Actuator Uncertainty Estimation And Optimal Control For Vehicle Magnetorheological Suspension System

The magnetorheological(MR)damper has attracted great attention and obtained considerable applications in the vehicle suspension systems due to its mechanical simplicity,high dynamic range,rapid response speed,low power requirements,large force capacity and robustness.The performance of a vehicle suspension system mainly depends on the substantial cognizance about the MR actuator dynamics and the corresponding control methodology.It is difficult to obtain an accurate model for MR damper due to its complex mechanical behavior.The traditional control of a MR damper of the vehicle suspension is an open-loop control based on the model of the MR damper.Thus,it is difficult to effectively reject the negative influence of the MR actuator uncertainties on the dynamic behavior of the system.Moreover,it is not clear about the questions:does the saturated version of the optimal control law derived from the corresponding un-constrained system,i.e.‘‘clipped-optimal’’,remain optimal for the constrained case as suggested in some previous publications,and what is the structure of the optimal control law?To answer these questions,it is essential to improve the reliability and evaluate the potential benefit of a semi-active control system.To solve these issues the major research in this dissertation are as follows:(1)To improve the reliability of the vehicle suspension system,a high order extended state observer(ESO)has been designed to estimate the MR actuator uncertainties.In the simulations,two cases that the temperature perturbations and the oil leakage of a MR damper are considered.Simulation results reveal that the high order ESO could estimate the arbitrary uncertainties.The comparisons have been made with the sliding mode observer(SMO)and the sliding mode disturbance observer(SMDO).The research results have established the theoretical foundation for the reliability improvement of the MR vehicle suspension system.(2)A constraint model for the control feasible region of a MR damper considering the hysteresis has been established and an essential distinction between the MR damper and the the semi-active actuators in some previous publications has been discussed.The constrained optimal control for vehicle suspension system based on two-point boundary value problem(TPBVP)with the established constrained model of the MR damper has been investigated compared with the the clipped-LQR.(3)An H_?adaptive optimal control based on policy iteration for the constrained semi-active vehicle suspension system with a MR damper has been presented based on the two kinds of objection function,respectively.Simulations demonstrate that the nonlinear optimal control of MR vehicle suspension outperforms greater than that of the clipped-LQR controlled.Therefore,the essential problem to the semi-active optimal is that the closed-loop dynamics become nonlinear and the LQR is not optimal anymore when the linear system is input constrained,and the optimal control of nonlinear continuous time systems is to solve either the nonlinear HJB or HJI partial differential equations as accurately and efficiently as possible(4)The experiment platform of a quarter vehicle suspension system with a MR damper based on Dspace has been set up.The necessity to consider the actuator uncertainties has been clarified by testing the uncertainties of the MR damper experimentally.The effectiveness of the ESO,SMDO and SMO to estimate the actuator uncertainties have been investigated based on the open-loop control and closed-loop control,respectively.