Neural adaptive tracking controller for robot manipulators with unknown dynamics

A neural network (NN)-based adaptive control law is proposed for the tracki ng control of an n-link robot manipulator with unknown dynamic nonlineariti es. Basis-function-like networks are employed to approximate the plant nonl inearities, and the bound on the NN reconstruction error is assumed to be u nknown. The proposed NN-based adaptive control approach integrates the NN a pproach and an adaptive implementation of the discrete variable structure c ontrol, with a simple estimation mechanism for the upper bound on the NN re construction errors and an additional control input as a function of the es timate. Lyapunov stability theory is used to prove the uniform ultimate bou ndedness of the tracking error, and simulation results demonstrate the appl icability of the proposed method to achieve desired performance.