Opportunity Knocks: A Community Navigation Aid Henry Kautz Don Patterson Dieter Fox Lin Liao University of Washington Computer Science & Engineering
Mar 29, 2015
Opportunity Knocks:A Community Navigation
Aid
Opportunity Knocks:A Community Navigation
Aid
Henry KautzDon Patterson
Dieter FoxLin Liao
University of WashingtonComputer Science & Engineering
OutlineOutline
1. Need for navigation aids2. Limitations of current devices3. Advancing the State of the Art4. Prototype: Opportunity Knocks5. Future plans
The Need: Community Access for the Cognitively
Disabled
The Need: Community Access for the Cognitively
Disabled
Problems in Using Public Transportation
Problems in Using Public Transportation
•Learning bus routes and numbers
Problems in Using Public Transportation
Problems in Using Public Transportation
•Learning bus routes and numbers
•Transfers, complex plans
Problems in Using Public Transportation
Problems in Using Public Transportation
•Learning bus routes and numbers
•Transfers, complex plans
•Recovering from mistakes
ResultResult
•Need for extensive life-coaching
•Need for point-to-bus service
ResultResult
•Need for extensive life-coaching
•Need point-to-bus service
•Isolation
Current GPS Navigation Devices
Current GPS Navigation Devices
Designed for drivers, not bus riders!Should I get on this bus?Is my stop next?What do I do if I miss my stop?
Requires extensive user inputKeying in street addresses no fun!
Device decides which route is “best”
Familiar route better than shorter one
VisionVisionCan we build a system that...
Automatically learns the daily pattern of the user’s transportation plans – no typing!Provides proactive help in carrying out the plansHelps user recover from mistakesIs compatible with today’s transportation infrastructure
ApproachApproach
User carries GPS cell phoneSystem infers bus use from
Position (near bus stop?)Velocity (on foot? in a vehicle?)Bus route information
Over time system learns about user
Important places Common transportation plans
Mismatches = possible mistakes
Inferring GoalsInferring Goals
Transportation Plans Transportation Plans
BA
Goalswork, home, friends, restaurant, doctor’s, ...
Trip segmentsHome to Bus stop A on FootBus stop A to Bus stop B on BusBus stop B to workplace on Foot
Work
User ModelUser Model
System learns a probabilistic model of the user’s pattern of transportation useRobust even if...
Data is missingBehavior varies
User is predictable, but not rigid!
xk-1
zk-1 zk
xk
mk-1 mk
tk-1 tk
gk-1 gk
Goal Prediction
Error Detectio
n: Missed
Bus Stop
Prototype: Opportunity KnocksGPS camera-phone“Knocks” when opportunity to help
Can I guide you to a likely destination?I think you made a mistake!This place seems important – would you photograph it?
ExampleExample
ExampleExample
ExampleExample
Example(continued
)
Example(continued
)
Example(continued
)
Example(continued
)
Example(continued
)
Example(continued
)
StatusStatus
Proof of concept prototypeCan use machine learning to create a smarter, more useful personal navigation system for disabled persons
Basic scientific contributions on predicting human behavior
Best paper award at national computer science conference, AAAI
Next StepsNext Steps
User interface design, testingAudioGraphicalAdaptive
Develop & deployHave begun discussions with METROSeek commercial partnerships
ENDEND
Probabilistic Model:Dynamic Bayesian
Network
Probabilistic Model:Dynamic Bayesian
Network
xk-1
zk-1 zk
xk
mk-1 mk Transportation mode
x=<Location, Velocity>
GPS reading
tk-1 tk
gk-1 gk Goal
Trip segment
Error DetectionError Detection
Approach: model-selectionRun two trackers in parallel
Tracker 1: learned hierarchical modelTracker 2: untrained flat modelEstimate the likelihood of each tracker given the observations
Novelty DetectionNovelty Detection