Adaptive Neural Control of A Flexible-Joint Manipulator via Recursive Sliding-Mode Dynamic Surface Approach

In this paper, a recursive adaptive neural tracking control method with nonlinear gains is proposed for a flexible-joint manipulator containing unknown dynamics. A recursive sliding-mode and a nonlinear gain function are introduced in the traditional dynamic surface control method, and radial basis function (RBF) neural networks (NNs) are employed to approximate the unknown functions. A new neural network weight law with σ-modified leakage term and an adaptive law are designed to estimate the bounds of approximation errors. It is proved that the control law can guarantee the ultimate boundness of all signals in the closed-loop system. Finally, the effectiveness of the proposed controller is verified by a simulation example.