Robust flight control system using neural networks: Adaptive dynamic surface design approach

In this paper, we propose an adaptive control method for micro aerial vehicle(MAV) flight system with model uncertainties. The proposed control system is constructed by the combination of the adaptive dynamic surface control(ADSC) technique and the self recurrent wavelet neural network(SRWNN). The ADSC technique which make the virtual controller using the first order filter provides us with the ability to overcome the explosion of complexity problems of the backstepping controller. The SRWNNs are used to observe the arbitrary model uncertainties of MAV flight system, and all their weights are trained on-line. From the Lyapunov stability theory, we derive the on-line tuning algorithms for all weights of SRWNNs and prove that all signals of a closed-loop system are uniformly ultimately bounded(UUB). Finally, we perform simulations to demonstrate the tracking performance and robustness of the proposed MAV control system during the pursuit guidance landing.