Automotive safety verification under temporal failure of adaptive cruise control system using statistical model checking

Abstract

Recently, a significant progress has been achieved in developing advanced driver assistant systems that improve the driver's comfort and safety. Adaptive Cruise Control (ACC) system is an example of these assistant systems which automatically controls the throttle and braking systems to maintain a safe distance between vehicles. Existing safety verification techniques for these systems consider the operation in normal conditions without taking the failure in sub-components into account for the safety assessment. In this work, we introduce an automotive safety analysis to investigate the impact of the temporal failures of the adaptive cruise control system. New probabilistic model of the ACC system is proposed based on Priced Timed Automata (PTA). This model is verified using statistical model checking technique. In this analysis, two temporal failure scenarios are considered: i) missing-control scenario, where neither throttle nor braking system receive control signals; ii) error-control scenario, where erroneous control signals due to failure in the distance sensors are introduced. The modeling and safety analysis of ACC system are fully automated by utilizing UPPAAL SMC. The analysis shows that the error-control failures have a worse impact on the automotive safety comparing with missing-control failures. In this work introduce a new insight to the automotive safety analysis under advanced driver assistant systems.