Model Reference Adaptive Neural Control for nonlinear systems based on Back-Propagation and Extreme Learning Machine

In this paper, a Model Reference Adaptive Neural Control (MRANC) that uses both off-line and online learning strategies and Single Hidden Layer Feedforward Networks (SLFNs) is proposed for a class of nonlinear systems. In the proposed scheme, one SLFN is used as the identifier to identify the unknown nonlinear system and then the other SLFN is used as the controller to construct the control law based on the information of the identified model. The neural-network parameters of the NNI and NNC are adapted off-line. The off-line trained neural controller ensures the stability and provides the necessary tracking performance. If there is a change in the system dynamics or characteristics, the trained neural identifier and controller are also adapted online for providing the appropriate control input to maintain the system's satisfactory tracking performance. Different from the existing technology where the Back-Propagation (BP) is employed to train the two SLFNs, the identifier is trained using a fast neural algorithm developed recently, namely Extreme Learning Machine (ELM) while the controller is trained using the Dynamic BP method. Simulation results show that the proposed approach has faster learning speed and higher tracking performance than the existing method.