Adaptive robust bank-to-turn missile autopilot design using neural networks

An adaptive robust neural-network-based control approach is proposed for bank-to-turn missile autopilot design. Feedforward neural networks with sigmoid hidden units are analyzed in detail for controller design. Without prior knowledge of the so-called optimal neural networks, we design a controller that exploits the advantages of both neural networks and robust adaptive control theory. For this scheme, a stable adaptive law is determined by using the Lyapunov theory, and the boundedness of all signals in the closed-loop system is guaranteed. No prior offline training phase is necessary, and only a single neural network is employed. It is shown that the tracking errors converge to a neighborhood of zero. Performance of the controller is demonstrated by means of simulations.