Abstract
This article presents a prescribed performance optimal control method for flexible-joint (FJ) robots with input delay. First, an auxiliary system and the dynamic surface control method are utilized to construct coordinate transformations, which effectively mitigate the adverse effects of input delay on system stability. Then, to enhance the system's control performance, a prescribed-time prescribed performance method is proposed. This method not only improves tracking accuracy but also predetermines the error convergence time, significantly enhancing the system's transient and steady-state performance. Finally, by designing the update laws for the identifier, actor, and critic neural networks (NNs) using the revised term and prediction error, this article effectively resolves the optimal tracking problem for FJ robots through a simplified reinforcement learning algorithm. The effectiveness of the proposed scheme is validated through simulation results.