Investigation of the contribution of intelligent suspension system on ride comfort and road holding quality in autonomous vehicles : A case study at Volvo Cars Corporation

Sammanfattning: Transformation to automated drive implies a switch in roles from driver to passenger, which means that the driver goes from complete flexibility to maneuver the car to almost none. There would be a mismatch between the human's preferred driving style and the driving behavior provided by an autonomous car, leading to both mental and physical discomfort amongst the passengers. Hence, ride comfort will be an extra significant factor amongst customers when self-driving cars are launched and plays an essential role in marketing. The introduction of autonomous driving and the increased demand for ride comfort have sparked interest in this research area in this master thesis project. The main objective of this thesis is to develop robust dynamics control to improve ride comfort and guarantee an enhanced dynamic behavior. According to ISO international standard 2631, acceleration is defined as an objective performance indicator that reflects human perception and comfort. Therefore, the main goal is to optimize the car's movement to minimize the resulting acceleration acting on the passenger. It turned out that an active suspension system is an alternative to minimize acceleration. Introducing active actuators allows continuous application of external forces to the vehicle and improves ride comfort to a higher degree. However, controlling the active actuators is a challenging task. For the active actuators to generate the required forces, they must be regulated with the help of an advanced control algorithm. The entire vehicle is a complex system with six degrees of freedom and is constrained by many physical limitations. The model predictive controller is suggested in this thesis as a potential solution to regulate this complex MIMO system. A full vehicle model is implemented in SIMULINK, and the model predictive controller is developed in MATLAB with the help of an algebraic modeling language called YALMIP. To evaluate the performance of the MPC controlled active suspension different test scenarios with various road profile has been investigated. The results are compared to a vehicle with a passive suspension, and the improvement of ride comfort has been calculated with the weighted RMS acceleration method, suggested by ISO 2631. The results for different test scenarios indicate that the active system always reduces acceleration and can guarantee a more comfortable drive than the passive system. This thesis concludes that the active suspension system is a step forward to a more comfortable ride for our future autonomous cars.