Tool-Supported Dependability Analysis of Semi-Markov Processes with Application to Autonomous Driving

For all safety-critical systems, modelling approaches allowing accurate dependability analysis is of utmost importance. Moreover, in order to accurately and realistically model the real-world behaviour of safety-critical systems, Semi-Markov Processes (SMPs) are highly useful. SMPs generalize Markov processes to give more freedom in how a system can be modelled. While SMPs are highly useful, previous literature fail to provide an intuitive modelling approach for them. As the first contribution, an intuitive novel modelling approach based on transition timers is presented. Once the systems of interest has been modelled as a SMP the problem still remains to analyze the model. As the second contribution, the Matlab app “SMP-tool” is presented. SMP-tool can analyze SMPs in several manners, perhaps most importantly by performing a dependability analysis through calculating the reliability and availability. Both contributions are applied to Highway Pilot, a typical system for autonomous driving.