Home

Task analysis - ePrints Soton

Contents

1. This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 3 3 2 Utility of the method Within each system many of the same errors were identified for each task because the tasks consisted of similar steps All of the errors identified for both systems would have been likely to show up from analysis of only one or two representative tasks which would reduce analysis time considerably This should be a consideration for future development of error analysis techniques in this context SHERPA was useful for investigating VIS interactions in a dual task environment i e performing IVIS tasks at the same time as driving For example instances of incomplete tasks and failure to start tasks were predicted for situations in which the demand from primary driving was high however SHERPA provided no way of estimating the severity of these errors or the frequency with which they may occur Although SHERPA follows a fairly rigid structure for assigning errors the suggestions for remedial strategies for addressing those errors are likely to differ between analysts SHERPA would benefit from repeated analyses by different personnel on a small sample of representative tasks A focus group scenario comprising a mix of ergonomists designers and engineers would also be a useful addition to the method to generate more useful remedial strategies 3 4 Heuristic Analysi
2. 1 hr analysis 1 hr analysis per menu screen 6 8 hrs data analysis MCPA 2 4 hrs data collection 8 10 hrs data analysis SHERPA 2 4 hrs data collection 8 10 hrs data analysis These times are similar to those observed in previous studies which have applied these methods 22 31 with the exception of the application time for Layout Analysis which was slightly longer than that predicted by Stanton and Young 22 This is likely to be caused by difference in the interfaces tested in the two studies Dynamic screen based interfaces analysed in the current study comprise many different menu layouts in a single system and analysis is therefore likely to be more complex compared to the static dashboard mounted controls analysed by Stanton and Young In comparison with empirical methods which usually require a sample of users interacting with a prototype product the time and resource requirements of the analytic methods are significantly lower This supports their application at an early stage in the product development lifecycle Performance issues identified at this stage can then be further investigated if necessary using empirical techniques at a later stage of development when prototypes are more accessible Quantitative and qualitative information was extracted from each of the methods in order to make comparisons between the two interaction types under investigation These data are presented in table 12 34 This is the p
3. 2004 pp 103 112 55 Stanton N A and Stevenage S V Learning to predict human error issues of acceptability reliability and validity Ergonomics 1998 41 11 pp 1737 1756 56 Blandford A and Rugg G A case study on integrating contextual information with analyltical usability evaluation International Journal of Human Computer Studies 2002 57 1 pp 75 99 57 Bevan N and Macleod M Usability measurement in context Behaviour and Information Technology 1994 13 1 pp 1994 43 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 58 Burns P C Trbovich P L Harbluk J L and McCurdie T Evaluating one screen one control multifunction devices in vehicles Proc 19th International Technical Conference on the Enhanced Safety of Vehicles Washington DC 6 9 June 2005 pp 1 9 59 Butters L M and Dixon R T Ergonomics in consumer product evaluation an evolving process Applied Ergonomics 1998 29 1 pp 55 58 60 Cherri C Nodari E and Toffetti A Review of existing tools and methods AIDE deliverable 2 1 1 Volvo Technology Corporation VTEC 2004 61 Jeffries R Miller J R Wharton C and Uyeda K M User interface evaluation in the real world a comparison of four techniques Proc CHI 1991 Conference on Human Factors in Computing Systems Pittsburgh PA 15 20 M
4. 2810 E Z Make selection Make selection Q S 310 310 1300 2130 310 3120 Figure 4 Excerpt of MCPA diagram for touch screen audio task with operation timings given in ms In the MCPA diagram time flows from left to right therefore a succeeding operation which is dependent on the completion of a preceding operation is positioned to the right of the operation upon which it is dependent Parallel operations are located in the same vertical position in the diagram and are separated into rows to represent the different interaction modes visual manual cognitive auditory After modalities and dependencies are defined durations can be assigned to each operation In this study these operation duration times were derived from a review of the HCI literature and are listed along with their sources in Table 4 14 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Table 4 Operation timings from HCI literature Mode Task Time ms Reference Time used in model Stanton and Baber Visually locate single target Locate target on screen 1300 3600 2008 1300 Recognise familiar words or 314 340 Olsen and Olsen Visually locate sequential objects 1990 alphanumeric target 340 Visual Check if on screen target is 600 1200 Pickering et al Check if target is high
5. 1 3 2 1 Move pointer to AM tab 1 3 2 2 Press enter button 1 3 3 Open DAB menu Plan 1 3 3 Do 1 2 in order 1 3 3 1 Move pointer to DAB tab 1 3 3 2 Press enter button 1 4 Check for station in list Plan 1 4 IF station is one of six presets THEN 1 IF not THEN 1 4 1 Select radio station preset Plan 1 4 1 Do 1 2 in order Move pointer to radio station preset button de Axel g N 1 4 1 2 Press ENTER button 1 4 2 Search for radio station from frequencies list Plan 1 4 2 Do 1 2 3 in order 1 4 2 1 Move pointer to STATION LIST 1 4 2 2 Press enter button 1 4 2 3 Find and select station from list Plan 1 4 2 3 To search through lower frequencies THEN 1 to search through higher frequencies THEN 2 do 3 to select station 1 4 2 3 1 Scroll up through radio station frequencies Figure 3 Excerpt of HTA for play radio station task using remote controller IVIS 11 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 3 1 1 IVIS evaluation HTA is a task description method Task description is a necessary precursor for further analysis such as MCPA which will produce measurable results 22 HTAs for two or more systems may be subjectively compared in order to identify differences in task structure however this exercise is useful for task design exploration rather than as a method for cont
6. in the HTA were further categorised as visual manual cognitive or auditory MCPA was selected here in preference to other theoretical modelling methods because it is the only one capable of modelling parallel activities in multiple modes 26 37 38 Parallel activities can be described according to the multiple resource model 12 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 39 which proposes that attention can be time shared more effectively between operations across different modes compared with operations which utilise the same mode 39 Two visual operations for example locating a control on the vehicle s dashboard and reading a label on screen cannot occur in parallel however one of these visual operations may take place at the same time as a physical operation such as moving the hand towards the screen The structure of a MCPA model is also affected by the dependency of operations A dependent operation is one which cannot begin until a previous operation has been completed 26 40 For example the driver cannot press the enter key on the remote controller until the pointer has been moved to the on screen target Figure 4 shows an extract from a MCPA diagram for the touch screen IVIS task play radio station Each operation is represented pictorially as a node and the relationships between operations are denoted by their relative po
7. interface performance task structure interaction times error rates usability issues and interface design Principles to guide the selection of evaluation methods were defined in a previous study 18 consider the type of information required from the evaluation consider the stage of the design process at which evaluation is needed consider the resources required and those available and consider the people that will need to be involved in the evaluation Today usability evaluation is encouraged in academia and industry however there have been suggestions that it can be ineffective and even detrimental if applied blindly and according to rule This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset rather than as a method of encouraging thought and consideration in designers and developers 20 Automotive manufacturers also tend to employ two distinct approaches to IVIS evaluation driving performance measures in relation to safety of driving whilst using an IVIS and customer satisfaction measured by surveys 21 The analytic methods presented in this study were selected to meet a requirement for measures which give an indication of interface usability before a product is sent to market and which encourage designers to explore how the design of an interface influences the user experience A review of analytic methods was conducted and the five presented in
8. made This will be a focus of future work References 1 Fastrez P and Hau J B Editorial designing and evaluating driver support systems with the user in mind International Journal of Human Computer Studies 2008 66 3 pp 125 131 2 Hedlund J Simpson H and Mayhew D International conference on distracted driving summary of proceedings and recommendations The Traffic Injury Research Foundation and The Canadian Automobile Association 2006 3 Young K L Regan M A and Lee J D Factors moderating the impact of distraction on driving performance and safety in Regan M A Lee J D and Young K L ed Driver distraction theory effects and mitigation CRC Press 2009 Ist edn pp 335 351 4 Gu Ji Y and Jin B S Development of the conceptual prototype for haptic interface on the telematics system International Journal of Human Computer Interaction 2010 16 1 pp 22 52 5 Krems J F Editorial selected papers from the 2nd European conference on human centred design for intelligent transport systems IET Intelligent Transport Systems 2011 5 2 pp 101 102 6 Mitsopoulos Rubens E Trotter M J and Lenn M G Usability evaluation as part of iterative design of an in vehicle information system IET Intelligent Transport Systems 2011 5 2 pp 112 119 7 Pickering C A Burnham K J and Richardson M J A review of automotive human machine interface t
9. off between these factors 52 Layout Analysis was performed for a number of IVIS menu screens which were identified by the other analytical methods as having usability issues 3 5 1 IVIS evaluation A Layout Analysis for one example menu screen is presented to illustrate the process see figure 7 The MCPA showed that the task time for adjusting fan speed with the remote controller system was reduced when using hard controls compared with the screen based controls which suggested that the design of this menu screen was not optimal The most significant recommended design change to this menu screen was to reduce the size of the fan speed controls which had low frequency and importance of use and to increase the size of the air direction controls which were used more frequently In order to make a quantitative comparison between the two input types the number of layout changes made to each system was used as a metric The two poorest performing menu screens for each IVIS including the remote controller climate screen were identified according to the results of the other analytic methods Layout Analysis was performed on the four menu screens and the number of changes recorded There were eleven changes to the touch screen menus compared with eighteen for the remote controller menus This could be an indication that in their current forms the remote controller menu screens would produce a less effective and efficient interaction than the touch s
10. same target selection activity as illustrated in Figure 6 Taw Locate target Make selection Total task segment time 1300 720 570 2590ms Figure 6 Excerpt from remote controller MCPA diagram to show a single target selection segment The location options are of equal duration for both IVIS as this operation requires the user to visually locate a target on screen and the target and screen sizes were approximately equivalent for the two systems The difference in segment time is produced by the second and third operations in the sequence which involve the user either homing their hand fingers to the touch screen target and pressing the target or manipulating the remote controller to move the cursor to an on screen target and pressing the enter button on the side of the controller The touch screen operation times were based on the time for moving the hand 320ms and pressing a key 200ms reported by Stanton and Baber 37 and there is also some assumed movement of the hand which occurs in parallel with the visual search operation Previous studies have reported times of between 368ms and 583ms for physical selection of on screen targets combining movement of the hand and pressing a target 42 44 which are commensurate with those used in the current study 320 200 520ms Card et al 42 reported a time of 570ms for pressing a pushbutton 20 This is the post print after peer review version of the paper
11. systems Table 9a Issues identified by Heuristic Analysis of touch screen IVIS Negatives Positives Minor Major Minor Major Little use of colour coding Glare and reflections on Auditory feedback for Text information is easy screen touch screen button and quick to read and Climate menus are presses understand no long words cluttered Discomfort from holding arm outstretched when Easy to activate wrong control especially with small buttons and those which are close together Audio volume adjustable for some functions or cluttered buttons Easy to go back between menu levels and to return operating touch screen Pop up screens indicate to HOME menu No non visual location of extra information touch screen controls Activation of functions via Small delay in screen hard controls is confirmed response to touch for some by a message on screen functions Table 9b Issues identified by Heuristic Analysis of remote controller VIS Negatives Positives Minor Major Minor Major Menu exits after a relatively short time Colour coding not helpful Auditory feedback volume not adjustable for button presses Cluttered appearance of navigation screen Unclear labeling of buttons Temperature units not display Back button located in top right corner of screen not easy to access and no hard back button Relative complexity of navigation menus Usefu
12. this procedure was applied to both interfaces it is likely that the relative comparisons had high construct validity On the other hand we cannot be sure that the results represent accurate measures of absolute task times because they have not been validated against real interactions There is potential for the MCPA method to model absolute task times accurately if a comprehensive and valid database of IVIS operation types could be developed MCPA in its current form fails to address the issue of the dual task driving environment as it does not account for breaks in task performance caused by the driver s attention reverting back to the primary driving task Although static task times have been found to correlate well with eyes off road time 29 30 incorporating the split in visual attention into the model would produce more accurate predictions of IVIS task times in a dynamic environment This study has highlighted further potential for developing the 21 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset method to enable dynamic dual task environments to be accurately modelled This will be investigated in future work 3 3 Systematic Human Error Reduction and Prediction Approach SHERPA SHERPA was applied to the two IVIS and operations were classified into one of five types action retrieval checking information communication and selectio
13. this study were identified as most suitable in an IVIS context given the constraints of the automotive industry described above For a more detailed discussion of method selection see Harvey et al 18 Descriptions of the five methods and the related VIS performance factors are presented in table 1 Table 2 lists the inputs and outputs of each method Heuristic Analysis and Layout Analysis yield mainly qualitative data Multimodal Critical Path Analysis MCPA and Hierarchical Task Analysis HTA are used to generate mainly quantitative data and Systematic Human Error Reduction and Prediction Approach SHERPA produces both quantitative error rate and qualitative remedial strategies data 22 Both types of information are discussed in relation to all five methods in order to explore the most useful applications of each method e g for making direct comparisons or generating design recommendations This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Table 1 Analytic methods and related factors of IVIS interface performance Factors Analytical methods Description Task structure Interaction times Error rates Usability issues Interface design Hierarchical Task Analysis HTA Multimodal Critical Path Analysis MCPA Systematic Human Error Reduction and Prediction Approach SHERPA Heuristic Analysis Layout Analy
14. touch which were rated as significant was used as a metric by which to compare the two IVIS interfaces There was little difference in the number of significant error descriptions identified for the touch screen 6 and remote controller 7 25 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Table 8a Errors identified by SHERPA for touch screen IVIS including probability criticality and frequency ratings E 3 F Description a 1S 5 S 5 a g fa A2 Consecutive presses are too quick M L A4 System does not recognise touch H L 24 A4 Press centre console button with too little force M L 2 Repeat centre console button press too many times whilst A4 is L M 2 waiting for accurate feedback AS User moves hand to wrong area of screen M L 17 A6 Touch incorrect button or other part of screen M M 24 A6 Touch incorrect button or other part of centre console H M 2 Driver cannot remove hand from wheel due to high primary Be task demand M M a A8 Driver does not move hand back to steering wheel L L 9 A9 Driver starts to move hand towards screen but has to replace M H 9 on wheel due to sudden primary task demand Operation incomplete due to increased demand from pe primary task M M y R1 Visual check is not long enough to locate icon M L 26 R2 Incorrect icon is located by mistake L M 26 C1 Ch
15. which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset and this value was used in the remote controller model for the time to press the enter button on the side of the controller An assumption was made that pressing a hard enter key located on the side of the remote controller 570ms would take longer than touching a target on screen 200ms due to the increased resistance from the remote controller button and the reduced ease of access Card et al 45 reported positioning time for a mouse controlled cursor as 1290ms which included target selection via a button press The movement of the remote controller was very similar to a mouse and it was assumed that this value provided a good approximation of positioning time for the remote controller Time to press the enter button 570ms was subtracted from total mouse positioning time 1290ms to give a value of 720ms which was assigned to the positioning of the remote controller in the model This combination of positioning the cursor and pressing the enter button resulted in longer task segment times for the remote controller compared with the touch screen demonstrating that the nature of the interaction styles of the two devices were different and that this had an effect on total task times 3 2 2 Utility of the method MCPA enabled a quantitative comparison of task times to be made between the two IVIS following a structured procedure based on information from the HTA As
16. Critical path analysis for multimodal activity in Stanton N Hedge A Brookhuis K Salas E and Hendrick H ed Handbook of human factors and ergonomics methods CRC Press 2005 1st edn pp 41 41 41 48 41 Olson J R and Olson G M The growth of cognitive modeling in human computer interaction since GOMS Human Computer Interaction 1990 5 2 pp 221 265 42 Card S K Moran T P and Newell A The psychology of human computer interaction Lawrence Erlbaum 1983 first edn 42 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 43 Ackerman P L and Cianciolo A T Psychomotor abilities via touch panel testing measurement innovations construct and criterion validity Human Performance 1999 12 3 4 pp 231 273 44 Rogers W A Fisk A D McLaughlin A C and Pak R Touch a screen or turn a knob choosing the best device for the job Human Factors 2005 47 2 pp 271 288 45 Card S K English W K and Burr B J Evaluation of mouse rate controlled isometric joystick step keys and text keys for text selection on a CRT Ergonomics 1978 21 8 pp 601 613 46 Alliance of Automobile Manufacturers Statement of principles criteria and verification procedures on driver interactions with advanced in vehicle information and communication systems AAM 2006 47 Bhise V D
17. Dowd J D and Smid E A comprehensive HMI evaluation process for automotive cockpit design Proc SAE World Congress Detroit MI March 3 6 2003 pp 61 70 48 Commission of the European Communities Commission reccomendation of 26 V 2008 on safe and efficient in vehicle information and communication systems update of the European statement of principles on human machine interface Commission of the European Communities 2008 49 Japan Automobile Manufacturers Association Guideline for in vehicle display systems version 3 0 JAMA 2004 50 Nielsen J Ten usability heuristics 2005 cited 2009 20 03 09 Available from http www useit com papers heuristic heuristic_list html 51 Mendoza P A Angelelli A and Lindgren A Ecological interface design inspired human machine interface for advanced driver assistance systems IET Intelligent Transport Systems 2011 5 1 pp 53 39 52 Stanton N A and Young M S Is utility in the mind of the beholder A study of ergonomics methods Appl Ergon 1998 29 1 pp 41 54 53 Hancock P A and Szalma J L On the relevance of qualitative methods for ergonomics Theoretical Issues in Ergonomics Science 2004 5 6 pp 499 506 54 Mackay W E The interactive thread exploring methods for multi disciplinary design Proc 5th Conference on Designing Interactive Systems Processes Practices Methods and Techniques Cambridge MA 1 4 August
18. This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset The Trade Off Between Context and Objectivity in an Analytic Approach to the Evaluation of In Vehicle Interfaces Catherine Harvey Neville A Stanton Transportation Research Group Faculty of Engineering and Environment University of Southampton Highfield Campus Southampton Hampshire SO17 1BJ UK This paper presents a case study to explore an analytic approach to the evaluation of In Vehicle Information Systems IVIS usability aimed at an early stage in product development with low demand on resources Five methods were selected Hierarchical Task Analysis HTA Multimodal Critical Path Analysis MCPA Systematic Human Error Reduction and Prediction Approach SHERPA Heuristic Analysis and Layout Analysis The methods were applied in an evaluation to two IVIS interfaces a touch screen and a remote controller The findings showed that there was a trade off between the objectivity of a method and consideration of the context of use this has implications for the usefulness of analytic evaluation An extension to the Multimodal Critical Path Analysis MCPA method is proposed as a solution to enable more objective comparisons of IVIS whilst accounting for context in terms of the dual task driving environment 1 Introduction This case study explores the use of analytic modelling in the I
19. as implicit knowledge of the context of use 56 this is often not the case However Layout Analysis still adds to the analytic approach by providing a strategy for exploring existing GUI 37 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset layouts this is important as the GUI should be optimised with task structure and input device to produce ideal system performance It also provides designers with a structured method for addressing the types of usability issues identified by SHERPA and Heuristic Analysis 5 3 6 2 Context Versus Objectivity Usability evaluation should account for the specific context within which systems are used 18 however the results showed that not all of the methods addressed this issue HTA and MCPA were developed for application in a single task environment which means that in this case the effects of driving on IVIS effectiveness were not modelled Based on this case study it appeared that the more a method accounts for the broad effects of context the more subjective it becomes On the other hand a narrow and more objective focus produces quantitative models which enable direct comparisons between systems to be made 56 For example MCPA allows detailed quantitative comparable predictions for a very specific aspect of usability however the focus on only one aspect of system effectiveness task times in a singl
20. ay 1999 pp 119 124 62 Nielsen J Finding usability problems through heuristic evaluation Proc CHI 1992 Conference on Human Factors in Computing Systems Monterey CA 3 7 May 1992 pp 373 380 63 Nielsen J and Phillips V L Estimating the relative usability of two interfaces heuristic formal and empirical methods compared Proc CHI 1993 Conference on Human Factors in Computing Systems Amsterdam 24 29 April 1993 pp 214 221 64 Nielsen J Logic versus usage the case for activity centered design Interactions 2006 pp 45 44
21. ccount for the dual task driving scenario There is however potential to extend the method to address this issue SHERPA was expected to yield a comprehensive list of potential errors guided by its structured taxonomic approach however assessment of error frequency and severity are still largely open to analyst bias Data triangulation against the results of the heuristic analysis also showed that neither method was comprehensive Heuristic Analysis is not suitable for comparisons between systems however there is potential for development as product or brand specific guidance Heuristic Analysis also has an 39 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset advantages of low training and application times which supports its use for early identification of potential usability issues Layout Analysis appears to be useful for bridging the gap between evaluation and design and has only moderate time and resource demands which will enable analysts to not only make quick decisions about product performance but also to make recommendations to improve usability The findings of this exploratory study have highlighted a trade off between subjectivity and focus on context of use An extension of the MCPA modelling method has been suggested to incorporate analysis of context into a quantitative technique so that more useful predictions of IVIS performance can be
22. changes across two menu screens 18 layout changes across two menu screens Layout Analysis Qual In both devices menu targets with highest importance and frequency of use should be placed in the most accessible place on screen Sequence of use of targets in IVIS interactions should also be accounted for Touch screen Taking the quantitative data in isolation it could be concluded that HTA and MCPA support the use of the remote controller over the touch screen and the other three measures SHERPA Heuristic Analysis and Layout Analysis favoured the touch screen over the remote controller Exploration of the individual methods however has shown that it is not sensible to evaluate the IVIS based on this data alone and that 35 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset some of the methods were unsuitable for making direction comparisons between the touch screen and remote controller interfaces The findings of this study underline the importance of considering the relevance of outputs on a method by method basis 22 if the results are used solely to identify which system is superior then richer information about wider aspects of usability could be lost Gray and Salzman 19 warned that the advantages and disadvantages of analytic methods must be understood in order to mitigate against erroneous claims about system usabi
23. creen menus However Layout Analysis is highly subjective and in this study was more useful for producing design recommendations rather than direct comparisons of usability 31 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset INITIAL DESIGN FUNCTIONAL GROUPINGS C OUTSIDETEMP TEMPERATURE TEMP AIR ARDIRECTION e TEMP AIR DIRECTION FAN SPEED FAN SPEED IMPORTANCE OF USE SEQUENCE OF USE TEMPERATURE AIR DIRECTION AIR DIRECTION FAN SPEED FAN SPEED TEMPERATURE OTHER FREQUENCY OF USE REVISED DESIGN TEMPERATURE A C TEMP OUTSIDE TEMP TEMP T T T AIR DIRECTION AIR DIRECTION FAN SPEED E FAN SPEED zi Figure 7 Layout Analysis for the remote controller climate menu 32 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 3 5 2 Utility of the method Layout Analysis was included in the method set to provide a technique for specifying design changes based on Ergonomics principles It is not a useful technique for contrasting different systems as number of layout changes is very subjective and is therefore not a valid metric by which to make quantitative comparisons One use of the technique would be to bridge the gap between evaluation and design the selection of menus which require redesign
24. ct development The interfaces under investigation were a touch screen and a remote controller input device Figure 2 illustrates the typical layout of these interfaces in a right hand drive vehicle showing the position of the display screen and additional control pad this was only present in the remote controller system The schematic shows the approximate positions of the interface features and is not to scale The control pad had a similar function to a joystick moving a pointer on screen This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Remote controller Figure 2 VIS schematic 2 1 Procedure Expert walkthroughs of two existing VIS were performed by a Human Factors analyst These were based around a scenario of interacting with several in vehicle secondary tasks in a stationary vehicle The term task is used to refer to a sequence of operations performed by a user to achieve a goal such as selecting a radio station or reducing fan speed A single analysts applied all five methods reported in this study Heuristic Analysis was performed first whilst the analyst was interacting with each interface the other four methods were applied after the data collection phase using the information gathered from each IVIS in the order HTA MCPA SHERPA Layout Analysis The analyst trained in each of the methods prior to the data collect
25. de and moving from right to left back through the task diagram The LST of operation X is determined by the LST of the succeeding activity minus the duration of operation X 28 If there is more than one succeeding operation that links directly into operation X then the earliest possible LST should be used LST of operation X LST of succeeding operation Duration of operation X The LFT of an operation is determined by the sum of the LST and duration of an operation LFT of operation X LST of operation X Duration of operation X After calculating the values from the forward and backward passes the free time or float is calculated All paths through the task network with the exception of the critical path will have some associated float time Float time of operation X is the difference between the LST and EST of operation X Float time of operation X LST of operation X EST of operation X The final stage of MCPA involves defining the critical path and calculating total task time The critical path occurs along the path of operations which has the most minimal float time in figure 4 this is denoted by the solid lines The durations of all operations on the critical path are summed to produce the total task time 3 2 1 IVIS evaluation Total task times were calculated for the touch screen and remote controller interfaces and are presented in table 5 a
26. duct or task i e task structure and results in predictions about performance including task interaction times error rates usability issues and interface design 19 Analytic methods require data about the tasks users and system which comes from paper based specifications and expert knowledge In contrast empirical evaluation methods measure performance directly 19 using existing or prototype systems under simulated or real world conditions This study focusses on the application of analytic methods only Need Definition of usability criteria Selection of evaluation methods Usability prediction analytic methods HTA CPA SHERPA Heuristic analysis Layout analysis Usability measurement empirical methods Driving performance metrics System performance metrics Subjective performance estimates Subjective usability estimates Evaluation feedback informs revision of interface Usability criteria met No Final product Figure 1 Framework for usability evaluation adapted from Harvey et al 18 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset The aim of this study was to explore the usefulness of analytic methods for interface usability evaluation in the context of IVIS Analytic methods were selected to meet a requirement for an approach to evaluation which can be applied at an early stage of product develop
27. e task environment means that contextual factors are not accounted for 57 Subjective techniques enable a broader approach which aims to capture the whole essence of user device interaction 22 and these methods therefore account for context to some extent However the qualitative nature of the outputs means that these methods do not drill down to a deep level of detail and are therefore more suited to usability checks e g heuristic analysis or design recommendations e g layout analysis and SHERPA rather than direct comparisons 38 58 63 5 3 6 3 Extending MCPA To address the trade off between context and objectivity an extension to MCPA which allows consideration of the context of use is proposed MCPA measures performance via quantitative predictions of task time rather than relying on the assumption that poor performance will follow on from identification of usability issues 19 51 57 Another advantage of MCPA is that it takes a taskonomic approach to modelling HMI 64 which means that systems are analysed in terms of the activity or task being 38 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset performed On the other hand the heuristic checklist applied in this study took a taxonomic approach because it analysed elements of an interface based on functional rather than task based categories 22 36 Nielsen 64 ar
28. echnologies and techniques to reduce driver distraction Proc 2nd IET International Conference on System Safety London 22 24 October 2007 pp 223 228 8 Pereira M Hamama H Dapzol N Bruyas M P and Sim es A Simultaneous interaction with in vehicle systems while turning left comparison among three groups of drivers IET Intelligent Transport Systems 2009 3 4 pp 369 378 9 Miller R B Human ease of use criteria and their tradeoffs IBM Corporation 1971 40 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 10 Dehnig W Essig H and Maass S The adaptation of virtual man computer interfaces to user requirements in dialogs Springer Verlag 1981 Ist edn 11 Dzida W Herda S and Itzfeldt W D User perceived quality of interactive systems Proc 3rd International Conference on Software Engineering Atlanta GA 10 12 May pp 188 195 12 Shackel B Ergonomics in design for usability Proc 2nd Conference of The British Computer Society Human Computer Special Interest Group York 23 26 September 1986 pp 44 64 13 Norman D A Design principles for human computer interfaces Proc SIGCHI Conference on Human Factors in Computing Systems Boston MA 12 15 December 1983 pp 1 10 14 Nielsen J Usability engineering Academic Press 1993 1st edn 15 Shneiderman B D
29. eck omitted L L 9 C2 Check is not long enough to obtain accurate feedback L M 2 S2 Wrong selection made L M 26 TOTAL ERRORS 207 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Table 8b Errors identified by SHERPA for remote controller IVIS including probability criticality and frequency ratings 3 3 3 Description a O 5 a x g ea A4 Press button with too little force L M 22 A4 Repeat button press too many times whilst waiting for accurate L M 2 feedback A5 Pointer misses icon button letter number H L 20 A6 Select incorrect icon button letter number M M 20 A6 Press down controller instead of enter button located on side of H L 20 controller A6 Touch incorrect button or other part of centre console H M 2 Driver cannot remove hand from wheel due to high primary task a demand M M 1 A8 Driver does not move hand back to steering wheel M L 9 A9 Driver starts to move hand towards controller but has to replace on M H 9 wheel due to sudden primary task demand A9 Driver cannot locate controller after physical search L H 7 R1 Visual check is not long enough to locate icon M L 22 R2 Incorrect icon is located by mistake L M 22 C1 Check omitted L L 7 C2 Check is not long enough to obtain accurate feedback L M 1 S2 Wrong selection made L M 22 TOTAL ERRORS 194
30. ed on the analyst s experience and knowledge of the VIS tasks There are also a number of rules and assumptions which support the use of these timings in the MCPA models 15 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Time to visually locate a target is 1300ms following 37 for any single target and the first alphanumeric target in a sequence Time to visually locate a target is 340ms for any sequential alphanumeric target after the first target in a sequence It is assumed that users would be more familiar with the layout of an alphanumeric keyboard than with the other menu screens in each system therefore search time for alphanumeric targets was reduced to 340 ms following the time to recognize familiar objects reported by 41 No cognitive make selection operation occurs in parallel with a sequential alphanumeric visual search 340ms following the heuristics for Mental operators devised by Card et al 42 Entering a word into the system is assumed to be a single chunk users make a decision about the sequence of letters or numbers at the start of the chunk therefore individual decisions for each subsequent alphanumeric entry are assumed to be unnecessary There is always some movement of the hand fingers touch screen or the cursor remote controller during visual search This movement follows the direction of ga
31. eeds to be evaluated with real users A method for evaluating an interface based on the position of related functions according to frequency sequence and importance of use 24 It is related to the dual task criterion because the location of an IVIS in relation to the driver will affect the optimisation of layout It is also related to frequency of use because familiarity of users with the interface is a factor which determines layout This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Table 2 Inputs and outputs for analytic methods Inputs Quantitative outputs Qualitative outputs HTA Task specification Number of operations Understanding of task hierarchical task structure goals and plans MCPA HTA Task interaction times Operation dependencies SHERPA HTA Error types and frequencies Remedial strategies Heuristic Experience of system task Number of usability issues Types of usability issues Analysis specification identified potential problems Layout System layout diagrams Number of layout changes Changes to interface layout Analysis required 2 Method An evaluation of two existing VIS was performed using the five analytic methods in order to explore the utility of this approach in terms of information inputs and outputs training times resource demands and possible extensions in the context of early stage produ
32. esigning the user interface strategies for effective human computer interaction Addison Wesley 1992 2nd edn 16 Bevan N International standards for HCI and usability International Journal of Human Computer Studies 2001 55 4 pp 532 552 17 ISO 9241 Ergonomic requirements for office work with visual display terminals VDTs part 11 guidance on usability 1998 18 Harvey C Stanton N A Pickering C A McDonald M and Zheng P A usability evaluation toolkit for in vehicle information systems IVISs Appl Ergon 2011 42 4 pp 563 574 19 Gray W D and Salzman M C Damaged merchandise A review of experiments that compare usability evaluation methods Human Computer Interaction 1998 13 3 pp 203 261 20 Greenberg S and Buxton B Usability evaluation considered harmful some of the time Proc CHI 2008 Conference on Human Factors in Computing Systems Florence 5 10 April 2008 pp 21 Brostr m R Bengtsson P and Axelsson J Correlation between safety assessments in the driver car interaction design process Applied Ergonomics 2011 42 4 pp 575 582 22 Stanton N A and Young M S A guide to methodology in ergonomics designing for human use Taylor and Francis 1999 1st edn 23 Kirwan B and Ainsworth L K A guide to task analysis Taylor and Francis 1992 1st edn 24 Stanton N A Salmon P M Walker G H Baber C and Jenkins D P H
33. et however with the remote controller system the Settings menu is eliminated and the user moves directly from the Audio menu to the Sound screen This analysis shows that the time differences between the two IVIS for the air direction navigation and increase bass tasks resulted from differences in task design between the two systems in other words it is the extra steps involved in the touch screen tasks which were responsible for the observed differences in task times rather than differences in the nature of the interface The effect of interface design can however be seen in the predicted times for the other tasks including play radio station and reduce fan speed MCPA predicted shorter times for these tasks with the touch screen compared to the remote controller When the individual task segments are examined it appears that the nature of inputs to the touch screen system supports quicker performance because the individual operations have shorter durations as illustrated in figure 5 Teo J Locate target Cio Move hand Total task segment time 1300 320 200 1820ms Figure 5 Excerpt from touch screen MCPA diagram to show a single target selection segment 19 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset This can be compared to a task segment from a remote controller MCPA diagram showing the
34. f the concept of usability in particular see Shackel 12 Norman 13 Nielsen 14 Shneiderman 15 Bevan 16 and The International Organization for Standardization 17 One of the most widely used definitions is found in ISO 9241 part 11 guidance on usability 17 which defines usability as The extent to which a product can be used by specified users to achieve specified goals with effectiveness efficiency and satisfaction in a specified context of use 17 p 2 The reason for the wide adoption of this definition is probably the reference to context of use Despite many attempts to produce a universal definition of usability the importance of the context of use in defining specific usability criteria means that definitions need to be constructed according to characteristics of individual products and the environments in which they are used 6 1 2 Usability evaluation Harvey et al 18 proposed a framework for the evaluation of IVIS as shown in figure 1 The framework follows a mixed methodology approach incorporating analytic and empirical methods to model different aspects of usability Analytic methods can be used to evaluate IVIS without needing to simulate the This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset interaction with real users This evaluation involves the examination of the intrinsic features of a pro
35. gued that both taxonomies and taskonomies are necessary in design however in a dual task driving context where interaction with secondary tasks is so dependent on the concurrent demand from driving the activity based approach 36 appears to be the most useful for usability evaluation MCPA in particular has potential for analysing these dual task interactions because the driver s interaction with primary driving tasks can be incorporated into the models in parallel to VIS operations This technique could be used as a direct measure of the effectiveness of the user system interaction in a dual task driving environment 5 Conclusions The aim of the case study presented in this paper was to explore an analytic approach to IVIS usability modelling to meet a requirement for early stage low resource product evaluation The methods were selected to model important aspects of HMI performance task structure interaction times error rates usability issues and interface design 19 The findings of the study have been discussed in terms of TVIS comparisons utility of the methods time and resource demands and potential for further development HTA was not useful for making relative comparisons between systems however it was found to be an essential starting point for MCPA and SHERPA and was also useful for the exploration of task structure MCPA modelled task interaction times as a measure of performance however in its current state it does not a
36. hich 35 is an example however it is possible that heuristics could be aimed at an earlier stage in design eliminating the need for high fidelity prototypes For example Nielsen s Ten Usability Heuristics 50 encourage a more general approach to usability evaluation which could be applied in the very earliest stages of product development Based on these limitations it is proposed that Heuristic Analysis could be a useful tool for reminding designers about important usability issues 38 51 rather than for making direct comparisons between interfaces The technique has potential for further development by individual automotive manufacturers for making checks to a design to ensure that certain brand or product specific targets have been met The flexibility of Heuristic Analysis means that specific usability criteria defined by manufacturers for particular products could be built in to a checklist 3 5 Layout Analysis Layout Analysis is a technique used to evaluate an existing interface based on groupings of related functions 24 It can assist in the restructuring of an interface according to the users perceived structure 30 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset of the task Functions are grouped according to three factors frequency importance and sequence of use 24 and a revised design is based on the optimum trade
37. ion phase and spent approximately 4 5 hours using the two IVIS interfaces before applying the methods A set of nine typical IVIS tasks was defined for this study as shown in table 3 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Table 3 IVIS tasks analysed in evaluation Categories Tasks Audio Play radio station 909AM radio is currently set to 97 9FM Increase bass by two steps Climate Increase temperature by 1 C via centre console controls not IVIS Reduce fan speed by two steps Direct air to face only air direction is currently set to face and feet Direct air to face and feet air direction is currently set to windscreen only Activate auto climate via centre console controls not IVIS Navigation Set navigation destination from system memory Home Set navigation destination from previous destinations University Road Southampton These nine tasks were selected from a set of over 130 tasks which were identified for existing VIS from a review of automotive manufacturers IVIS manuals which was conducted by the analyst prior to data collection Four factors defined by Nowakowski and Green 33 were used to guide task selection use whilst driving availability in existing systems frequency of interaction and compliance with the 15 second task time rule see 29 The nine tasks were all likely to be used whilst d
38. is based on the results of the analytic models MCPA and SHERPA and the redesign is aimed at addressing the issues identified Layout Analysis would also be useful at very early stages of design before the prototyping phase to assist in initial layout decisions 22 4 General discussion The analytic methods applied in this case study were selected to model the performance of two IVIS interfaces Training data collection and application times were estimated based on the current study these will be useful to designers and analysts in future applications of these methods Training time estimates are presented in table 10 and data collection and analysis times are presented in table 11 Table 10 Training time estimates for the analytic methods Not much time Some time Lots of time Heuristic Analysis Layout Analysis MCPA Familiarisation with checklist Learn layout factors Learn rules calculation method lt 1 hour 1 2 hours gt 2 hours SHERPA Familiarisation with error codes gt 2 hours HTA Learn structure and notation gt 2 hours 33 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Table 11 Data collection and application time estimates for the analytic methods Not much time Some time Lots of time Heuristic Analysis Layout Analysis HTA 1 hr data collection 1 2 hrs data collection 2 4 hrs data collection
39. isplay screen There were also instances where the two methods did not agree and this has also been found in previous studies e g 55 For example glare on the touch screen would lead to a R1 SHERPA error information not obtained however the SHERPA method which is based on the HTA specification does not support the analyst in accounting for environmental factors and therefore this was not identified as a potential error The issue of false positive error detection has also been found in studies of SHERPA 31 55 this could encourage unnecessary changes to a design Heuristic analysis identifies usability issues and the assumption is that these will lead to poor usability when the IVIS is used by consumers however identification of usability issues is not a guarantee of poor performance 19 This problem is compounded by the lack of information about frequency of occurrence of issues in this type of analysis Layout analysis was only applied to the two worst performing menu screens in both IVIS therefore it is very difficult to make quantitative comparisons between the touch screen and remote controller based on this information alone The subjectivity of techniques like Layout Analysis and also Heuristic Analysis and SHERPA is a disadvantage in situations where quantifiable metrics are needed so that two or more competing systems can be compared These techniques also suffer from problems associated with the assumption that the analyst always h
40. l audio feedback to indicate incorrect entries Large text size Sensible use of abbreviations Screen is recessed protected from glare and located in a natural viewing position Hard button for home menu Easy non visual location of hard controls 29 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 3 4 2 Utility of the method The Heuristic Analysis generated qualitative data relating to positive and negative features of each IVIS according to the checklist 35 There are a number of checklists and guidelines for IVIS design for example see 35 46 47 49 however no single set of criteria has been accepted as the industry standard This reflects the difficulty in defining a set of heuristics which is capable of providing a comprehensive checklist for VIS usability One of the main problems with the method was the lack of information regarding the frequency with which particular usability problems would occur in everyday usage A further limitation of the heuristic method is the requirement for a fully developed product or prototype in order to evaluate some aspects of usability This includes the effect of glare on the IVIS display screen which cannot be assessed without exposing the IVIS to particular environmental conditions This is a constraint imposed by the design of many existing checklists for VIS evaluation of w
41. lighted highlighted 2007 900 Read p age o oe pew 5000 average to Read navigation warning 5000 e g navigation warning performing task Read number e g temperature 1000 1200 Wierwille 1993 Check temperature display on centre console display 1000 Moye hond one SIRE me ies Moturant et al Move hand to touch screen or to touch screen remote touch 900 3 1980 remote controller 900 controller and vice versa Press button target 200 Sat Renee Press touch screen target 200 400 Card et al 1983 Ackerman and 505 583 Cianciolo 1999 Homing on target movement Move hand betweentaroets time during visual search 8 520 Stanton and Baber assumed extra 320 total 520 Manual 2008 with touch target time Rogers et al 368 512 2005 1290 includes pressing enter re CEE Aree 1290 Card et al 1978 therefore positioning time 1290 570 720 Press enter button on remote controller 570 Press hard enter button 570 Card et al 1983 Press button on centre console 570 Listen for feedback to confirm Average from Listen for radio station 3000 ake kani correct radio station performing task confirmation 3000 Listen for change in audio 3000 Average from Listen for audio settings settings e g bass performing task confirmation 3000 Cognitive Make simple selection 990 Prania ane Bebe Make selection 990 2008 Where a range of values has been reported an operation time was estimated from within that range bas
42. lity 5 3 6 1 Analytic Methods for IVIS evaluation HTA produced a hierarchical outline of tasks which described the smallest operations which a user performs when interacting with a particular interface This analysis showed that the basic task segments for selecting a menu target consisted of the same number of operations for both systems and highlighted the effect of task structure on interaction strategies with the two IVIS i e the touch screen generally required more target presses to complete tasks than the remote controller Operations identified by the HTA were then fed in to the MCPA and assigned duration times in order to calculate predictions of total task times Like the HTA MCPA also highlighted the differences in task structure between the two IVIS however the MCPA also showed that although the number of operations in a task segment was consistent the operation timings assigned to these operations produced differences between the task times of the touch screen and remote controller Although there was some overlap between the output of these two methods both are recommended in IVIS evaluation HTA is a necessary precursor for other methods including MCPA and SHERPA and is thought to be a useful exercise for familiarising designers and evaluators with task structures MCPA expands the output of HTA by assigning predicted times to the tasks and the task time metric is useful for comparing IVIS and for making estimates about the effect
43. long with the differences between the two devices for each task 17 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Table 5 Total task times for secondary tasks performed via the touch screen and remote controller IVIS Task time ms Difference Task Touch Remote Remote controller screen controller Touch screen k Play radio station 909AM 8460 10770 2310 27 30 Increase bass by 2 steps 11380 11100 280 2 46 Increase temperature by 1 degree 4860 4860 0 0 00 Reduce fan speed by 2 steps 5340 6650 1310 24 53 Direct air to face and feet 8880 6080 2800 31 53 Direct air to face only 7060 6080 980 13 88 Turn on auto climate 3090 3090 0 0 00 Enter destination from system memory 16820 11260 5560 33 06 Enter destination from previous entries 16820 11260 5560 33 06 Total task time 82710 71150 11560 13 98 The MCPA method predicted that five tasks would take longer with the touch screen than the remote controller and that two tasks would take longer with the remote controller than the touch screen There was no difference between the two systems for the increase temperature and auto climate task times this was because they were performed via centre console controls rather than the IVIS interfaces and the task design was identical in both cases The two air direction tasks were predicted to be shorter with the remote cont
44. ment with little demand for resources however currently these methods are not widely used in the automotive industry for VIS evaluation This study therefore attempts to explore the utility of analytic methods including advantages and disadvantages identify training and application times and address shortcomings by proposing the extension to one or more of the techniques to increase their utility in a driving context The findings will be useful to interface designers and evaluators working within the automotive industry but also in other domains to support the selection and application of analytic methods with the overall objective of encouraging early stage evaluation and design for usability The study evaluated two IVIS interfaces a touch screen which is one of the most commonly used interface types and a remote controller which works like a joystick to control a cursor on screen and was recently introduced to the market It is important for automotive manufacturers to evaluate the performance of a new IVIS interface technology like the remote controller against their current system as a benchmarking activity The results of this comparison are reported in the case study however the main aim was to explore the intrinsic attributes of analytic methods in the context of IVIS interface evaluation 19 rather than as a direct comparison of systems 1 3 Analytic Methods Analytic methods were selected to model the different aspects of IVIS
45. n 25 This classification was based on the analyst s judgement Within each error type there are a number of error modes which are shown in table 6 22 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Table 6 Error modes and their descriptions Error mode Error description Action Al Operation too long short A2 Operation mistimed A3 Operation in wrong direction A4 Operation too much little AS Misalign A6 Right operation on wrong object A7 Wrong operation on right object A8 Operation omitted A9 Operation incomplete A10 Wrong operation on wrong object Information retrieval RI Information not obtained R2 Wrong information obtained R3 Information retrieval incomplete Checking Cl Check omitted C2 Check incomplete C3 Right check on wrong object C4 Wrong check on right object CS Check mistimed C6 Wrong check on wrong object Information communication Il Information not communicated p Wrong information communicated Information communication B incomplete Selection S1 Selection omitted S2 Wrong selection made Credible error modes were determined for each task step in the HTA along with the form that the error would take the consequences of the error and the recovery potential Next the analyst estimated the probability P of the error occurring during the task and also the criticality C
46. n IVIS the packaging of the product compliance with traffic regulations system maintenance and information referring to the road network 2 2 Data Analysis The data collected on each IVIS were modelled using the five analytical evaluation methods described previously During the modelling phase close attention was paid to the utility of each method and to the training times execution times and resources required Each of the methods and their application in this particular context of use is described further in the following sections 3 Results and discussion 3 1 Hierarchical Task Analysis HTA HTA was conducted for the two IVIS under investigation and an example of a HTA for the remote controller VIS task play radio station is presented in figure 3 10 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 1 Play radio station Plan 1 Do 1 4 in order WHEN radio station is set THEN 5 1 1 Move hand from steering wheel to controller 1 2 Open audio menu Plan 1 2 Do 1 2 in order 1 2 1 Move pointer to AUDIO symbol 1 2 2 Press ENTER button 1 3 Select source from FM AM or DAB Plan 1 3 For FM THEN 1 for AM THEN 2 for DAB THEN 3 1 3 1 Open FM menu Plan 1 3 1 Do 1 2 in order 1 3 1 1 Move pointer to FM tab 1 3 1 2 Press enter button 1 3 2 Open AM menu Plan 1 3 2 Do 1 2 in order
47. n Vehicle Information System IVIS development cycle The motivation for the work was to understand how to deliver an approach to modelling aspects of IVIS usability working with inevitable commercial constraints to provide useful information on which to base design decisions VIS have been used by automotive manufacturers particularly in the premium vehicle sector for the last decade These systems integrate many secondary This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset vehicle systems including entertainment comfort communications and navigation into a single screen based interface VIS are most commonly controlled by the driver via a touch screen remote controller hard buttons or a combination of these devices Recently the issue of IVIS usability has received growing attention 1 3 This is commensurate with the increase in functionality of these systems 4 6 which has been accompanied by the introduction of new approaches to facilitate the user system interaction 7 8 The issue of usability of such interfaces is now more significant than ever 1 1 Defining Usability An essential starting point for assessing usability is to define criteria against which to evaluate Since terms such as ease of use 9 user friendliness 10 and user perceived quality 11 were first introduced there have been many definitions o
48. of IVIS tasks on concurrent tasks such as driving SHERPA highlighted a number of potential errors with both systems which would be useful to a designer at the early stages of product development the remedial strategies devised as part of the analysis would guide any necessary redesign activities in order to reduce errors in the driver IVIS interaction There was however quite significant overlap between the errors 36 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset which were identified for the two systems which does not support the use of SHERPA as a comparative evaluation tool SHERPA is based on an objective task description and the analysis follows a rigid structure that produces quantifiable results however the assignment of error frequency and severity is dependent on the analyst s judgement The remedial strategies recommended as part of SHERPA are also an example of qualitative output Comparison of heuristic analysis with the results of SHERPA in a process of data triangulation 19 36 53 54 showed that both methods identified some of the same usability issues however SHERPA errors tended to relate to individual operations and issues which may prevent these being performed successfully whereas the heuristic analysis identified more general issues relating to the system and wider environment e g glare and reflections obscuring the d
49. of the error using an ordinal scale Low L Medium M High H Finally the analyst proposed strategies to reduce the 23 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset identified errors An extract of a SHERPA output table for the touch screen IVIS task play radio station is presented in table 7 Table 7 Extract of SHERPA output table for touch screen IVIS task play radio station 9 S gt S 5i Task fa a 5 S 3 Z mo a 3 S T os 5 3 amp m EE mm 1 Play radio N A N A N A N A N A N A N A 1 1 Open AUDIO TV N A N A N A N A N A N A N A menu Reduce need for removing hands Immediate Driver cannot remove Cannot perform from wheel x when hand from wheel due any interaction 3 increase number of A8 primary M M to high primary task with touch steering wheel demand demand screen allows controls increase 1 1 1 Move automation of hand to secondary tasks touch Reduce need for screen removing hands Driver starts to move Immediate Cannot perform from wheel hand towards screen j when any interaction increase number of A9 but has to replace on primary M H with touch steering wheel wheel due to sudden demand 3 screen controls increase primary task demand allows automation of secondary tasks 1 1 2 Pre
50. ost print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset Table 12 Quantitative and qualitative comparisons between the two IVIS Method Touch screen Remote controller Best performance Quant 125 total operations 113 total operations HTA Qual Most touch screen and remote controller tasks have similar structures but nature of individual operations is different Remote controller Quant 82710 msec total task time 71150 msec total task time MCPA Qual Remote controller times are dependent on the speed of movement through menu options and number of options scrolled through before reaching target When examined at a task segment level the touch screen is predicted to produce shorter task times Remote controller Quant 6 significant errors 7 significant errors SHERPA Qual Remedial measures include increasing the sensitivity and allowing better differentiation between targets for the touch screen increasing precision of the pointer and moving the enter button for the remote controller Touch screen Quant 7 ive 6 ive issues 8 ive 6 ive issues Heuristic Analysis Qual Usability issues include glare on the screen and lack of tactile feedback for the touch screen poor location of back button and complexity of menus for the remote controller Touch screen Quant 11 layout
51. pare to N A N A N A N A N A N A N A open menu 1 1 1 1 incoifeci meni Ensure labels Make S2 Wrong selection made Immediate L M clearly relate to opened selection function Pe Make icons and Visual check is not when Cannot open S labels larger to R1 long enough to locate primary M L A 1 1 1 2 desired menu ensure quick icon demand Locate allows identification AUDIO TV Wrong menu is me Incorrect icon is opened if Ensure icons clearl R2 Aer Immediate L M ae located by mistake mistake is not relate to function realised A4 System does not Audio TV does Immediate H L Increase sensitivity 1 1 3 Touch recognise touch not open of touch screen AUTEN Touch incorrect button R aput Increase size of button A6 made or no Immediate M M or other part of screen buttons input made 24 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 3 3 1 IVIS evaluation SHERPA was performed for each of the nine tasks for both systems Tables 8a and 8b present all identified errors and error modes for the touch screen and remote controller respectively The tables also include the probability P and criticality C ratings for each error shown in bold Significant errors were defined as those with either high P or C ratings or where both P and C were rated as medium The number of error descriptions i e system does not recognize
52. rasting products It is possible to compare two or more different products or individual tasks based on the number of operations identified by HTA A system which requires the user to perform a large number of operations per task is likely to be less efficient than a system with fewer operations however this will also depend on the time taken to perform each operation and the error potential of the tasks involved 3 1 2 Utility of the method HTA is a fairly time consuming method to carry out as each individual operation in a task needs to be analysed however creating a comprehensive HTA can considerably reduce the time required for other modelling methods such as MCPA and SHERPA A problem facing Human Computer Interaction HCI is that interfaces are often engineering focussed and are therefore not optimised for activity patterns 36 HTA provides an activity based classification of user behaviour which in itself can be used to improve interface design The process of conducting HTA can also provide the analyst with important information about task structure and menu design A deeper understanding of the links between task design and usability should increase focus on good HMI design we therefore also recommend that designers and Human Factors specialists within manufacturing companies use the process as a learning tool 3 2 Multimodal Critical Path Analysis MCPA HTA is used as a starting point for MCPA In this study operations identified
53. riving unlike other functions such as vehicle or display settings Based on information from automotive manufacturers and the analyst s personal experience it was expected that the tasks would all be used at least once during a typical medium long journey All of the tasks were available in existing IVIS including the two systems under investigation Finally preliminary investigations conducted by the analyst indicated that it should be possible to complete each of the nine tasks in less than 15 seconds The 15 second rule which is commonly referred to in the design and evaluation of IVIS tasks 29 33 34 states that no navigation tasks involving a visual display and manual controls and available during driving should exceed 15 seconds in duration 29 Tasks were performed using each system and the inputs from user to system and outputs from system to user were recorded Pictures were taken of the VIS menu screens and controls at each stage of the interaction and the analysts recorded a description of each interaction For the Heuristic Analysis each IVIS was assessed against a checklist developed by Stevens et al 35 The This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset checklist was adapted for this evaluation by removing sections which were not directly connected to usability including those relating to the documentation supplied with a
54. roller than the touch screen In the remote controller system the user is allowed to select the exact options directly because there are separate options for air to face and feet and face only however the touch screen presents three options face feet and windscreen and the user therefore needs to select multiple options to set air direction to face and feet this involves extra operations to complete the task The destination entry tasks were also predicted to take longer with the touch screen compared to the remote controller This is because the touch screen system required users to read a warning about using the navigation function whilst driving and this contributed a large amount of time to the task 5000ms to read the warning 1300ms to locate the Agree button 320ms homing time to target 200ms to touch target 6820ms total extra time Without this extra task segment the touch screen would have produced shorter task times than the remote controller for the two navigation tasks Similarly the 18 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset time difference in the increase bass task can be attributed to an extra task segment in the touch screen task with this system the user had to select the Audio TV button then Settings followed by Sound in order to access the Bass targ
55. s The Heuristic Analysis was applied by the analyst using an adapted IVIS checklist originally developed by Stevens et al 35 The checklist was organised into nine sections covering integration of the system into the vehicle input controls auditory properties visual properties of the display screen visual information presentation information comprehension menu facilities temporal information and safety related aspects of information presentation The evaluation was based on the analyst s experience gained from four five hours of interaction with each system in a stationary vehicle 3 4 1 IVIS evaluation Tables 9a and 9b list the issues identified via the Heuristic Analysis for the touch screen and remote controller IVIS The issues were categorised by the evaluator as positive or negative and further 28 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset categorised according to the estimated severity of each issue Negative major and minor and positive major and minor issues were identified for both IVIS using the heuristic checklist There were slightly more negative issues identified for the remote controller 8 compared to the touch screen 7 however because the difference was so small and the analysis was purely subjective it is not possible to use these values to make a valid quantitative comparison between the two
56. sis A task description method used to break down tasks into their smallest components and structure them in a hierarchy of goals sub goals and plans 23 24 Although HTA normally needs to be combined with other techniques to produce meaningful analysis 25 it can illustrate where tasks might lead to ineffective interactions due to poor structure Used to model task times based on the interactions between operations performed in different modes 26 28 MCPA was selected over other time prediction methods as it enables operations to be modelled in parallel Task times produce high correlations with eyes off road time 29 30 which is a measure of the interference of secondary tasks in the dual task environment Predicts error rates and types for particular systems and tasks 31 32 Errors will be useful in assessing the level of training which is needed for successful use of a product or system The nature of the dual task driving environment will also give rise to specific errors such as failing to complete an operation due to a sudden increase in primary task demand Uses a checklist of principles as a guide for identifying usability issues with an interface 14 The content of the analysis is set according to the criteria of interest dual task environment environmental conditions range of users and training provision Because it is a subjective technique it is less easy to predict factors such as uptake which n
57. sitions in the MCPA diagram and by the arrows connecting each operation 28 In figure 4 the operations locate audio TV icon visual move hand to touch screen manual and make selection cognitive are performed in parallel and are therefore presented one above the other The operation homing on target manual is dependent on the user having already decided on the correct button located it on screen and moved their hand to the touch screen it is therefore presented as a sequential operation in the diagram The operation touch audio TV button manual is dependent on the user having homed in on the target with their finger therefore this is presented to the right of the preceding homing operation MCPA diagrams were generated in this way until all operations in each task had been represented 13 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 2 O 900 900 1300 320 1620 1620 200 1820 1000 2820 3120 320 3440 gt Moyo band to EE Homing on target Toue AU DIOLEY f Move hand gt Homing on target gt touch screen 7 button 4 2 400 400 1300 i O 1620 1620 0 1820 300 j 3120 0 3440 gt i z f A i 2 i lt H S o 990 990 990
58. ues for error prediction supporting novice users in the healthcare sector Applied Ergonomics 2009 40 3 pp 379 395 33 Nowakowski C and Green P Prediction of menu selection times parked and while driving using the SAE J2365 method University of Michigan Transportation Research Institute UMTRI 2001 34 SAE J2365 proposed draft Calculation of the time to complete in vehicle navigation and route guidance tasks 2002 35 Stevens A Board A Allen P and Quimby A A safety checklist for the assessment of in vehicle information systems a user s manual Transport Research Laboratory 1999 36 Wilson C E Triangulation the explicit use of multiple methods measures and approaches for determining core issues in product development Interactions 2006 pp 46 63 37 Stanton N A and Baber C Modelling of human alarm handling response times a case study of the Ladbroke Grove rail accident in the UK Ergonomics 2008 51 4 pp 423 440 38 Olson J S and Moran T P Mapping the method muddle guidance in using methods for user interface design in Rudisill M Lewis C Polson P G and MacKay T D ed Human computer interface designs success stories emerging methods and real world context Kaufmann 1996 edn pp 269 300 39 Wickens C D Multiple resources and performance prediction Theor Issues Ergonomics Sci 2002 3 2 pp 159 177 40 Baber C
59. uman factors methods a practical guide for engineering and design Ashgate 2005 first edn 25 Stanton N A Hierarchical task analysis developments applications and extensions Applied Ergonomics 2006 37 1 pp 55 79 26 Baber C and Mellor B Using critical path analysis to model multimodal human computer interaction Int J Hum Comput St 2001 54 4 pp 613 636 27 Wickens C D Processing resources and attention in Damos D L ed Multiple task performance Taylor and Francis 1991 Ist edn pp 3 34 41 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset 28 Harrison A A survival guide to critical path analysis Butterworth Heinemann 1997 1st edn 29 Green P The 15 second rule for driver information systems Proc ITS America Ninth Annual Meeting Washington DC 9th April 1999 pp 1 9 30 Nowakowski C Utsui Y and Green P Navigation system destination entry the effects of driver workload and input devices and implications for SAE recommended practice University of Michigan Transportation Research Institute UMTRI 2000 31 Baber C and Stanton N A Human error identification techniques applied to public technology predictions compared with observed use Applied Ergonomics 1996 27 2 pp 119 131 32 Lyons M Towards a framework to select techniq
60. ze so only a small homing movement is needed after the target is found 41 This movement time varies with the visual search time It is assumed that the movement starts just after visual search begins therefore a value of 1000ms has been assigned in the models Duration modality and dependency information is used to calculate Early Start Time EST and Early Finish Time EFT as part of the forward pass through the network Late Start Time LST and Late Finish Time LFT as part of the backward pass through the network and finally Float Time according to the following rules The forward pass calculates the EST and EFT of each operation moving progressively through the task diagram from left to right 28 The EST of operation X is determined by the EST of the preceding operation plus its duration If there is more than one preceding operation which links into operation X then the EST is determined by the latest EFT of the preceding activities 16 This is the post print after peer review version of the paper which appeared in IET Intelligent Transport Systems 6 3 243 258 not typeset EST of operation X EST of preceding operation Duration of preceding operation The EFT is the EST of an operation plus its duration time EFT of operation X EST of operation X Duration of operation X The backward pass calculates the LST and LFT of each operation starting from the End no

Download Pdf Manuals

Related Search

Related Contents

Kayzero for Windows Manual,first draft  MANUEL D`UTILISATION - Green Light Cycle Ltd.Urban Ryder  Guía del usuario  Quick Start Guide for Airborne Package with Inventek  

Copyright © All rights reserved. Failed to retrieve file