Implementation of a Simulation-Based Evaluation Environment for Autonomous Driving Service Verification

In this study, we developed a simulation-based evaluation environment using Vehicle-In-the-Loop (VIL) to assess the safety and performance of autonomous vehicles. Our approach combines real road tests with simulation to provide a comprehensive assessment of autonomous driving systems in various scenarios. We focused on testing the safety response function by simulating a LiDAR sensor failure. The results demonstrated that the safety response node accurately detected the sensor failure, activated the risk mitigation mode, and successfully brought the vehicle to an emergency stop using the AEB function. This validates the effectiveness of the VIL in evaluating safety response functions. This study highlights the potential of VIL in enhancing the evaluation and improvement of autonomous driving technologies, suggesting future research directions for expanding scenario complexity and simulation accuracy.