Mobile Millennium Part of the Connected Traveler Project

1
Mobile MillenniumPart of the Connected Traveler
Project

• Quinn A. Jacobson, Ph.D.
• Nokia Research Center Palo Alto

2
Connected Traveler Project

• Part of the SafeTrip-21 initiative of the VII
  Program
• Unique Public-Private Partnership
• Caltrans, UC Berkeley, Nokia and Navteq
• Leveraging existing technology
• Cell phones and after market automotive
  accessories
• Early deployment consistent with VII long term
  vision
• Making the traveler more informed and aware
• Two Distinct Parts/Tracks
• Group Enabled Mobility and Safety (GEMS)
• Mobile Millennium (MM)

3
Group Enabled Mobility and Safety

• Framework for sharing information to provide
  mobility and safety information to travelers
• Through a variety of platforms available today
• Internet connected computers
• Mobile devices connected through WiFi or cellular
  network
• After-market DSRC devices (not yet widely
  available)
• Supports trusted content
• Routing/scheduling info, warnings and
  informational bulletins
• Authenticated updates from DOTs, transit
  authorities, etc
• Supports community generated content
• Rich data to enhance the experience of the
  connected traveler
• Partnership UC Berkeley, PATH, Caltrans, Navteq

4
Mobile Millennium

• Is a set of technologies to get better traffic
  data
• Innovative way to collect consumer GPS probe
  data(from participants with GPS-equipped mobile
  phones)
• New algorithms for traffic modeling(including
  arterial roads)
• New ways to get real-time traffic data back to
  users
• Is a large scale pilot
• To drive the refinement of the technologies
• Prove the robustness, quality usefulness of the
  service
• Evaluate the quality and usefulness of the data
• Partnership UC Berkeley, CCIT, Caltrans, Nokia,
  Navteq

5
Community Enhanced Traffic

• Users contribute locally observed traffic data
• Accurate GPS data shared in privacy preserving
  way
• Aggregated with other sources and processed
• Navteq (proven system), Berkeley (next-gen
  algorithms)
• Users get benefit of enhanced quality and
  coverage
• Real-time traffic information on their connected
  device
• Quality and coverage naturally grows with user
  base

Local Data from Users(User Privacy Protected)
Probe Data
GPS Positioning Data
Sensor Data(DOT Private)
Traffic Model
Traffic Service
Traffic (TMC) Feed
Existing Road Sensors
Global Data to Users(Visual and Audio Interfaces)
6
Current Traffic Sensing in use
Good, but we can do better

• Current infrastructure sensors for traffic
• Inductive loops in pavement, wireless pavement
  detectors, cameras, license plate readers, toll
  transponders
• High cost to install and maintain
• Only cover a small fraction of highways in a few
  countries
• Cellular network monitoring approach
• Monitor cell phones connected to towers
• Must calibrate and recalibrate for each area
• Large variations in accuracy

7
GPS-enabled Mobile Devices
The perfect Traffic Monitoring Device
Device knows its location and speed from on-board
GPS
Device can compute appropriate times and places
to send information
Device can perform appropriate actions to protect
privacy of the user
Device can talk to the outside world through the
wireless network
8
Virtual Trip Lines for Probe Data Collection

• Virtual lines at intelligently placed locations
• Phone/device uses GPS to detect crossing of trip
  lines
• Anonymously reports crossing (with some
  probability)with position, speed travel time
  info
• Benefits
• Very efficient (only send data in areas of
  interest)
• Good privacy
• No infrastructure required
• Dynamic in space and time

9
Privacy Features
Ensure PRIVACYVTL Placement Restrictions
End-to-end encryption
VTLDataBase
ID ProxyServer
VTLServer
Cache of local VTLs
Detect crossing
VTLUpdate Log
Make Updates ANONYMOUSStrip IP addressCheck
right to use systemFilter suspicious updates
LOCAL FILTER Quantize, temporal shift, prob.
updates
Ensure PRIVACYAggregation Privacy
FilteringQuery Restrictions
10
Traffic Data Processing

• Navteq has extended their processing engine
• To handle new consumer GPS probe data feeds
• Integrated with all other data feeds
• Proven quality and robustness
• UC Berkeley
• Developing next generation traffic processing
  algorithms
• Specifically designed around consumer probe data
• Both highway and arterial road models
• Both real-time and predictive support

11
Service Back to Users

• Users benefit from high-quality, real-time
  traffic data
• Delivered to their cell phone (or other connected
  device)
• Developing innovative visualization of traffic
  data
• Good coverage can lead to information overload
• Leveraging existing Navteq technology for audio
  reports

12
February 8th Field Trial (Mobile Century)

• UC Berkeley / CCIT / Nokia Collaboration
• Funded by Caltrans and Nokia
• 100 cars, 160 drivers, 8 hours, 20 miles of road
• Carrying connected mobile device (Nokia N95
  phone)
• Giving a glimpse into the future
• 2-5 penetration of vehicles on the road
• Real-time data collection, processing and
  reporting

13
Example of GPS probe potential
14
GPS probe compared to Loop Detectors
15
The Mobile Millennium Pilot

• Large scale public pilot to be run soon
• Jointly run by California Center for Innovative
  Transportation (CCIT) and Nokia
• In collaboration with UC Berkeley, Caltrans and
  Navteq
• Other partners to be named soon
• Leveraging the prevalence of GPS-enabled cell
  phones