Probabilistic Analysis of Weakly-Hard Real-Time Systems
This work addresses the need for more flexible timing analysis in safety-critical automotive systems, but it is incremental as it builds on previous methods.
The paper tackles the problem of analyzing timing constraints in automotive real-time systems by extending a specification language to support probabilistic analysis for weakly-hard constraints, demonstrating the approach on an autonomous vehicle case study.
Modeling and analysis of non-functional properties, such as timing constraints, is crucial in automotive real-time embedded systems. EAST-ADL is a domain specific architectural language dedicated to safetycritical automotive embedded system design. We have previously specified EAST-ADL timing constraints in Clock Constraint Specification Language (CCSL) and proved the correctness of specification by mapping the semantics of the constraints into Uppaal models amenable to model checking. In most cases, a bounded number of violations of timing constraints in automotive systems would not lead to system failures when the results of the violations are negligible, called Weakly-Hard (WH). Previous work is extended in this paper by including support for probabilistic analysis of timing constraints in the context of WH: Probabilistic extension of CCSL, called PrCCSL, is defined and the EAST-ADL timing constraints with stochastic properties are specified in PrCCSL. The semantics of the extended constraints in PrCCSL is translated into Uppaal-SMC models for formal verification. Furthermore, a set of mapping rules is proposed to facilitate guarantee of translation. Our approach is demonstrated on an autonomous traffic sign recognition vehicle case study.