Neural net control of rigid robots

Model based controller design involves extensive, time consuming modelling and identification of the process. This problem might be overcome by a learning controller that is able to adapt to the process in order to optimise its performance. In this paper, a learning controller is used to control a rigid robot. The stability of the system is proved using Lyapunovs direct method. From the stability analysis rules have been obtained for the controller design. The theoretical results have been verified by simulation experiments. In the simulations the control system adapted to the process and remained stable.