AutoSup: Driving Behaviors Recognition Supervision via Incentive Mechanism
Sheng Cao, Yashuang Mei, Peng Xiangli, Xiaoming Huang, Xiaosong Zhang · 2021
Reducing driving accidents is usually the most concerned goal of all the taxi sharing platforms, like DiDi, Uber, etc. Fatigue driving is one of the important factors causing driving accidents. Current solution to avoid fatigue driving is to force the drivers to stop and rest for a period of time after driving continuously for more than designated hours. It’s a one-size-fitsall solution and there is no different treatment policy according to the specific situation of each driver. In this paper, we propose AutoSup that is a blockchain-based method to recognize abnormal driving behaviors of drivers and give each driver the personalized rest reminder with incentive approach. First, we propose a new way for head posture recognition to detect abnormal driving behaviors. A rest reminder alert is issued when the recognition threshold is triggered via smart contracts, which lets taxi sharing platforms perform precisely to automatically inform tired drivers instead of fixed time limit. Furthermore, we apply a novel credits system to record abnormal driving behaviors, the taxi sharing platforms could give the corresponding driver bonus credits or penalty credits respectively, which is to simply and effectively reward the compliant drivers and punish the illegal drivers. Experiments are conducted to verify the effectiveness. This work distinguishes itself from others by helping taxi sharing platforms to supervise and evaluate the driving behaviors of drivers more openly, fairly and efficiently, and it could improve harmonious service relationship between the sharing platforms and drivers in the real world.