Security Verification Approaches For Navigation Of Autonomous Vehicle

The autonomous vehicle was an integrated intelligence system which included many functions such as environmental awareness, planning and decision-making, and multi-level driver assistance. It involved the using of the technologies like computer, modern sensing, information fusion, communications, artificial intelligence and automatic control, and so on. It was a typical high-tech complex and the major research part of the future intelligent transportation systems. The autonomous navigation was the signal technology of the autonomous vehicle. Safe driving was the ultimate goal.In this paper, the safety of the navigation behavior for autonomous vehicle was the object of study. It used two different methods, simulated verification and formal verification, to research on the safety verification problems of the autonomous vehicle’s navigation behavior decision.The research of this paper:(1) Security authentication method of the autonomous vehicle’s navigation behavior decision based on system simulated verification technology.This paper proposed a security authentication method of the autonomous vehicle’s navigation behavior decision based on system simulated verification technology. This method introduced the application of road curvilinear coordinate system, and considered random variables as the control input of the vehicle. Then it used Monte Carlo sampling techniques to simulate the random behavior of the system.(2) Security authentication method of the autonomous vehicle’s navigation behavior decision based on formal verification technology.For the fact that autonomous vehicles were hybrid systems with high security requirements, the paper established a hybrid automaton model of vehicle control based on the basic concepts and theories of the hybrid system. Then it proposed a online security authentication method of the autonomous vehicle’s behavioral decisions based on the using of the reachability analysis techniques and methods.Innovation and uniqueness of the work:(1) For the actual changing road shape, the paper introduced the road curvilinear coordinates system to simplify the dynamic model of vehicle and road models.(2) This paper proposed an iterative sampling process based on the uniform sampling which could improve the efficiency in the use of the sample.(3) The paper applied hybrid systems verification techniques to the field of vehicle control, and the Markov Chains were used to abstract the future stochastic reachable sets of the vehicle and analysis and calculate the security of vehicle behavior decisions.