Exploration of Distributed Automotive Systems Using Compositional Timing Analysis

This chapter presents a design space exploration method for mixed event-triggered and time-triggered real-time systems in the automotive domain. A design space exploration model is used that is capable of modeling and optimizing state-of-the-art automotive systems including the resource allocation, task distribution, message routing, and scheduling. The optimization is based on a heuristic approach that iteratively improves the system design. Within this iterative optimization it is necessary to analyze each system design where one of the major design objectives that needs to be evaluated is the timing behavior. Since timing analysis is a very complex design task with high computational demands, it might become a bottleneck within the design space exploration. As a remedy, a clustering strategy is presented that is capable of reducing the complexity and minimizing the runtime of the timing analysis. A case study gives evidence of the efficiency of the proposed approach.