Validation of Software Testing Results for Real-Time Vehicle Control Software

Monte Carlo black-box software testing of an experimental advanced vehicle control system revealed anomalous system behavior in certain slices of the real-time control system algorithm performance subspace. In this study, we validate these results by experimentally checking the control system using identical faults injected at the outputs of the sensors. The validation experimental results provide confirming evidence for the hypothesis that Monte Carlo black-box software results are representative of the actual performance of the PATH lateral control system.