Research On Hierarchical Chassis Integrated Control Based On Model Predictive Control Method

To vehicle running at high speed,slight disturbance may lead to losing stability. The mechanical structure of the traditional vehicle has been perfect. Enhancing vehicle stability by improving the mechanical structure has limited space. Rapid development of Electronic technology and control theory provides new method to improving vehicle stability by active chassis control such as Anti-Lock Braking System, Electrical Stability Control, and Active Steering. The functions of chassis control systems ultimately are achieved by the tire force, but the vertical load of the tire and the ground shear force are in coupling. This leads to the coupling of various active chassis control systems. This thesis adopted model predictive control theory and hierarchical structure to research the chassis integrated control of braking, steering, and suspension. The main work of this thesis are as follows:1. To verify the effectiveness of the algorithm, a 15-dof nonlinear vehicle model together with Magic-formula tire model is built on the MATLAB/Simulink? platform. The model and Carsim? were compared in the lateral dynamic response, and the results show that the coincidence between the two models is well. So the model built in this theory can reflect the dynamic characteristics of the vehicle. In order to implement active brake control, reflecting the wheels locking condition, a braking model is built. Cooperation with Carsim? was made.2. The hierarchical integration control structure was adopted. The integrated control structure was divided into four layers: signal disposal layer, integrated control layer, control allocation layer, actuation layer. In this thesis, the integrated control layer and control allocation layer were studied. This thesis divided the integrated control layer into horizontal integration and global integrated. The global integrated involves the active suspension control, active steering and active braking. Horizontal integration only involves active steering and active braking. In the horizontal integration, model predictive control which has good adaptability of distortion is adopted, and the 2-dof linear vehicle with extra yaw moment and extra front steering angle as input is used to predict the future response of the vehicle. The global integration is based on the rules: To start active suspension control to improve yaw-rate response when the lateral acceleration beyond a certain limits. Suspension control adopts yaw-rate deviation as input, the control potential was analyzed based on a two degrees model.3. Control allocation layer solved the control allocation problem in hierarchical integrated control structure. The main task of the part is realizing the yaw moment by individual braking. After comparing several common linear allocation methods, we choose QP as the allocation method based on some reason as: flexible in target function setting and constraint processing. A comparison between single wheel brake methods and QP was implemented in stability control on low adhesion road. Results show that: the constrained optimization (QP) can guarantee the constraint as well as realized the allocation task, and expanded the scope of the system stability.The main innovation in this thesis includes:1) Adopt 2-dof vehicle model as prediction model, design the horizontal integration model predictive controller.2) Suspension control potential was studied based on 2-dof linear model.