Adaptive neural network control based on nonlinear disturbance observer for BTT missile

For the BTT missiles with parameter perturbations and disturbance, an adaptive RBF neural network based on the nonlinear disturbance observer is proposed, using the ability of the RBF neural networks to approximate the arbitrary nonlinear function. The on-line adjust algorithm is designed to compose the adaptive RBF neural network with dynamical structure. The nonlinear disturbance observer is introduced to compensate the approximation error of neural network and decrease the influence to control system effectively, so this method is robust to the parameter perturbations and external disturbance. Simulation results indicate that the method has better tracking accuracy and robustness.