Direct adaptive neural network control of robots

Abstract

Neural network modelling of robots is introduced using the GL matrices and operator (Ge et al. 1994), and a new adaptive neural network controller for robots is presented. The controller is based on direct adaptive techniques, and there is no need for matrix inversion. Unlike many neural network controllers in the literature, inverse dynamical model evaluation is not required and no time-consuming training process is necessary, except for initializing the neural networks based on approximate parameters of the initial posture at time t = 0. It is shown that if gaussian radial basis function networks are used, uniformly stable adaptation is assured and asymptotic tracking is achieved.