Decentralized Robust Optimal Tracking Control for Robot Manipulators with State and Input Constraints
摘要
ADP (Adaptive Dynamic Programming) has been rapidly developed and has become an effective tool for solving robot manipulators’ robust optimal tracking control problem. However, a significant challenge of ADP is the control inputs and the state of the constrained system. Because in the early stage, high-amplitude excitation is needed to explore the system. This makes the system susceptible to constraint violations, causing the control performance to decrease or even become unstable. This paper proposes a decentralized robust optimal tracking control scheme for robot manipulators with state and input constraints, where the control input is asymmetrically constrained. First, a local feedforward control law is designed to transform the control problem for a robot manipulator into a control problem for an affine nonlinear system, in which the state constraints are solved, and the uncertainties from the physical connection effect between joints are eliminated. Then, a decentralized robust optimal control law is proposed based on ADP, where the input constraints are solved by introducing an asymmetrically constrained energy cost function. The proposed controller ensures fast convergence speed, the tracking errors are UUB (Uniformly Ultimately Bounded), and the cost function converges to a near-optimal value. Finally, the performance of the proposed controller is assessed through comparative simulations with existing approaches, specifically the decentralized optimal controller (DOC) and the robust optimal controller (ROC). The simulation results and performance analysis indicate that the proposed method achieves improvement rates of