Self-triggering evolutionary optimal control of multi-modular robot manipulators

Abstract

Leveraging the unique flexibility, modular robot manipulators are typically suitable for complex operational tasks in extreme environments, which frequently require the cooperative operation of multi-modular robot manipulator. Herein, the load distribution policy for the heterogeneous cooperative operation tasks of multi-modular robot manipulator is proposed in this article, which allocates appropriate wrenches to multi-modular robot manipulator to achieve the desired motion of the manipulated object. Subsequently, a novel self-triggering optimal control method via evolution computing is proposed to address the optimal regulation problems of multi-modular robot manipulator. The evolution computing algorithm can search for a superior policy during policy improvement when calculating gradient information becomes infeasible or system dynamic is not differentiable, overcoming the limitations of gradient-dependent adaptive dynamic programming. The proof of convergence for the evolution computing algorithm further enhances the rigorousness of the evolution computing-based self-triggering optimal control method. Additionally, to reduce communication bandwidth, energy consumption, and computational load, a self-triggering control scheme is introduced into the controller, and an appropriate self-triggering condition is designed, which solely utilizes the current state of the system to determine the next triggering moment for the modular robot manipulators. Compared with traditional event-triggering control, self-triggering control does not require dedicated hardware to monitor whether triggering rules are violated. Hence, the introduction of self-triggering control significantly broadens the application scenarios for modular robot manipulators. Ultimately, the modular robot manipulator system is proven to be uniformly ultimately bounded with the Lyapunov theory. The visualization data of experimental results verifies the superiority of the evolution computing-based self-triggering optimal control method.