Robust adaptive control of unknown plants using recurrent high order neural networks-application to mechanical systems

We extend our previous results on control of unknown dynamical systems using dynamic neural networks. The proposed algorithm is divided into two phases. First a recurrent high order neural network (RHONN) identifier is employed to perform “black box” identification and then a dynamic state feedback is developed to appropriately control the unknown system. Although the method is applicable to many classes of nonlinear systems, we concentrate our attention to the case where the unknown system is a mechanical system. This is since the control of mechanical systems is of great importance in many areas of engineering (e.g. robotics); moreover mechanical systems possess special properties that can be appropriately utilized in order to establish a very efficient identification and control scheme