Tracking Stolen Bikes through Everyday Mobile Phones and Participatory Sensing Ted Tsung-Te Lai Chun-Yi Lin Ya-Yunn Su Hao-Hua Chu National Taiwan University

  • View
    215

  • Download
    2

Embed Size (px)

Text of Tracking Stolen Bikes through Everyday Mobile Phones and Participatory Sensing Ted Tsung-Te Lai...

  • Slide 1
  • Tracking Stolen Bikes through Everyday Mobile Phones and Participatory Sensing Ted Tsung-Te Lai Chun-Yi Lin Ya-Yunn Su Hao-Hua Chu National Taiwan University 1
  • Slide 2
  • Bikes are everywhere 2 Cyclists face many problems
  • Slide 3
  • Safety (CyberBike, HotMobile10) 3
  • Slide 4
  • Route quality (BikeStatic, CHI10) 4
  • Slide 5
  • Fitness (BikeNet, SenSys07) 5 Sensors: -Heart rate -GPS -Accelerometer etc
  • Slide 6
  • Bike Theft (BikeTrack) 6
  • Slide 7
  • Bike Theft Survey (208 students) 1 out of 1.8 person has bike stolen experience 1 out of 3.7 stolen bikes was recovered Mostly found on campus Is it possible to use participatory sensing to recover stolen bikes? 7
  • Slide 8
  • 1.Motivation 2.BikeTrack System design 3.Evaluation and preliminary results 4.Future work 5.Conclusion Outline 8
  • Slide 9
  • BikeTrack overview 9 BluetoothBikeUsers use phone to scan Bluetooth Log BluetoothID/Location/Timestamp Server for bike location query data
  • Slide 10
  • Spec: 20-meter radio range 1.5-month lifetime 16 USD/tag Customization: Only broadcast beacon ID Why Bluetooth? Available on almost every phone Bluetooth beacon tag 10
  • Slide 11
  • Bluetooth tag mounting on a bike 11
  • Slide 12
  • Phone implementation Android 2.1 Scan Bluetooth ID every 20secs in background When a Bluetooth ID is found, it logs Auto-upload data during network availability Bluetooth IDLocationTimestamp 12
  • Slide 13
  • Server implementation Linux + Apache + MySQL Web interface to query bike location on google map 13 Bike locations
  • Slide 14
  • 1.Motivation 2.BikeTrack system design 3.Evaluation and preliminary results 4.Future work 5.Conclusion Outline 14
  • Slide 15
  • User study Two-week during summer 11 CS grad students Dataset: 3700 bluetooth/location/times entries 3500 self-detection; 200 detection of other users Constraint: 15 CS department layout
  • Slide 16
  • 1.How well does participatory sensing work in tracking bikes? 2.Is it possible to locate stolen bike on campus? 3.Is it possible to reduce battery consumption based on user behaviors ? Evaluation and preliminary results 16
  • Slide 17
  • Avg. Bluetooth detections/day All bikes were detected Avg. detection rate: 5.1 times/day 17
  • Slide 18
  • 1.How well does participatory sensing work in tracking bikes? 2.Is it possible to locate stolen bike on campus? 3.Is it possible to reduce battery consumption based on user behaviors ? Evaluation and preliminary results 18
  • Slide 19
  • Bike location distribution in Taipei 19
  • Slide 20
  • Bike location distribution at NTU 20
  • Slide 21
  • 1.How well does participatory sensing work in tracking bikes? 2.Is it possible to locate stolen bike on campus? 3.Is it possible to reduce phone battery consumption based on user behaviors ? Evaluation and preliminary results 21
  • Slide 22
  • Avg. user detection pattern during a day 22 Detection happened at noon, dinner, end of a day Detection pattern varies with users Future optimization (currently scan/20 seconds)
  • Slide 23
  • 1.Motivation 2.BikeTrack System design 3.Evaluation and preliminary results 4.Future work 5.Conclusion Outline 23
  • Slide 24
  • Formulating deployment strategy 24 How to incorporate user spatial-temporal model to reduce phone overhead? How to incentivize participation?
  • Slide 25
  • 1.Motivation 2.System design 3.Evaluation and preliminary results 4.Future work 5.Conclusion Outline 25
  • Slide 26
  • BikeTrack - A low cost participatory sensing system for bike tracking Preliminary result shows that BikeTrack is a promising system to locate bikes Conclusion 26
  • Slide 27
  • Questions & Answers BikeTrack: Tracking Stolen Bikes through Everyday Mobile Phones and Participatory Sensing Ted Tsung-te Lai Chun,Yi Lin, Ya-Yunn Su, Hao-Hua Chu National Taiwan University 27