Home
Sample Headline 1
Contents
1. 41 fW i Time a a2 53 fM Hardness was Language saved successfully Options Options i F z User touches the appliance The menu shows up User selects User selects the desired Setting is changed on with the mobile phone automatically Water Hardness value and saves the the appliance User selects Settings changes with OK Fig 4 Example of task solving using the AID The water hardness can be changed more easily using the hierarchical menu of the AID typically without consulting any manual The actual task order can be found in Table 1 for each device the control problem solving and everyday tasks were performed Finally the column repeated control was tested from top to bottom 7 For practical reasons we did not move back to the lounge area again to test the dishwasher a second time 11 C AUTO ID LABS 180 Interaction method Traditional 150 E AID 120 Mean time s Error bars 2 SE Control Problem solving Everyday Repeated control Fig 5 Mean time of task completion While the use of the AID device significantly cut down the execution time for exceptional control and problem solving tasks it was two to four times slower for everyday tasks We collected both qualitative and quantitative results Qualitative results used both the explicit answers from a post test questionnaire administered after all tasks were completed reported in Section 3 2 as well as notes
2. F Blackwell A F The fuzzy felt ethnography understanding the programming patterns of domestic appliances Personal Ubiquitous Computing 8 3 4 2004 161 176 Nichols J Myers B A Studying the Use of Handhelds to Control Smart Appliances In ICDCSW 03 Proc of the 23rd Int Conf on Distributed Computing Systems IEEE Computer Society 2003 274 279 Norman D The Design of Everyday Things 2nd edn Basic Books 2002 Lewis C Rieman J Task Centered User Interface Design A Practical Introduction University of Colorado 1993 Roduner C The Mobile Phone as Universal Interaction Device Are There Limits In MobileHCI 06 Workshop on Mobile Interaction with the Real World 2006 20
3. Interface Provision The mobile device s programmable screen and keyboard can be used to extend or even personalize the appliance s user interface especially in situations where space on the appliance is limited The use of handheld devices for controlling the environment has already a long tradition based on the usually infrared based remote controls that provide access from afar to audio video equipment such as TVs and stereos but also lights shades garage doors or even cars Given the many remotely controllable appliances found in today s households however it is becoming less and less practical to have a separate remote control for each of them Also users increasingly need to carry out tasks that involve more than a single appliance e g switching on the DVD player while also turning on the TV at the same time Last but not least many of today s remote controls are overloaded with functionality that users hardly ever need resulting in large unwieldy devices and hard to use interfaces A number of research projects e g 12 14 as well as commercial products e g Philips Pronto have grown out of these needs They often use a PDA to dynamically download a user interface from the appliance Nichols et al and Zimmermann et al developed these ideas further and proposed to use a PDA as a personal universal controller 15 or universal remote console 16 respectively which in turn led to the standardization of the universal r
4. allow even printed paper to send information to a mobile device e Computational Resources Mobile phones and PDAs have become powerful computing devices often featuring processors with hundreds of megahertz and considerable RAM and Flash memory Given their energy demands for sustained wide area communication provision their powerful batteries can often easily support substantial calculations and short range communications without significantly affecting the primary usage e g telephone or organizer of the device Users are also accustomed to recharging their devices periodically thus providing a virtually unlimited energy supply for locally installed applications e Programmable Screen and Keyboard Many devices already feature large color displays and programmable soft keys 5 way navigation buttons click wheels or even touchscreens allowing system designers to quickly build a wide range of attractive and versatile interfaces Built in microphones and speakers together with dedicated MP3 processors can additionally support the use of speech commands and audio cues Together these four aspects can provide two important novel provisions to the control of appliances namely e Information Provision The mobile device can provide additional information regarding the state of the appliance either by providing a display to an otherwise displayless appliance or by extending an already existing but smaller embedded display e User
5. and overloading of some buttons with multiple functions of the appliances under test With respect to the AT amp T phone they used for some of their experiments Nichols et al mention that this phone has several buttons that can be pressed and held to activate another function There is no text on the telephone to indicate this Nichols et al specifically addressed these drawbacks of the physical interface in the handheld implementation by providing an intuitive virtual interface which explains the superiority of the handheld approach In our experiments the physical user interface on the four appliances is very intuitive and convenient for all four different everyday tasks In the paper by Nichols et al there is also little information on the actual tasks that were performed The result is that we cannot compare the one or two button tasks mentioned by Nichols et al with the everyday one button tasks we studied In our opinion the value of the mobile phone to complete exceptional tasks stems from the shortcomings of the physical user interfaces and the corresponding manuals User interface designers commonly have to deal with the conflicting constraints of cost size and usability On the one hand expensive high resolution displays with many buttons help to build a user interface that is convenient and intuitive even for uncommon tasks On the other hand the cost and size restrictions typically limit the number of buttons available and the
6. designed user interface of the printer used in our study Several users made the informal comment that they liked the layout of the printer s control panel and that it was relatively easy to navigate in its menu because buttons were labeled in a helpful and familiar way Apart from the repeated control tasks we found the results of the problem solving tasks particularly interesting While in one task the appliance supplied context i e an error code that the AID used to automatically show the relevant instructions it did not do so in the two other repeated control tasks We merely gave participants some unspecific information about a problem that they had to find manually using either the printed user guide or the AID Even though we made sure the AID covered all the topics we had found in the printed documentation participants were significantly faster when they used the AID printer Mya 253s Map 25s dishwasher Mya 91s Map 29s This is surprising especially for the dishwasher as its user manual is very compact and as we find well made However the mobile phone s menu hierarchies that allow for the structuring of content seem to prove beneficial for this task Table 2 Mean time of task completion While the use of the AID device significantly cut down the execution time for exceptional control and problem solving tasks it was two to four times slower for everyday tasks Control Problem solving Ev
7. office All devices where ready to use and equipped with the appropriate handbook in German language For each of the four appliances we handed the test subjects small cards with their assignment printed on They were asked to work through each task twice One time users should use the traditional method to solve the task i e they should interact directly with the device using the physical interface The other time they should use the AID device Users were explicitly told that they could but would not have to use the user manual when completing a task in the traditional way In order to minimize potential learning effects we used counterbalancing i e participants were divided into two subgroups that worked through the cards in different orders Group A was asked to complete every task first with the traditional method and then again with help of our AID device Group B was asked to begin 10 AUTO ID LABS ie Em EN with the AID device and then use the traditional interface afterwards To get comparable results the order of the tasks was the same for all participants Learning effects were compensated for by letting the users perform the first task of each appliance again at the end of the study We measured the time needed for each task and then asked users whether it was easy or difficult for them to solve the task as well as which method they liked better For tasks that required users to find a solution to a problem an
8. size and resolution of the displays For most appliances the result is that interface designers are forced to realize uncommon tasks by a combination of buttons The corresponding instructions are then listed in a complementary manual This requires a significant effort from the user however The manual might not even be close by there is no direct link to the corresponding section in the manual and the limited expressiveness of a manual makes the association of the printed instructions with the actual physical interface of the appliance non trivial The AID addresses all of these shortcomings since it represents a cost effective way to equip any appliance with a high resolution display and a multitude of buttons It thus enables user interface designers to address the conflicting constraints of usability cost and size associated with user interface design by leveraging the external mobile phone with its significant display and input capabilities The AID provides not only a high resolution display and input capabilities it also permits the personalization of the user interface due to the personal nature of the mobile phone While designers usually have to develop an interface for the average user the AID concept allows them to build a software interface that automatically adapts the language adjusts the features available according to the capabilities of the user and lists the history of recently performed tasks The user interface designer wo
9. was unavailable 14 AUTO ID LABS Agreement liked using the AID The AID made me feel If had an AID for my Better user manuals more confident when appliances I d use would render the AID used the appliances functions that find too unnecessary complicated today Fig 7 Subjective Likert scale responses Participants rated the usefulness of the AID device after they finished all tasks 0 strongly disagree 4 strongly agree Error bars 1 SD We ended our questionnaire with the following question Could you imagine accessing all functions offered by an appliance only through the AID Why why not While some participants could imagine giving up the traditional interface 4 users i e 17 most replied that a software only user interface was not an option for them 17 participants i e 74 2 participants were unsure Most users answered that they would rather not use the AID for simple tasks However two respondents stated that they might be willing to accept an appliance without a traditional user interface if its price were lowered in turn Two participants added that they could think of appliances that would benefit from a full replacement of the user interface as this would make ugly control panels redundant and improve visual appearance 3 3 User feedback In the course of the study there were a number of issues that were brought up by the participants In this section we re
10. 001 Fig 34 823 everyday tasks p lt 001 F 4a 98 302 repeated control tasks p lt 001 F41 124 44 851 12 AUTO ID LABS i We therefore find our hypotheses confirmed that mobile phone based interaction significantly reduces completion times for control and problem solving tasks while significantly slowing down everyday tasks Interestingly controlling an appliance with the mobile hone was significantly faster even after participants had familiarized themselves with the task and could be considered experienced users of the respective appliance We also studied the effects of other factors namely age gender and experience with advanced phone features We did not find an interaction effect with interaction method for any of these factors We therefore conclude that for the 23 participants in our study the use of the AID was beneficial irrespective of their gender age or previous mobile phone experience An individual analysis of our 18 tasks showed a consistent pattern For each of the everyday tasks traditional interaction proved faster while users were faster using the AID in all other cases However there was a single exception from this picture namely the repeated control tasks performed on the printer While interaction with the AID took slightly less time Map 16 9s Mya 21 68 this difference fell short of significance at the 05 level p 177 F4 40 1 885 We presume that this is due to the well
11. AUTO ID LABS Christof Roduner Marc Langheinrich Christian Floerkemeier Beat Schwarzentrub Auto ID Labs White Paper WP BIZAPP 039 Christof Roduner Marc Langheinrich Institute for Pervasive Computing ETH Zurich Institute for Pervasive Computing ETH Zurich Auto ID Lab Switzerland Christian Floerkemeier Beat Schwarzentrub Auto ID Lab Massachusetts Institute of Institute for Pervasive Computing ETH Zurich Technology Contact Institute for Pervasive Computing ETH Zurich Clausiusstr 59 8092 Zurich Switzerland E Mail roduner inf ethz ch Internet www vs inf ethz ch Q Springer Verlag Berlin Heidelberg 2007 Reprinted from LaMarca A Langheinrich M Truong K N Eds Pervasive Computing 5th International Conference PERVASIVE 2007 Toronto Canada May 2007 Lecture Notes in Computer Science vol 4480 pp 198 215 Available from www springerlink com content xn30k60395152055 With kind permission of Springer Science and Business Media Business Processes amp Applications AUTO ID LABS 1 IRPOCDUCTION 20 e cece cccceeccecceccceccecceeeceeceecaueceeeaeeceecueeaeeceecaeeseeceeeaeeseeciesneeeeeeaeeaeens CEA dogs ccc 2 1 Hypotheses 2 2 Appliances and Tasks 2 3 Participants 2 4 Apparatus 2 5 Procedure SD S MR 3 1 Quantitative results 3 2 Qualitative results 3 3 User feedback d DISCUSSION anneanne Cond eon reed Conc FU ov PASEO eva CU P d dens 5 Conclusion and Outlook oct occ
12. a x Dia cp Va a a patena a ua avt References RINT EET RERO REC DOC RC CORPER AUTO ID LABS Mobile phones are increasingly becoming ubiquitous computational devices that are almost always available individually adaptable and nearly universally connectable using both wide area and short range communication capabilities Until Star Trek like speech interfaces are fully developed mobile phones seem thus poised to become our main devices for interacting with intelligent spaces and smart appliances such as buying train passes operating vending machines or controlling smart homes e g TVs stereos and dishwashers as well as heating and light But how much can a mobile phone simplify our everyday interactions before it itself becomes a usability burden What are the capabilities and limitations of using mobile phones to control smart appliances i e operating things like ATMs or coffee makers that typically do not benefit from remote control This paper presents a user study investigating the use of a prototypical mobile phone based interaction system to operate a range of appliances in a number of different task settings Our results show that mobile devices can greatly simplify appliance operation in exceptional situations but that the idea of a universal interaction device is less suited for general everyday appliance control Personal mobile devices such as mobile phones and PDAs represent an important building block in many ubicomp
13. ance s traditional user interface would be faster than mobile phone based interaction We used four typical appliances for which we found a number of use cases where mobile phone based interaction might be beneficial a dishwasher V ZUG Adora S 55 a coffee maker Jura Impressa S70 a printer HP LaserJet 4050tn and a radio Sangean ATS 505 For each appliance we defined a number of tasks for participants to work through once using the appliance s native controls once using our AID device We grouped these tasks 18 in total listed in Table 1 into the following four categories AUTO ID LABS i I L OL LPPALAAA DRE Em e Control tasks involve the adjustment of a special device setting e g setting the water hardness for the coffee maker or the invocation of an unusual operation e g create a printer cleaning page Theses tasks reflect the use of a mobile phone for user interface provision e Problem solving tasks confront users with an abnormal situation e g a malfunctioning dishwasher displaying an error code that must be dealt with These tasks correspond to the use of a mobile phone for information provision e Everyday tasks are control tasks that are most typical for an appliance e g brewing a coffee and are performed very often e Repeated control tasks are control tasks that a user has performed very recently and is still very familiar with Using our prototype presented in 25 a small pilot study was r
14. and recordings taken during the tasks that captured the participants thoughts through a thinkaloud technique which they were instructed to employ see Section 3 3 The quantitative results simply measured the time between the begin and end of a task though for tasks involving an answer rather than asking the user to operate an appliance i e the problem solving tasks see Table 1 on page 5 the time until the user started to reply was measured This data is reported in the following section Figure 5 and Table 2 show the average task completion times we measured in each condition As predicted the mean time of task completion decreased for control tasks and problem solving tasks whereas it increased for everyday tasks We further examined the collected data using analysis of variance ANOVA We ran a two way ANOVA with factors interaction method and task type which confirmed a significant main effect of interaction method p lt 001 F 73s 00 23 Also a significant main effect of task type p lt 001 Fa73s 661 18 and a significant interaction effect between task type and interaction method p 001 Fa73 amp 218 68 was found Focusing on this interaction effect we continued by analyzing the effects of interaction method for each of the four task types For each task type an ANOVA showed that there was a significant difference between interaction method control tasks p lt 001 Fi 36 600 1 problem solving tasks p lt
15. articipants if a well made quick reference guide could make this content obsolete on the AID Most of them agreed that this was not the case don t want to go find a manual that anyway may have thrown away Two participants also said that the AID could be improved by adding search capabilities because then can search for keywords with CTRL F just like when I read a manual in PDF format Overall while having some concerns on the ever increasing dependence on technology participants generally liked the AID and gave very positive feedback on its use In one participant s words In the beginning was sceptical but now I m very excited 3 2 5 2 1 0 5 0 Accessing Adjusting settings Explanation of Looking up arbitrary everyday functions error codes information user manual on AID Perceived value E Fig 8 Concluding questions Participants were asked to rate the perceived value of the AID device in various situations 0 no value 3 great value Error bars 1 SD 16 AUTO ID LABS At first sight our measurements seem to disagree with the results of the user study by Nichols et al 22 which showed that their participants took twice as long interacting with the actual physical interface even for those everyday tasks that required only one or two button presses on the actual appliance Nichols et al attribute this superiority of the handheld approach to the poor labelling insufficient feedback
16. ck change channel set clock store preset station See Section 1 AUTO ID LABS We tested a total of 23 participants 10 4396 of which were male and 13 5796 of which were female Most of them were undergraduate or graduate students recruited through mailing lists from different universities There were 10 participants with a background in sciences or engineering 10 participants from the humanities or social sciences and 3 participants with no academic background All of them spoke German as their native language and owned a mobile phone Except for two subjects all participants used their mobile on a daily basis The average age of participants was 29 8 years ranging from 21 to 50 SD 6 1 years None of them had any relevant previous experience with the appliances used in our experiment Participants were compensated for their time with a USB memory drive or a cinema voucher according to their preferences In order to evaluate the usefulness of a universal interaction device we had to provide our test subjects with a corresponding mobile unit that would let them properly experience the use of such a remote control unit Our AID prototype system supports mobile phone based interaction for all of the tasks outlined above Our system is implemented as a Java MIDlet that we deployed on a Nokia 6630 mobile phone This mobile phone features a color display with a resolution of 176 x 208 pixel Unlike the device
17. d telling us about it time stopped with their first utterance Figure 2 shows some of our participants carrying out tasks Figures 3 and 4 show an example of an entire task adjusting the water hardness in the coffee maker being done using the traditional method and using the AID device respectively The final part of our study asked our participants to complete seven different tasks in a row see Figure 6 on page 13 for a list which they had to perform as fast as possible They were however free to choose any method to operate the appliances i e they could pick up the AID consult the printed manual or directly operate the physical appliance interface to accomplish the assignment We then recorded their individual choices for each of the seven tasks In order to get participants to truthfully choose the methods they thought would be most effective we offered a portable MP3 player to the user with the fastest time The study ended for each participant with a final questionnaire that collected their opinion and suggestions on the AID device A single session typically lasted about 80 120 minutes v9 m v _ e 1 _ er _ i i d F T f ef Cotfee naker ater Hardness bd 5 om Coffee Powder Quantity 0 off 0 off Settings Coffee Temperature lo 1 1 13 fn 1f 3 fu Maintenance amp Care Water Quantity lo 2 14 27 fn 2 14 27 fH Help water Hardness i 3 28
18. eation of macros to facilitate repeated tasks User studies investigating the performance of personal mobile devices for the spontaneous interaction with appliances were carried out by Nichols et al 11 22 as part of the evaluation of their personal universal controller 15 They studied the use of a PocketPC to control a stereo and a telephone digital answering machine In particular they compared the performance of 12 subjects when accessing the appliance either using the PDA interfaces or the interface on the actual appliance The authors found that compared to the user interface on the PocketPC interaction based on the physical appliance interface took twice as long entailed twice as many errors and required five times more external help While these results seem very encouraging to the vision of using mobile phones as universal interaction devices in the future they might strike one as somewhat counterintuitive Why would a softkey based interface of a generic PDA be more efficient for playing back voice messages of a real answering machine than the machine s own buttons Why wouldn t the direct interaction with the physical machine help users with understanding and operating the device by making use of the machine s perceived affordances 23 Obviously using a PDA or mobile phone as an interaction device will be greatly beneficial when this presents the only way to access an otherwise invisible device e g for information ser
19. emote console framework within the INCITS V2 technical committee 17 Among the appliances that are typically considered controllable by such universal remote controllers are video recorders DVD players TVs video projectors stereos answering machines light See www pronto philips com AUTO ID LABS i Em switches home security systems ATMs elevators copy machines cars vending machines heating control systems microwaves ovens and washing machines Researchers have recently begun to look at the suitability of different possible interaction techniques such as scanning a barcode pointing with a laser beam or touching an RFID tag for such scenarios 18 19 However there is surprisingly little work addressing the question of which appliances are actually suitable for this new paradigm of interaction and under which circumstances they are so Koskela et al 20 have studied the use of mobile phones PCs and media terminals in a household over six months However the handheld device could only be used to control lights and curtains in their setting Rode et al 21 conducted an ethnographic study to find out which household appliances users choose to program Their research gives however no indication for which appliances a personal mobile device may be an appropriate interaction tool Moreover they do not consider spontaneous interaction with an appliance but focus on the programming of actions for the future and the cr
20. ere asked to change the level of water hardness in the coffee maker The above pictured steps usually required an extensive study of the printed manual The experiment began for each participant with a brief introduction to the concept of the AID our user study and its goals Participants were then asked to fill out a profile questionnaire that allowed us to gather information on their background age education previous experience with devices used in the experiment etc We explained the basic concepts of the AID i e the user interface provision and the information provision and demonstrated it interactively using example tasks that did not reoccur in the course of the study We also told participants explicitly not to think of the AID as a remote control and that they would only be able to use it on an appliance after they had touched it followed by pressing the phone s main navigation button We finally handed them the AID device and guided them through a number of simple preparation tasks in order to familiarize them with the phone s controls and user interface The beginning of the study was conducted in a semi public lounge in our university as it offered a dishwasher that we could use There were only few distractions here and for the rest of the study we moved to a quiet office where we had set up the coffee maker the laser printer and the radio The introduction and initial explanations described above were also conducted in this
21. eryday Repeated control Traditional 162 112 121 131 2 4 28 24 AID 24 21 27 16 8 12 12 6 13 AUTO ID LABS i DRE m Printer Activate EconoMode Radio Select preset station Printer Fix error 64 TAID BI Traditional Coffee maker Adjust water hardness DIBoth Printer Cancel print job ONeither Radio Store preset station Printer Clear paper jam 0 10 20 30 40 50 60 70 80 90 100 Fig 6 Interaction method usage during contest For the final contest participants were free to use any method that they felt most comfortable with for each task Out of the seven tasks four were previously completed tasks while three Activate 55 GG EconoMode Fix error 64 Clear paper jam were new tasks Figure 6 summarizes the results of the final contest in which participants had to solve a list of tasks in the shortest possible time but for which they could freely choose the interaction method Overall the AID was used in 69 of the control and problem solving tasks even when participants had previously used the traditional user interface in similar situations just as effectively as the AID Most of the participants stated that already slightly different tasks would make them insecure fearing that they would not know where to find it in the appliance s menu structure They would therefore opt for the AID as it gives me a better overview than the printer s
22. he phone s main navigation button As users performed the tasks in separate steps the experimenter had time to use an undocumented button combination on our AID to quickly select the proper context dependant reactions for the upcoming task thus giving users the illusion of having our AID actually detect and read out the appliance Obviously this setup made it impossible to really control the appliance in any way through the AID a shortcoming that we pointed out to participants indicating that we were only interested in seeing the right buttons being pressed not an actual appliance being controlled Figure 1 shows four example screenshots of the AID during the study one for each of the four appliances Invoking our AID device on an idle appliance brings up the appliance s main menu as shown in Figure 1 a For each appliance this main menu would offer all tasks that are available through the appliance s physical interface in a hierarchical menu The user interface for a typical task is shown in Figure 1 b We also included a troubleshooting menu entry for each task which would contain the contents found in the corresponding section of the appliance s user manual Figure 1 c shows such a list of common problems that might occur at the dishwasher Finally we provided several step by step assistants that help users with physical appliance manipulations The assistant supporting the task of clearing a paper jam at the laser printer is illust
23. ilities will enable appliance manufacturers to overcome the conflicting constraints on cost size and usability of the user interface by leveraging the user s mobile phone 18 1 2 3 4 5 6 7 8 9 10 11 12 13 AUTO ID LABS Kindberg T Barton J Morgan J Becker G Caswell D Debaty P Gopal G Frid M Krishnan V Morris H Schettino J Serra B Spasojevic M People Places Things Web Presence for the Real World Mob Netw Appl 7 5 2002 365 376 Griswold W G Shanahan P Brown S W Boyer R Ratto M Shapiro R B Truong T M ActiveCampus Experiments in Community Oriented Ubiquitous Computing Computer 37 10 2004 73 81 Intille S S Designing a Home of the Future IEEE Pervasive Computing 1 2 2002 76 82 Davies N Gellersen H W Beyond Prototypes Challenges in Deploying Ubiquitous Systems IEEE Pervasive Computing 1 1 2002 26 35 Carter S Churchill E Denoue L Helfman J Nelson L Digital Graffiti Public Annotation of Multimedia Content In CHI 04 CHI 04 Extended Abstracts on Human Factors in Computing Systems 2004 1207 1210 Smith M A Davenport D Hwa H Turner T Object AURAs A Mobile Retail and Product Annotation System In EC 04 Proc of the 5th ACM Conf on Electronic Commerce ACM Press 2004 240 241 Rohs M Roduner C Camera Phones with Pen Input as Annotation Devices I
24. n Pervasive 2005 Workshop on Pervasive Mobile Interaction Devices PERMID 2005 23 26 Myers B A Stiel H Gargiulo R Collaboration Using Multiple PDAs Connected to a PC In CSCW 98 Proc of the 1998 ACM Conf on Computer Supported Cooperative Work 1998 285 294 Ballagas R Rohs M Sheridan J G Borchers J BYOD Bring Your Own Device In UbiComp 2004 Workshop on Ubiquitous Display Environments 2004 Beigl M Point amp Click Interaction in Smart Environments In HUC 99 First Int Symposium on Handheld and Ubiquitous Computing Springer Verlag 1999 311 313 Myers B A Nichols J Wobbrock J O Miller R C Taking Handheld Devices to the Next Level Computer 37 12 2004 36 43 Hodes T D Katz R H Servan Schreiber E Rowe L Composable Ad Hoc Mobile Services for Universal Interaction In MobiCom 97 Proc of the 3rd Annual ACM IEEE Int Conf on Mob Comp and Networking ACM Press 1997 1 12 Ponnekanti S Lee B Fox A Hanrahan P Winograd T ICrafter A Service Framework for Ubiquitous Computing Environments In UbiComp 01 Proc of the 3rd Int Conf on Ubiquitous Computing Springer Verlag 2001 56 75 19 AUTO ID LABS 14 15 16 17 18 19 20 21 22 23 24 25 Olsen D R Jefferies S Nielsen T Moyes W Fredrickson P Cross modal interaction using XWeb In UIST 00 Proc of the 13th Annual ACM Sympo
25. o it a few months later without any problems Another participant explained how the AID s menu system allowed her to interactively explore the full range of functions always like to gain an overview of all available functions of my appliances The AID would be ideal for that While they liked our prototype implementation of an AID some participants expressed doubt if an actual appliance manufacturer would be able to come up with an equally user friendly piece of software for their own products One participant suggested that every manufacturer should include a menu item called troubleshooting just like all Windows applications have a Help menu Many users especially liked the immediate reaction of the AID to special situations e g by displaying helpful instructions instead of just a status code As one participant put it It is also extremely convenient that I m immediately told what the cause of an error is However there were also cases where context sensitive behavior confused people more than it helped as they expected the normal menu to show up Most participants expressed their frustration with user manuals As one user said Most handbooks contain too much information that don t actually want to read They therefore found it very helpful to have the essential problem or task oriented information available on the AID and were not bothered by its relatively small display size We asked some p
26. rated in Figure 1 d For each step the system highlights the part of the printer the user must operate next When the user pushes the right arrow on the phone s keypad the assistant advances to the next step Assistants and troubleshooting tips can either be accessed manually through the menu or they are displayed automatically when the appliance is in a state where such help is needed The device used in our pilot study 25 was an NFC enabled Nokia 3220 that actually performed a true wireless identification of each appliance However as both the computational resources and the screen estate of the Nokia 3220 were limited in comparison with the Nokia 6630 and because no discernable difference was noticeable between the real and simulated NFC action we decided on the setup described above 5 As pointed out previously this appliance state detection would be setup secretly by the experimenter prior to giving out a particular task to the user as no communication between the appliance and our AID device takes place AUTO ID LABS Display Display User enters the menu User selects the desired User opens the menu User presses to User presses P to User presses N to by pressing and hol menu item by pressing item to adjust the set the water save the changes leave the menu ding P for 2 seconds repeatedly value hardness to 2 Fig 3 Example of traditional task solving At one point study participants w
27. s used in other evaluations see Section 1 above the Nokia 6630 does not offer pen based input capabilities but features only a simple keypad a QE z r1 z of Dishwasher pg Device does not start White film Bad drying Rusty spots Remaining detergent Condensed water Options _v Backi A a Radio main menu b Adj water hardness c Dishwasher user manual d Printer troubleshooter Fig 1 AID prototype implementation The screenshots show examples of the AID user interface for each of the four appliances used in the study Appliance functions could either be selected by traversing the main menu a or by having the AID automatically i e simulated see Section 2 4 detect the appliance s state d Radio 1 0 1 13 fH 2 14 27 fH 3 28 41 fH 4 42 53 fH Settings Alarm clock a 3 Our pilot study included 9 participants none of which participated in our main study later AUTO ID LABS Fig 2 Participants interacting with appliances The images above show three of our participants carrying out tasks using the AID device to operate the coffee maker troubleshooting the printer using the printed manual and using the AID device to operate the radio Instead of actually coupling the AID with our four appliances we simulated both appliance identification as well as the transmission of appliance status by having the user press t
28. sium on User Interface Software and Technology ACM Press 2000 191 200 Nichols J Myers B A Higgins M Hughes J Harris T K Rosenfeld R Pignol M Generating Remote Control Interfaces for Complex Appliances In UIST 02 Proc of the 15th Annual ACM Symposium on User Interface Software and Technology ACM Press 2002 161 170 Zimmermann G Vanderheiden G Gilman A Prototype Implementations for a Universal Remote Console Specification In CHI 02 CHI 02 Extended Abstracts on Human Factors in Computing Systems ACM Press 2002 510 511 LaPlant B Trewin S Zimmermann G Vanderheiden G The Universal Remote Console A Universal Access Bus for Pervasive Computing IEEE Pervasive Computing 3 1 2004 76 80 Rukzio E Leichtenstern K Callaghan V Holleis P Schmidt A Chin J An Experimental Comparison of Physical Mobile Interaction Techniques Touching Pointing and Scanning In UbiComp 06 Proc of the 8th Int Conf on Ubiquitous Computing Springer Verlag 2006 Ailisto H Pohjanheimo L Valkkynen P Str mmer E Tuomisto T Korhonen l Bridging the physical and virtual worlds by local connectivity based physical selection Personal Ubiquitous Comput 10 6 2006 333 344 Koskela T Vaananen Vainio Mattila K Evolution towards smart home environments empirical evaluation of three user interfaces Personal Ubiquitous Comput 8 3 4 2004 234 240 Rode J A Toye E
29. systems discussed in the community 1 4 Their widespread use and their characteristics as a general purpose computing platform make them appear as ideal devices for implementing many interactive services in intelligent environments Scenarios involving personal mobile devices range from attaching digital annotations to physical objects 5 7 sharing public displays 8 9 and interacting with appliances of all sorts 10 11 Using mobile phones and PDAs to query and control smart environments and artifacts is attractive due to four main aspects of today s devices e Wireless Communication Apart from the continuously expanding wide area coverage mobile operators are also increasingly offering digital i e packet switched communication services such as GPRS EDGE or UMTS which can provide fast reliable and economic device communication from almost anywhere in the world both indoors and outdoors Moreover short range communication protocols such as infrared IR and Bluetooth allow local ad hoc communication between similar devices e Tag Detection The recent addition of Near Field Communication NFC technology not only improves intra phone communication i e simplifying the often complicated Bluetooth setup process but also allows mobile devices to detect and read out 3 AUTO ID LABS i 1 KCK G C C RbEAA p L OGAGL DAE m passive NFC compatible RFID Tags Moreover camera phones can use 2D barcodes to
30. t universal appliance controllers might be superior to traditional physical appliance interfaces in exceptional situations only but not for carrying out everyday tasks we had our 23 test subjects perform a series of 18 tasks distributed among four appliances By collecting quantitative measurements we could confirm that our users were significantly faster when having to solve exceptional tasks with our AID our prototypical universal appliance controller but slower when performing everyday tasks Our qualitative methods further confirmed that users would often prefer using the AID but still liked the natural interaction with a device if a simple straightforward task was to be solved and the tasks required them to be in the vicinity of the appliance anyway These findings seem to suggest that hybrid approaches that combine traditional haptic user interfaces with extended user interfaces on a mobile device offer the best of both worlds Users could continue to directly interact with appliances which is both faster and more convenient in most everyday situations However in special situations where users would have to remember complex and clumsy sequences of pushing buttons or manipulating the appliance it is much more intuitive to use a mobile device with its powerful and versatile user interface for interaction The results thus suggest that the proliferation of mobile phones with high resolution displays and short range communication capab
31. th analytically and anecdotically We close with a discussion and conclusions Our study tried to assess the benefits and limits of handheld devices in appliance operations by asking study participants to use a range of appliances in a variety of situations both traditionally using the appliances physical interface and with a specifically developed universal interaction device We then obtained quantitative data by measuring the time it took a test subject to complete a specific task as well as qualitative data by observing and recording the users actions thoughts through think aloud techniques 24 and opinions through a posttest questionnaire This section presents the hypotheses tasks and procedure of our study including a description of our prototypical universal interaction device the A D short for Appliance Interaction Device We began our project with a set of three hypotheses which together would either support or weaken our intuitive notion that the use of universal interaction devices has limits Specifically we hypothesized as follows e For controlling an appliance in exceptional situations interaction based on a mobile phone would be faster than interaction based on the traditional user interface e Looking up context dependent information on the handling of an appliance would be faster using a mobile phone than using traditional means e g user manuals e To carry out everyday tasks the use of an appli
32. two line display and allows me to complete the task faster as one participant explained On the other hand participants hardly used the AID to store or select a station in the radio s preset memory stating that they didn t even think of using the AID as it was just more natural to interact with the radio directly In the post test questionnaire we asked users to rate a number of statements on a five point Likert scale The statements and participant responses are shown in Figure 7 We also asked users to rate the value of the AID in some given situations The results of this question are depicted in Figure 8 Finally every participant was asked to answer three concluding questions At the outset we asked them Do you use appliances for which you would like to have the AID If so please name these appliances 22 participants 96 replied positively and listed both devices from our study as well as microwave ovens TVs DVD players car radios heating systems and appliances that you change often such as rental cars Only 1 participant 4 would not want our AID We also asked If you owned an AID for an appliance would it be ok if you could access rarely used functions only through the AID Why why not 18 78 of our users agreed 4 1796 disagreed and 1 participant was not sure Some participants expressed concern that they would no longer be able to interact with their appliances in case their mobile phone or PDA
33. uld also benefit from the availability of well supported software development platforms such as J2ME or Symbian The long range communication capabilities of the mobile phone might also facilitate software maintenance since the user interfaces can be updated remotely 17 AUTO ID LABS i LOELAAULuLL URRE E In our study the appropriate mobile phone software was preloaded on the mobile phone and we simulated the appliance identification with a mobile phone The rationale for simulating the identification was the limited display and computing capabilities of today s NFC enabled mobile phones and the fact that our pilot study indicated that there was no discernable difference between real and simulated NFC action As the NFC technology becomes available in mobile phones with high resolution displays future user studies could incorporate the appliance identification and possibly also the downloading of the software While we do not believe that this will impact the findings of our study it would make the overall application scenario even more realistic The goal of our study was to assess the benefits and limits of using mobile devices such as PDAs or mobile phones for appliance control While the idea of a universal remote control is an appealing one given the technical capabilities and prevalence of such devices we questioned the sheer limitless uses that today s designers and researchers often envision for them Hypothesizing tha
34. un in advance to identify tasks suitable for comparison For example the coffee maker s manual contained instructions on how to brew a latte macchiato which required the user to adjust various settings in no less than 10 steps Obviously brewing a latte macchiato could simply be offered on the mobile device as a single menu item As this would have drastically favored the AID device we took care to select only tasks that would require a comparable degree of interaction when executed directly on the appliance and on the mobile interaction device Similarly we omitted tasks that were so poorly supported by the appliance manufacturer that they proved very difficult and lengthy when tested in our pilot study e g changing the coffee maker s water filter As these tasks could be improved easily e g through better documentation we did not consider them for our main study Table 1 Appliance tasks in the user study Participants were asked to complete 18 tasks distributed among the four available appliances Not all appliances had suitable tasks in all four categories Control Problem solving Everyday Repeated control Dishwasher adjust water hardness Fix error F2 Start program activate child safety lock white film on dishes Coffee maker Adjuist water hardness brew coffee adj water hardness set switch on time Printer change paper type fix faded print cancel print job change paper type print cleaning page Radio Set clo
35. vices such as voicemail systems or online media libraries Similarly a universal interaction device might be the only means for users to discover the invisible information services available in a smart room or attached to a smart object And obviously as the success of the TV remote control has shown handhelds should be well suited to control appliances where interaction at a distance is desirable such as a heating control system However it is less clear whether personal mobile devices are beneficial for interacting with physical appliances that require the user s presence to operate such as ATMs elevators or microwave ovens With this in mind we set out to conduct a user study exploring the benefits and limits of using a mobile phone to operate physical appliances i e devices that would typically not benefit from being remotely controllable Our aim was to identify in particular the conditions under which devices like coffee makers printers or microwave ovens would benefit from being operated not directly but through a mobile phone or conversely when it would be a hindrance rather than an advantage to have such a separate interaction device AUTO ID LABS i n nh h ec6GCEC A A AABE E The remainder of this paper presents our user study in detail Section 2 starting with the experimental design and participants describing our apparatus and materials and outlining our procedure Section 3 contains the results of our study bo
36. view these informal comments The biggest concern that users expressed was that of increasing dependence on technology and in particular the mobile phone through the use of an AID device Most often participants wondered how they would use their appliances if their phone was misplaced lost stolen malfunctioning without network coverage or had simply run out of battery Two participants also mentioned that they did not want to always carry their phone with them just to be ready to use their appliances Other concerns were more diffuse don t fully trust the mobile phone and would like to have the buttons on the device Just in case Someone else perceived the AID as yet another burden in everyday life We ve got so many devices around us already so don t think we need another one just for rarely used functions What most participants pointed out as the AID s biggest advantage over traditional interfaces 5 d n was its menu structure It was described as easy to use well structured clearly laid out and intuitive Many participants felt that this was mainly due to the larger display size 15 AUTO ID LABS ie DD ii especially when compared to a small appliance LCD Several participants explained that they would hardly get lost in the menu thus enabling them to easily find even previously unknown functions One participant said If I had an AID could forget about how I did something and still be able to d
Download Pdf Manuals
Related Search
Related Contents
NEC Express5800/T110f-S Configuration Guide Amana ACS3380 User's Manual Selecção da tipologia de cruzamentos em função da procura Omega Speaker Systems FPU4500 User's Manual 製品保証書 取扱説明書 Copyright © All rights reserved.
Failed to retrieve file