Nonzero-Sum Game-Based Modular Manipulator Optimal Tracking Control: Performance-Index Function Without Control Input

To address the issue that the performance-index function encompassing both dynamic and control processes in traditional adaptive dynamic programming (ADP) cannot be optimized as time approaches infinity, this paper proposes an optimal tracking control method for a modular manipulator based on a nonzero-sum game. The dynamic model of the modular manipulator is established using the Newton–Euler iterative method. By treating each module of the manipulator as a player in a nonzero-sum game and employing adaptive dynamic programming, the trajectory-tracking problem is transformed into an optimal-control problem. The critic neural network approximates the performance-index function without control input to derive the optimal tracking control policy. A stability theorem proves the stability of the closed-loop system, and an experimental platform validates the accuracy and optimality of the proposed method.