The Problem of Stability in the Application of Neural Network to Continuous-Time Dynamic Systems - Intelligent Robots and Systems 95. 'Human Robot Interaction and Cooperative Robots', Proceedings.

Using neural network to identify a function in the dynamic equation brings about additional difficulties which are not generic in the other function approzimation problems. First, training samples can not be arbitrarily chosen due to hard nonlinearity, so are apt to be nonuniform over the region of interest. Second, the system may become unstable while attempting t o obtain the samples. This paper deals with these problems in continuoustime systems and suggest an effective solution, which provides stability and uniform sampling b y the virtue of supervisory controller. The supervisory control algorithm can be applied to the robot system dynamics. Of course the algorithm can be applied to the n-th order robot system, the simulation result will be given for the simple two link robot.