CRAG at the SBFT 2023 Tool Competition - Cyber-Physical Systems Track
Paolo Arcaini, Ahmet Cetinkaya · 2023
CRAG is a combinatorial testing-based generator of road scenarios for autonomous driving testing. The parameters of the combinatorial model are used for determining lengths and curvatures of a road in the Frenet frame representation. CRAG first generates a test suite of roads achieving pairwise coverage, and then incrementally searches for roads achieving combinatorial strength k, but only for the k-tuples that contain (k-1)-tuples having optimal test performance in the previously generated roads.