Development of Complex Scenarios and Control Algorithms for Autonomous Driving Functions (ADFs) in a Driving Simulator

Autonomous vehicles and Advanced Driver Assistance Systems (ADAS) applications are emerging fields of research and development that require specialized controllers to ensure safe and efficient functionality. However, testing these controllers in real-world scenarios can be costly and risky, making the use of simulators necessary for their design and validation. This thesis presents a framework for the control and simulation of fundamental ADAS applications for autonomous vehicles, using Simulink as controller design environment and Carla as simulation environment. Firstly, the thesis describes how these tools were integrated. Then, the focus shifts to the development and testing of two control methods: PID and Nonlinear Model Predictive Control (NMPC). On the one hand, the PID method is applied to the lateral control of the vehicle using a single-track model for the tuning process. On the other hand, the NMPC method is applied to both the lateral and longitudinal control of the vehicle. To address limitations of the single-track model, which lacked direct throttle and brake control capabilities, a new vehicle model is derived through system identification from the vehicle data collected in the Carla simulation environment. Finally, a comprehensive testing scenario is designed within the Carla simulator, consisting of a straight road for acceleration and a curve that requires deceleration and steering. Simulation results verified the effectiveness of the proposed control strategies.