of similar growth in other in vehicle technologies such as vehicle navigation systems wireless Internet capabilities and wireless messaging In the summer of 2000 the Na tional Highway Traffic Safety Administration NHTSA sponsored an Internet Forum on the safety implications of such in vehicle technologies Llaneras 2000 Technical experts as well as the general public contributed to the Forum which drew widespread national and international participation In May of 2001 the U S Congress scheduled hearings before a House subcommittee on the topic of Driver Distractions Electronic Devices in the Automobile at which a number of presentations were made in
151. vents by definition only occurred while the vehicle was moving Lane wanderings were most common occurring at a rate of 4 3 per hour lane encroachments occurred at a rate of 2 1 per hour and sudden brakings at 0 11 per hour or just over one every ten hours of driving For the analyses that follow these three outcomes were summed to provide an overall Adverse Vehicle Event total Tables 14 16 present the results of 30 independent bootstrap calculations one for each outcome measure by distraction variable combination Table 14 shows the results for percentage of time no hands are on the steering wheel Table 15 for percent age of time the eyes are directed inside the vehicle and Table 16 for the total number of adverse vehicle events per hour The bootstrap procedures yielded estimated propor 49 50 tions along with 95 and 99 confidence intervals The 95 confidence intervals are shown in the tables along with the significance level for comparisons of each level of distraction with the reference in that category For example in Table 14 when com pared to not using a cell phone or pager both dialing answering and talking listening were associated with significantly higher levels of having no hands on the steering wheel dialing answering at the 01 level and talking listening at the 05 level In general the models reveal fairly consistent trends of higher levels of no hands on the steering wheel and eyes directed insi
152. y distracting 8 6 0 38 Child distracting 12 9 0 29 Adult distracting 22 9 0 27 Conversing 77 1 15 32 Reaching leaning etc 97 1 Manipulating vehicle controls 100 0 3 78 Other internal distraction 67 1 External distraction 85 7 1 62 Combined categories Age differences in the likelihood of engaging in a particular distraction on at least one occasion were generally small although sample sizes were too small for valid statis tical testing Compared to males females were more likely to groom themselves and to attend to things outside the vehicle Altogether excluding any time spent simply conversing with other passengers in the vehicle drivers were engaged in some form of potentially distracting activity up to 16 1 of the total time that their vehicles were moving Eating and drinking includ ing preparing to eat or drink and holding food in one s hands headed the list followed by internal distractions external distractions and smoking Less total time was devoted to manipulating audio controls using a cell phone other occupant distractions read ing or writing and grooming see table The occurrence of driver distractions varied according to a number of contextual variables with the most influential of these being whether the vehicle was stopped or moving at the time Thus drivers were more likely to read and write manipulate ve hicle controls attend to an external distraction reach for objects inside their vehic

Report - AAA Foundation for Traffic Safety

Contents

Download Pdf Manuals

Related Search

Related Contents