Research On Distributed Model Predictive Control For Active Suspension

The vertical,pitch,and roll motions generated by the vehicle will cause the occupants to move accordingly,which seriously affects the ride comfort and driving safety of the vehicle.So the research on the suppression of the vertical,pitch,and roll motions of the vehicle has a strong actual meaning.Taking the active suspension system as the research object,this paper proposes a multi-agent-based distributed coordinated control method for active suspension,and designs a Kalman filter vehicle vibration state observer.This method can solve the problems of multi-actuator coupling,engineering application computing power requirements,ride comfort,and driving safety requirements in the active suspension system.The following are primary research portions:(1)The control input and its functions are redefined in accordance with the operating principle of the active suspension system.The seven subsystems of the vehicle model are regarded as seven agents,and the active suspension dimensionality reduction model based on graph theory is established.(2)By considering the influence of the state of other adjacent agents on its own agents,a distributed coordination control method for active suspension based on multi-agents is proposed.Design individual model predictive control controllers for each agent.Through cooperation between each agent,the vertical vibration acceleration of the unsprung mass,the vertical acceleration at the center of mass of the vehicle body,the roll angular acceleration of the vehicle body,and the pitch angular acceleration of the vehicle body can follow their ideal values.(3)In the quadratic programming solution process of the rolling optimization part of the model predictive control,there will be a high-dimensional matrix inversion operation problem.This problem will make the method proposed in this paper difficult to implement in real-time engineering applications.Therefore,this paper proposes a fast optimization solution method for the i-th agent based on the RBF neural network.The model predictive control algorithm’s calculation performance can be increased by using this method to quickly find the rolling optimization solution.And this method effectively balances the performance index of the system and the computing power requirement of the algorithm in engineering applications.(4)This paper designs a Kalman filter vehicle vibration state observer to observe the state variables required in the distributed model predictive controller.The controller can improve the precision of the suspension system’s response to vehicle movement and solve the problem of high costs caused by installing sensors in the actual vehicle system.Through the joint simulation of CarSim and Matlab/Simulink,it is verified that the control method proposed in this paper can effectively suppress the vertical motion,pitch motion,and roll motion of the vehicle body.Especially in steering conditions,this method can simultaneously take into account the ride comfort,safety,and handling stability of the vehicle and also meet the computing power requirements of engineering applications for the algorithm.