Fuzzy adaptive nonlinear MIMO control for rigid coupled multibody robots using reinforcement learning model

Robots and robotic arms are examples of mechatronic systems with numerous degrees of freedom (DOF), which are essential to modern industrial and daily living. However, it is difficult to accurately characterize these systems because of their strong coupling, unknown dynamics, and external perturbations, which makes typical model-based control methods impracticable. This study addresses this problem by presenting a reinforcement learning-based adaptive MIMO control approach that enhances the stability and flexibility of robot and robotic arm movements in the face of environmental disturbances and uncertainties. For quick and precise trajectory tracking, the control integrates the starfish optimization algorithm (SFOA), fuzzy reinforcement learning, and finite-time convergence. The primary innovation of the suggested control technique is the incorporation of dynamic adaptation, real-time learning, and instantaneous feedback into a multivariate feedback architecture, which offers higher real-time performance, accuracy, and robustness in crucial situations. We used joint space modeling to express the governing equations of the problem and applied quaternion modeling, the control strategy is demonstrated on two robots with two and five degrees of freedom and incomplete excitation. The simulation's outcomes demonstrate outstanding performance.