Slides from the final project presentation for a prototype created as part of University at Buffalo's Department of Industrial and Systems Engineering IE 535 Human-Computer Interaction course taught by Sudeep Hegde.
Lead/graduate: Lauren Madar
Undergraduate members: Scott Baron, Kaitlyn Franz, Tiffany Standish
IE 535: Improving Smartphone Interactions During Vehicle Operation prototype project final presentation
1. Improving Smartphone
Interactions During
Vehicle Operation
Final Project Report
University at Buffalo - IE 435/535 Fall 2018
Graduate Student: Lauren Madar
Undergraduate
Students: Scott Baron, Kaitlyn Franz, Tiffany Standish
3. Background
Distracted driving is a large issue. Smartphones have become a large source of distraction.
Smartphones have become an important part of everyday life to the point that some people
feel they must answer or make phone calls, respond to messages, and perform other
activities on their mobile devices while driving. As of right now there is no easy way to change
phone applications without taking attention away from the road.
4. Goal
Explore…
● The types of tasks users commonly perform on their phones while driving
● How easy or difficult they find using their vehicle’s phone integration functions
● Provide options for improvement over existing vehicle-phone controls.
5. Data Gathering
Survey allowed us to gather information about:
● Users
○ Age, driving experience, etc
● Common phone-related functions performed (while driving)
● Level of difficulty of perform tasks
We examined different cars to see the inconsistency between steering wheel
controls.
8. Storyboard & Paper Prototype
● Low fidelity
● Ease of button placement/movability
● Limited resources● User selecting a music album from
phone to play through car
● Many different pathways to do this
9. Feedback for Paper Prototype
● Learnability factor
● Initial run - 53 seconds
● Second run - 24 seconds
● Why did speed increase so much?
12. Prototype Requirements & Criteria
● As a result from our survey and paper prototype, the interactive prototype was
required to have:
○ Steering wheel controls
○ Touchscreen view
○ Mobile device view
○ HUD view
● Wanted to simulate an actual car dashboard/windshield
● Contain distracted and non-distracted views
○ Goal: reduce/limit time in distracted views
13. Usability Test Plan
To assess the usability of the final interactive prototype, we examined:
● Task completion times
● Time spent in a distracted modes
● Subjective ratings from users
15. Prototype Methodology
● Considered Axure and other prototyping software
● Needed to allow online distribution to users
● Desire to collect detailed analytics made a website with database
necessary - not available in Axure/AdobeXD out of the box
● 2 Google Forms surveys embedded to capture demographic info and
subjective measures, one at app start and one after task completion
● ExtJS used for UI, PHP on Apache server for backend,
MySQL for database
● Specific interaction events stored on server
16. Reduced # of scenarios due to constraints
Planned to do 3 scenarios of type media, communication, navigation.
Only had time to complete menu/interactions for the media/music category.
Created 3 music tasks:
1. Play a specific song (Yellow Submarine, The Beatles - Album: Sgt. Pepper)
2. Change radio station to a specific station (FM 102.5)
3. Play a specific album (Brass Monkey by Beastie Boys)
17. Data Captured
● Users
○ Demographic attributes: ID, Email, IP
○ State control attributes: Current Mode (practice, task),
Current View (undistracted, phone, steering, touch)
CurrentHUDstate (off, on)
● Events - actions that users perform
○ Attributes: ID
Timestamp (date/time event happened)
Type
Value (used for some events)
18. Controls
Keyboard (simulate steering buttons without
needing to divert gaze), available in all views
including undistracted & HUD.
● left arrow (back)
● up arrow/down arrow (next/prevsibling)
● right arrow (select or get children)
● enter (ok or select)
Touch Simulation (mouse click)
● select menu item to navigate in Phone and
Touchscreen views
● trigger arrows/select in Steering view
● activate different views to simulate
looking/paying attention to different
controls
19. Controls - continued
‘Look at Steering Controls’ button fallback.
● analogous to keyboard arrows/enter
● allows users who don’t have sufficient
hand-eye coordination to explore controls
on-screen
● only available in a distracted ‘look down’
view.
20. Practice vs Task Modes
● Wanted to account for repeated uses having better performance in
our paper prototypes
● Created an option for users to get familiar with the controls before
starting the timed tasks, but did not force users to do a practice task
32. Final Prototype Outcomes
What went well:
● Easy to send to many users
● Interactivity scope could be focused
● Time needed from designers to interact
with respondents was minimal
● Able to collect variety of data events and
capture in DB
Events captured:
Total: 2021, usable: 1001
What was problematic:
● Higher fidelity took longer than anticipate
to finish
● Training/instruction for users limited to a
few paragraphs of text
● Limitations on type of devices that the
prototype could be used on caused
confusion (required laptop/desktop with
keyboard and mouse or touchpad, not
touchscreens)
● Little time for debugging resulted in data
not being recorded for first 10
respondents
● Timestamp configuration in database only
captured seconds, not milliseconds.
● Lots of data gathered, not enough
time to analyze.
34. Response for Final Prototype
● Total Responses: 23
● Removed due to data
collection issues: 10
● Responses examined: 13
35. Users: 13, Survey Responses: 11 Inconsistencies or
confusion about need to
complete/submit
embedded
Google Forms
36. Age Ranges & Use of Video Games/Controllers
Sam
e
age
rangesasinitial
survey,no
responses51-65
Gameplay withcontrollers?
37. Usability Metrics-Task Completion Performance
SKIPPEDALL
SKIPPEDALL
SKIPPEDALL
SKIPPEDALL
Zero values for a
single task means
user started but
then quit/skipped
that task.
38. Summary Statistics - Tasks
Task 1 Completed
(seconds)
Task 2 Completed
(seconds)
Task 3 Completed
(seconds)
Mean 49.00 Mean 25.50 Mean 9.71
Standard Error 28.00 Standard Error 5.65 Standard Error 2.41
Median 21.00 Median 28.50 Median 9.00
Mode 21.00 Mode 40.00 Mode 9.00
Standard Deviation 48.50 Standard Deviation 13.84 Standard Deviation 6.37
Sample Variance 2352.00 Sample Variance 191.50 Sample Variance 40.57
Kurtosis -2.40 Kurtosis 0.53
Skewness 1.73 Skewness -0.32 Skewness 0.92
Range 84.00 Range 31.00 Range 19.00
Minimum 21.00 Minimum 9.00 Minimum 2.00
Maximum 105.00 Maximum 40.00 Maximum 21.00
Sum 147.00 Sum 153.00 Sum 68.00
Count 3 Count 6 Count 7
39. Summary Statistics - Total Time & Survey
Age Group Gameplay
Mean 31-50 Mode 3 (controller, sometimes)
Standard Error 0
Median 31-50
Mode 31-50
Standard Deviation 0
Sample Variance 0
Range 0 Range 3
Minimum 31-50 Minimum 1
Maximum 31-50 Maximum 4
Count 6 Count 6
Time Spent in Prototype
(minutes)
Mean 50.66
Standard Error 32.29
Median 3.28
Mode 316.02
Standard Deviation 116.41
Sample Variance 13550.20
Kurtosis 3.23
Skewness 2.18
Range 315.13
Minimum 0.88
Maximum 316.02
Sum 658.53
Count 13
40. Navigation into Modes
Perhaps forcing a
practice session
and/or additional
instructions would
have increased ease
of use.
45. Event Metrics - Keyboard
Used to navigate
media menu and
select media
options.
46. Did the prototype achieve goals? Some...
No ‘Easy’ responses,
result distribution inconclusive
Need to analyze
data further to find
‘time spent in distracted views’
47. Goals - easy to learn controls
While more analysis
needs to be done to
determine length of time
spent in distracted
modes, this MAY
indicate preference for
using keyboard controls
in the undistracted view!
48. Goals - user opinions
Users also commented that it was difficult to remember
what song/album/station to target and wanted an
on-screen reminder during task operation.
50. Interesting
Finding:
Time Spent in Prototype
vs
Game Controller Use
(per person)
If this user had been slower, might
have been an inverse correlation.
These are people at work leaving the
window open/multitasking, so may
skew results.
51. ● Analyze data more thoroughly, tracing user paths.
● Explore alternate menu structures & tasks.
● Resolve issues with user confusion about tasks, ‘force’ a practice
session or trial run, assure surveys are filled out for every
respondent.
● Explore if video game controller use (with directional pads)
affects performance, learning and preference in using steering
wheel directional-pad controls (mapping/
applying previously learned mental models,
hand-eye coordination, recognition)
Opportunities for Improvement/Further Research