autoui2013 rick swette-talk

32
Comparing three novel multimodal touch interfaces for infotainment menus

Upload: rick-swette

Post on 21-Jun-2015

357 views

Category:

Automotive


2 download

DESCRIPTION

Automotive UI 2013 - Design Talk - Auditory Menus and Interaction design For the Vehicle. 3 Design takeaways.

TRANSCRIPT

Page 1: Autoui2013 rick swette-talk

Comparing three novel multimodal touch interfaces for infotainment menus

Page 2: Autoui2013 rick swette-talk

“The New Standard”

http://www.youtube.com/watch?v=F-oE3ZzkqxY

Page 3: Autoui2013 rick swette-talk

“How can we use sound to improve the driving experience?”

The “experience” is some combination of allowing the user to engage in secondary activities while always prioritizing safety.

Page 5: Autoui2013 rick swette-talk

The Goal

Demonstrate a device that has the following qualities:

Reduce the negative impact of secondary task on driving performance or total visual demand,

At an acceptable cost for secondary navigation speed.

How

Compare them experimentally against common methods (i.e. direct touch) on a common task (i.e. menu navigation).

The Study

Page 6: Autoui2013 rick swette-talk

4 interfaces

see them in action

Page 7: Autoui2013 rick swette-talk

Measuring....

- Driving Performance- Total Visual Demand- Speed of operating the menu

see this in action

While.....

Selecting items which are either 1 – 3 levels deep in a hierarchical menu

(same video as on previous slide)

Page 8: Autoui2013 rick swette-talk

Results

Page 9: Autoui2013 rick swette-talk

Serial Swiping had significantly better driving performance than the

rest.

(lowest deviation)

Page 10: Autoui2013 rick swette-talk

...But under Direct Touch users were twice as fast to complete the menu operation task

as all the others.

2 TIMES AS FAST!!!!!

Best Driving Performance!!!!!

Page 11: Autoui2013 rick swette-talk

Total visual demand of the “visual based” Direct Touch screen was equal to that of the multimodal Serial Swipe.

.` =

Page 12: Autoui2013 rick swette-talk

No significant differences in the distribution of glance lengths either, but definitely some longer looks

Page 13: Autoui2013 rick swette-talk

Gaze Percentage:Time Looking at the Interface / Total Task completion time.

Page 14: Autoui2013 rick swette-talk

Same total visual demand distributed over a much shorter amount of time =

Worse Driving Performance

Page 15: Autoui2013 rick swette-talk

Results Part 2:

Diving deeper

Page 16: Autoui2013 rick swette-talk

~2.5 seconds

~1.25 seconds

That’s an average of about 1~ second of waiting around before hitting the target.

~4 seconds

~4 seconds

Users begin working immediately and in parallel to driving.

Time To Complete

while driving

while not driving.

Page 17: Autoui2013 rick swette-talk

In the lane change task every ~10 seconds a command is sent.

Often the user can execute a changing of the menu with little risk of missing a command.

Page 18: Autoui2013 rick swette-talk

but the story may change for multiple-step tasks.

It is likely the results would not have been significant if multiple step tasks were not introduced.

Page 19: Autoui2013 rick swette-talk

no change in total task time between 2 and 3 step tasks!

The willingness to commit such actions so speedily in succession is the real problem.

Page 20: Autoui2013 rick swette-talk

Users are either...

Adding on one more step and not waiting at all between one of the steps

or...

waited on average less between each selection simply because they were given more work.

Page 21: Autoui2013 rick swette-talk

User’s exhibiting wonderful stair-case like sub-tasking.

Compare to our winner:

Page 22: Autoui2013 rick swette-talk

Lessons Learned:

3 lessons for designing in the vehicle

Page 23: Autoui2013 rick swette-talk

Because Direct Touch is faster, the user may be

tempted to fit more selections

before the next command,

1.

and maybe make a miss.

Page 27: Autoui2013 rick swette-talk

For the GRUV prototypes, work is lost if you remove your hand from the touchpad. In other words, regaining your position is too difficult.

It’s exactly the opposite for Serial Swipe, which does not require “starting over” when interrupted.

3.

Page 28: Autoui2013 rick swette-talk

Lesson 3

Afford interruption.

Progress must not be lost if interrupted.

Page 29: Autoui2013 rick swette-talk

Discussion Questions

Page 30: Autoui2013 rick swette-talk

Discussion QuestionHow could these principles be applied to context where

safety should be considered when wearing Google Glass?

1. Built in slowness 2.Design out “shortcuts” 3. Design for interruption

http://www.youtube.com/watch?v=IZdkIVS53Uw

driving constructionhttp://www.youtube.com/watch?v=yWK-Uinxn40

Page 31: Autoui2013 rick swette-talk

Discussion QuestionCan voice do it all? I’d argue no.

In the following example, a user can choose up to 3 ways of doing the same thing.

So then, how do you afford the “right mode switching” without suffering the costs of modes?

http://www.youtube.com/watch?v=F-oE3ZzkqxY

Page 32: Autoui2013 rick swette-talk

Discussion Question

Is the touchpad where it’s at – how can the touchpad realize its full potential ?

http://www.youtube.com/watch?v=t3SpNJT88_o