New strategy to perturbation estimation and constrained control of vehicle suspension system

Designing a reliable model-based controller for the active suspension system, considering physical constraints of the system, requires exact knowledge about model uncertainties and unknown road disturbance. This study presents a new strategy to upgrade a nominal dynamic model of the suspension system and the road profile into the actual system. In this strategy, complementary terms, calculated online from the information of sprung and unsprung mass displacements, are appended to the state space model of the system to compensate for uncertainties and disturbances. The accuracy of the constructed model is confirmed by experiments on a suspension platform with the MacPherson mechanism. Comparison with the unknown-input Kalman filter, addressed in the literature, highlights the greater performance of the suggested algorithm in estimating of a precise reliable dynamic model for the suspension system. Accordingly, a continuous predictive control method is developed for design of a nonlinear controller using the estimated dynamic model to achieve ride comfort and safety for the vehicle in the framework of suspension system constraints. The proposed control method firstly transforms the state constraints of tire deflection and suspension deflection to the equivalent input constraints and then finds the control law analytically by solving an optimization problem through the Karush-Kuhn-Tucker theorem. The constrained stability of the controller is analyzed by the Lyapunov method. Co-simulation results from Adams/Matlab demonstrate the superior performance of the proposed controller, which adapts itself to real conditions through accurate estimation of suspension dynamics. Additionally, according to the simulation time and the closed-form solution, it is fast and suitable for online implementation compared to other methods, such as the firefly algorithm and the nonlinear model predictive control method.