Abstract:
Enhanced capabilities of sensors and digital maps for intelligent vehicles lead to a complex and multivariant system environment with a broad variety of situations and tr...Show MoreMetadata
Abstract:
Enhanced capabilities of sensors and digital maps for intelligent vehicles lead to a complex and multivariant system environment with a broad variety of situations and traffic scenarios. To assure the feature under development's valid behavior, the sample of scenarios evaluated for Verification and Validation (V&V) needs to proof substantial coverage of all possible situations. Currently applied V&V activities on system-level are in a large part based on real world tests. These are not scalable to sufficiently cover the variant system environment. Our previously introduced Reactive-Replay enables substantial coverage by reuse of recorded real world data in closed-loop simulation. In this contribution we present an approach to determine the relevance of recorded scenarios and derive efficient sets of test scenarios. Our two-step approach starts with a specification-based classification-tree for initial scenario selection. A data-driven reduction of the initial scenario set is achieved by the following analysis of covered parameter spaces. The final consolidated test set avoids repetitive situations while ensuring a significant diversity of the sampled system environment.
Date of Conference: 27-29 June 2017
Date Added to IEEE Xplore: 05 February 2018
ISBN Information: