Study on Full Scale Testbench Based Vehicle-In-The-Loop Simulation Testing for Automated Vehicles

Testing and validating automated vehicles is crucial but challenging due to constraints in space, time, cost, and the risks associated with using real vehicles. Driving simulators have emerged as an alternative way for this purpose, but they have difficulties in replicating the dynamic characteristics of the vehicle. To solve this issue, this study aims to apply a vehicle-in-the-loop (VIL) simulation verification method that reflects vehicle characteristics for testing and validating fully autonomous vehicles. This is achieved by receiving the responses of the autonomous vehicle controller via a testbench-based chassis dynamometer and providing the sensor outputs required for full autonomy to the autonomous vehicle controller through a driving simulator. The sensor outputs implemented in the driving simulator are provided to the existing advanced driver assistance system (ADAS) through over-the-air (OTA) stimulation, which uses physical sensor signals but originally generated in virtual world. In addition, for simultaneous transmission of high-level sensor data, such as multi-channel LiDAR and cameras, which are technically and financially challenging for OTA stimulation, a dedicated sensor data interface and packet structure designed for this purpose is used to inject the sensor data directly into the autonomous vehicle controller to validate fully automated vehicles.