Enhancing Safety and Handling of Autonomous Vehicles in Simulated Traffic: A Co-Simulation Framework Using Multibody Dynamics and CARLA

In recent years, there has been an increased emphasis on autonomous driving technologies to improve vehicle road safety amidst rising traffic congestion and the complexities of intersection jaywalking and diverse road conditions. Therefore, improving the vehicle's handling ability is crucial for safe and efficient traffic navigation, particularly emphasising collision prevention and safety in unforeseen circumstances. Evaluating safety perspectives in such situations, the lane change event serves as an important measure for addressing the matter and forms the focus of this paper. However, for such new-age technology, conducting proving ground tests replicating urban conditions is a costly endeavour. Hence, simulation is a better approach, which can mimic real traffic conditions, develop control systems, and simulate vehicle handling behaviour all working together within a closed-loop system. An autonomous lane change manoeuvre event of a four-wheeled vehicle is created in a simulation environment where the system will initiate the lane change manoeuvre based on the relative distance of the slow-moving preceding vehicle to the ego vehicle. The trajectory of the vehicle will be determined by the type of slow-moving vehicle and the dimensions of the road. A series of lane change events generated by MBD simulation will be converted into a reduced-order model (ROM) and integrated into the system. The control system will be developed using MBSE and co-simulated with CARLA for testing the system in a CARLA environment, replicating urban conditions. This enables the integration of accurate dynamics of the vehicle into CARLA, ensuring a reliable manoeuver. To check the robustness of the system, various traffic scenarios will be simulated.