A Novel Robotic GWO LDI Modeling and Control for Nonlinear Systems

Abstract

This works aims to develop a new and improved GWO (Grey Wolf Optimizer), the so-called Robotic GWO (RGWO). First, to improve GWO's update formula position with an optimal learning strategy, we adapt the algorithm to real mobile environments, including robots, so that tracking robots can move prey toward targets. Then, the nonlinear active suspension (AS) control system is linearized by a neural network (NN) based linear differential inclusion (LDI) using feedback and feedforward linearization. In theory, it is found that the general SM (Sliding Mode) optimal control cannot provide sudden optimal results for the active linearized suspension system, so a method is proposed to improve the shortcomings of the active linearized suspension system. By constructing an extended SM-optimal manifold function, an improved SM-optimal controller is designed, which incorporates information on the entire structure and the expected performance of the suspension. For comparison purposes, the performance of three kinds of controls: SM optimal refinement control, logic-fuzzy SM control, and PS (passive suspension), shows the proposed controller's advantages . Finally, our improved SM optimal control for nonlinear AS systems, in general, can achieve the actual nominal optimal suspension performance, as confirmed by the simulation results. The results also show that the improved SM optimal control method provides better robustness even when the operating conditions or parameters of the structure vary.