Formal verification of service-oriented adaptive driver assistance systems

Abstract

Many future Driver-Assistance-Systems (DAS) will use components not permanently mounted to the vehicle. Unlike state-of-the-art DAS with static configurations, the system and software architecture changes at runtime. To handle configuration changes, Service Oriented Architecture (SOA) and automatic orchestration is a promising approach. Whenever systems are set up automatically, they have to be validated. This paper presents an approach based on formal methods. Existing component models are annotated with Quality-of-Service parameters and transformed automatically to Hybrid Automata. These automata are then composed to an overall system model and model checking is used to check safety properties. The complete transformation-orchestration-validation process is executed without user interaction and thus can be performed at runtime.