Details of Research Outputs

Autonomous cars are the future of transportation. However, safety is one of the major obstacles to the success of autonomous cars. The studies have shown that unsafe driving is the biggest problem for traffic accidents. In this study, we analyzed the issues of unsafe driving for autonomous cars and suggested that identification and correction of unsafe driving behaviors is the key to reduce traffic accidents. We proposed an algorithm to identify unsafe driving behaviors by using P2P benchmarking, then used a statistical model to evaluate the reliability of benchmark results. We also proposed an algorithm to correct undesirable behaviors for autonomous cars. The whole process can be done automatically without human intervention. The simulations showed that the proposed algorithm is workable.