Transcript Slide 1

Enhance Routing Efficiency and
Semantics through Participatory Sensing
Liviu Iftode/Ruilin Liu
Dept. of Computer Science, Rutgers University
Oct. 29, 2014
Routing of connected vehicles
• Connected vehicles vs. traditional way of traffic routing
– Real-time traffic based route planning (Themis)
– No way to evaluate the route choice (DoppelDriver)
– Weak semantics sensible/acceptable by drivers (NaviTweet)
“What would
have happened
if I took the
other route”
“I’m taking
route 66”
n up
Themis: A Participatory Navigation
System for Balanced Traffic Routing
• R. Liu, H. Liu, D. Kwak, Y. Xiangy, C. Borcea, B. Nath and L. Iftode, "Themis: A Participatory Navigation System For Balanced
Traffic Routing," accepted at the 2014 IEEE Vehicular Networking Conference (VNC).
Balanced routing: the concept
Balanced routing solution
 Participatory Sensing Algorithm
 Traffic Flow Estimation Algorithm
 Balanced Routing Algorithm
Flow Estimation
 Total Traffic = controlled traffic + background traffic
 Extrapolate the controlled traffic into the total traffic
▪ dynamic ratio
▪ sample roads have ground truth
▪ Estimate the ratio based on similarity to the sample roads
Aslam, J., et al. "City-scale traffic estimation from a roving sensor network." ACM Sensys, 2012.
Balanced routing algorithm
• Modified cost:
City-scale synthetic experiment
 Synthetic (Real Data + Simulation)
▪ 26,000 fleet of taxis in three consecutive Tuesdays
▪ Generate traffic demand for 7%, 20%, and 40% penetration rate
▪ Traffic simulation based on traffic-delay model
DoppelDriver: Counterfactual Actual
Travel Times
• D. Kwak, D. Kim, R. Liu, B. Nath and L. Iftode, “DoppelDriver: Counterfactual Actual Travel Times for Alternative
Routes," submitted for peer review.
What if I have taken the other route?
• Problem
– no way to compare and assess a route decision
– no systematic way to learn from them to make to make more
efficient decisions in the future
• How to integrate drivers’ experience into route decision
 sharing the actual travel times drivers experienced
 logging daily trips of actual travel times for chosen and nonchosen routes into a personal trip diary
Counterfactual Thinking in Routing
You are
Taken Route ATA
Counterfactual Route ATA
You would be
here if you
chose this
What if your
navigation could
tell you where
you currently are
on and how long
it took if you had
taken the other
routes to your
DoppelDriver algorithm
Comparison of Travel Times (Black vs. ∑ (Blue, Red, Green)
Feasibility study over taxi dataset
NaviTweet: Social Vehicle Navigation
• W. Sha, D. Kwak, B. Nath and L. Iftode,“Social Vehicle Navigation: Integrating Shared Driving Experience into Vehicle
Navigation,” in Proc. 14th International Workshop on Mobile Computing Systems and Applications (HotMobile’13),
Feb. 2013.
• D. Kwak, D. Kim, R. Liu, B. Nath and L. Iftode,“Tweeting Traffic Image Reports on the Road,” in Proc. Sixth
International Conference on Mobile Computing, Applications and Services (MobiCASE’14), Nov. 2014.
Social Vehicle Navigation
• Share traffic reports using voice and image (NaviTweet)
• Traffic Digest sent to interested drivers
→ complements factors such as ETA in route choice
→ provides semantically richer information
Tweet Digest
"traffic is about
to be cleared"
Route 66 or Route 22??
Social Vehicle Navigation - NaviTweet
O & D input
Route selection
Traffic report view
Navigation and Tweet
• 97 participants are included in our pilot study
• 30% of drivers change their route choices
• The weight of tweet information over the route choice is 3.9/5.0
compared with 3.97 of ETA
• 67% of users thought this system would be useful
Conclusion & future work
 Changing the route planning from one-directional style to
two-way participatory style
 Providing more semantics towards route choice is urgent
required by drivers
 Incentivizing a community is the key to success
 Incorporate multiple data sources and sensing platforms
to diversify the routing planning
Thanks for listening!
For additional information and other related projects, please refer to: