Cellular-based Vehicle to Pedestrian (V2P) Adaptive Communication for Collision Avoidance Mehrdad Bagheri, Matti Siekkinen, Jukka K. Nurminen Aalto University - Department of Computer Science and Engineering - Espoo, Finland mehrdad.bagheri@aalto.fi, matti.siekkinen@aalto.fi, jukka.k.nurminen@aalto.fi Aalto University - Espoo, Finland ICCVE 2014 1
Realizing V2P for Road Safety in Autonomous Cars Pedestrian detection and V2P communication A crucial requirement for future autonomous cars Also for advanced driver assistance systems Different scenarios for road safety, collision warning and collision avoidance Bushes Trees Parked Car Parked Car Aalto University - Espoo, Finland ICCVE 2014 2
Cellular-based Active Road-safety Establish V2P communication, by means of smartphones, over cellular network. Integrate it with other detection methods. No need for new infrastructure Availability: for both vehicles (drivers) and pedestrians DSRC methods: still expensive, not deployed, not supported on phones Radar methods: dependent on line-of-sight Aalto University - Espoo, Finland ICCVE 2014 3
Cellular-based Active Road-safety System Setup Network nodes: Vehicles Cyclists Pedestrians Network nodes beacon periodically Reporting their location, speed and direction Processing made to evaluate the risks Alert the car to brake Aalto University - Espoo, Finland ICCVE 2014 4
Cellular-based Active Road-safety Deploy the road safety system Cloud back-end App on your mobile phone Aalto University - Espoo, Finland ICCVE 2014 5
Energy Consumption Problem A typical (non-adaptive) road-safety system: Network nodes beaconing periodically For road safety, recommended 10 Hz frequency (100 ms interval) [1],[2] Frequent beaconing consumes a large amount of power [1] ETSI EN 302 665 - Intelligent Transport Systems (ITS); Communications Architecture. [Online]. Available: http://www.etsi.org/deliver/etsi_en/302600_302699/302665/01.01.01_60/en_302665v010101p.pdf. [2] ETSI TS 102 637-2, Intelligent Transport Systems (ITS); Vehicular Communications; Basic Set of Applications; Part 2: Specification of Cooperative Awareness Basic Service. [Online]. Available: http://www.etsi.org/deliver/etsi_ts/102600_102699/10263702/01.02.01_60/ts_10263702v010201p.pdf. Aalto University - Espoo, Finland ICCVE 2014 6
Evaluate Negative Effect on Smartphone Battery Life Example: Messaging Frequency: 10 Hz (100 ms interval) Typical Battery Life = 16:00 Connection: LTE Battery Life: 04:39 (decreased more than 10 hours!) Battery life drops significantly Battery Life (hours) 18 16 14 12 10 8 6 4 Deployment with such setup is NOT practical Note: Problem with all radio-based methods (DSRC, WiFi) 2 0 Typical Battery Life Running Road Safety App Aalto University - Espoo, Finland ICCVE 2014 7
Adaptive Multi-level Communication Messaging Frequency: Adjusted according to collision risk Higher frequency Better precision and more reliability in prediction Higher frequency Worse battery life Try to minimize messaging frequency Messaging Frequency (Hz) 11 10 9 8 7 6 5 4 3 2 1 0 Time no risk low risk high risk no risk low risk no risk Aalto University - Espoo, Finland ICCVE 2014 8
Adaptive Multi-level Communication Simplified (Two Levels) Messaging Frequency: Adjusted according to collision risk Higher frequency Better precision and more reliability in prediction Higher frequency Worse battery life Try to minimize messaging frequency Messaging Frequency (Hz) 11 10 9 8 7 6 5 4 3 2 1 0 -- Low-rate beaconing -- -- High-rate beaconing -- Time ------------------- Low-rate beaconing --------------- no risk low risk high risk no risk low risk no risk Aalto University - Espoo, Finland ICCVE 2014 9
How to Set Message Frequency? High risk level (full-rate mode): 10 Hz recommended Other risk levels (other modes): Messaging frequency is influenced by risk factors such as: Vehicle Speed Distance to Pedestrian And by other factors such as: Communication Latency Reaction Time Aalto University - Espoo, Finland ICCVE 2014 10
How to Set Low-rate Message Frequency? Reaction + Brake: Min distance to alert the autonomous car Processing: Delay caused by communication (RTT) and computation Extra: chance to switch from low-rate to full-rate frequency higher speed shorter extra time need for a higher frequency Messaging Frequency (minimum possible) = f (relative speed, distance, cellular network specification) Aalto University - Espoo, Finland ICCVE 2014 11
How to Set Low-rate Message Frequency? Messaging Frequency = f (relative speed, distance, cellular network specification) Example: Set frequency for different max. car speeds Distance = 200 meters Communication: LTE, RTT 50 ms Messaging Frequency (Hz) 1.2 1 0.8 0.6 0.4 0.2 0 0 36 72 108 144 Car Speed (km/h) Aalto University - Espoo, Finland ICCVE 2014 12
Calculate Energy Consumption Electrical power consumption of different risk levels (modes): Pfullrate and Plowrate Calculated based on models for 3G or LTE [1]. They depend on: Energy required to keep radio on Energy required to send one beacon message Do not depend on: Amount of data sent (beacon message has a tiny size) Total electrical energy consumed, depends on: Cellular network specification Beacon messaging frequency Fraction of time spent in risky situations: tfullrate vs. tlowrate Padaptive = ( Pfullrate tfullrate + Plowrate tlowrate ) / ( tfullrate + tlowrate ) [1] M. A. Hoque, M. Siekkinen, J. K. Nurminen, S. Tarkoma, and M. Aalto, Saving energy in mobile devices for on-demand multimedia streaming -- a cross-layer approach, ACM Transactions on Multimedia Computing, Communications, and Applications, vol. 10, no. 3, pp. 1 23, Apr. 2014. Aalto University - Espoo, Finland ICCVE 2014 13
Results with an Example Example Scenario: Battery Original Life-time = 16:00 Battery Capacity = 7.98 Wh Connection: LTE, RTT 50 ms Max Car Speed = 108 km/s (30 m/s) Time spent in risky situation: 1 hr (out of 24 hrs) Low-rate Frequency: 0.29 Hz (3.4 seconds interval) Battery life: 13:38 (decreased 02:22) Practical smartphone battery life Enables deployment of pedestrian safety systems Battery Life (hours) 18 16 14 12 10 8 6 4 2 0 Non-adaptive Communication Typical Battery Life Adaptive Communication Aalto University - Espoo, Finland ICCVE 2014 14
Results Summary Battery Life = f (relative speed, distance, cellular network, tfullrate / tlowrate) Battery life depends on risky situations encountered (e.g. during 24 hour) Many pedestrians spend a limited time in risky situations (e.g. less than 1 hour) Full-rate mode duration can be reduced even more with context-aware filters and optimization Practical smartphone battery life Enables deployment of pedestrian safety systems 1 Aalto University - Espoo, Finland ICCVE 2014 15
Conclusion Dealt with limited energy source on smartphones Proposed adaptive multi-level method Results in practical energy consumption Enables deployment of pedestrian safety systems Aalto University - Espoo, Finland ICCVE 2014 16
Discussion and Future Work Simulation Analyze following aspects Energy-efficiency of Different Smartphones Urban Areas and Pedestrian Daily Mobility Pattern Prototype Aalto University - Espoo, Finland ICCVE 2014 17
Thank you! Any questions? Aalto University - Espoo, Finland ICCVE 2014 18