ADAPTIVE NEURAL-NETWORK CONTROL OF ROBOT MANIPULATORS IN TASK SPACE

In this paper, adaptive neural network control of robot manipulators i n the task space is considered. The controller is developed based on a neural network modeling technique which neither requires the evaluati on of inverse dynamical model nor the time-consuming training process, It is shown that, if Gaussian radial basis function networks are? use d, uniformly stable adaptation is assured, and asymptotically tracking is achieved, The controller thus obtained does not require the invers e of the jacobian matrix, In addition, robust control can be easily in corporated to suppress the neural network modeling errors and the boun ded disturbances, Numerical simulations are provided to show the effec tiveness of the approach.