Research on performance magnetorheological semi-active suspension based on improved gray wolf algorithm-optimized linear quadratic regulator control

In this paper, an improved LQR control strategy based on the improved gray wolf algorithm (IGWA) is used on a quarter car model with a magnetorheological (MR) semi-active suspension (SAS) system. The proposed control algorithm is utilized to overcome the shortcoming that the weight matrix Q and matrix R determined by experience in the traditional LQR control method. The inverse model of a magnetorheological damper (MRD) is designed using a fuzzy neural network (FNN) approach to describe the relationship between the desired damping force and the control current. Simulation results demonstrate that the improved gray wolf optimization algorithm can not only enhance the balance between local and global search to improve the problem of local optimum but also shortens the optimization time and improves the search efficiency. The inverse model demonstrates excellent agreement between the actual and predicted current signals. Compared to traditional LQR control and passive suspension, the improved gray wolf algorithm-optimized LQR (IGWO-LQR) control reduces the root mean square (RMS) values of vehicle body acceleration by 21.41% and 35.38%, respectively, under A-Class road conditions and by 19.04% and 39.02%, respectively, under B-Class road conditions. It is clear that the proposed controller makes the vehicle body acceleration greatly improved compared with the passive suspension, and the control effect is better than the conventional LQR controller.