Adaptive testing of controllers for autonomous vehicles

The authors discuss techniques for the evaluation of complex software systems, and for the identification of classes of vehicle faults that are most likely to impact negatively on the performance of a proposed autonomous vehicle controller. The approach involves subjecting a controller to an adaptively chosen set of fault scenarios within a vehicle simulator, and searching for combinations of faults that produce noteworthy performance by the vehicle controller. The search uses a genetic algorithm. The approach is illustrated by evaluating the performance of a subsumption-based controller for an autonomous vehicle. The preliminary evidence suggests that this approach is an effective alternative to manual testing of sophisticated software controllers.<
>