Abstract

This paper presents an approach to fuse embedded systems in automotive or industrial applications applying a combination of graphs and fitness landscapes for identification of significant test cases. To identify these test cases graph-based space exploration with simultaneous refinement and reduction by fitness landscapes is used. Metrics to achieve test cases with high quality are generated from specification requirements. The methodology is presented in an automotive example utilizing an optical sensor for lane detection. The sensor is embedded in an existing framework for generation of synthetic environmental conditions.