Embedding smartness into validation and verification of automated driving functions

Execution of hundreds of thousands of test runs with a trustworthy virtual vehicle in a cluster or cloud environment is a challenge in itself. With Simcenter Prescan360, we introduced an off-the-shelf solution to generate scenario variants, orchestrate the massive numbers of test runs and post-process these results for easy interpretation. However, by adding additional smartness, the process can be optimized and automated much further. In this presentation we will highlight how design exploration and AI techniques are applied to create critical but unseen scenarios that might make the system fail. We will also introduce smart metrics to automate accurate safety assessments of the automated driving functional performance.