OPTIMAL TRACKING CONTROL FOR ROBOT MANIPULATORS WITH ASYMMETRIC SATURATION TORQUES BASED ON REINFORCEMENT LEARNING

This paper introduces an optimal tracking controller for robot manipulators with asymmetrically saturated torques and partially - unknown dynamics based on a reinforcement learning method using neural network. Firstly, the feedforward control inputs are designed backstepping technique to convert problem into problem. Secondly, cost function of system input is defined, constrained Hamilton-Jacobi-Bellman equation built, which solved by online algorithm only single Then, asymmetric saturation rule determined. Additionally, concurrent used relax demand persistence excitation conditions. The built ensures that closed-loop asymptotically stable, approximation error uniformly ultimately bounded (UUB), converges near-optimal value. Finally, effectiveness proposed shown through comparative simulations.