Neural Network NARMAX Model Based Unmanned Aircraft Control Surface Reconfiguration

To guarantee the safety of an unmanned aircraft on the accident of the control surface failure, an algorithm of fault detection and surface reconfiguration is necessary. As the neural network can approximates arbitrary nonlinear functions, the neural network NARMAX model is able to capture the nonlinear dynamics. The proposed neural network NARMAX model based complementary filter provides an efficient and accurate diagnosis to the control surface failure. The simulation shows that the aircraft can quickly reconstruct the control surface and recover the safety with the help of the proposed algorithm when the ailerons fail.