Use Cases for Simulation in the Development of Automated Driving Systems

Abstract

Developing an automated driving system (ADS) is difficult, because a large, het-erogeneous set of traffic situations has to be handled by the system. At present, data collection does not provide the means to capture them sufficiently compre-hensive. And real-world testing is not feasible in the required amount. So virtual techniques, namely simulation, will play an important role in the development of an ADS, in particular in its verification and validation. Because such a system is highly safety critical, the development will have to follow at least the process recommendations of the ISO 26262 (or similar directives). This permits to identify activities which can be supported by simulation within the overall development context. This paper describes three such use cases for simulation in detail. Each use case is presented with its objective, the scope and level of the modeling, and tooling. Further, it is discussed what it takes or would take to realize the use case-what appears to be available and what might be lacking. The paper thus contributes to an overview of the role virtual testing and explora-tion techniques might play in the development of ADS. It provides guidelines for practitioners who are faced with the task of selecting the methods to support a complex development process. And it hints at shortcomings of present-day tools, or necessary prerequisites to use those which are available.