A design method for adaptive inverse control using NARX neural networks

According to NARX dynamic network, a learning algorithm of improved RTRL is presented in this paper and applied to adaptive inverse control system, which consists of two NARX neural networks: one is applied to identify the controlled plant; the other approximates inverse transfer function of the plant. The online training method using NARX is also described in detail. Practical simulation results show NARX-based identifier and controller are feasible and the given algorithm is efficient in the application of adaptive inverse control (AIC).