Adaptive flight control law based on neural networks and dynamic inversion for unmanned aerial vehicle

For the small-sized UAV, the paper puts forward a nonlinear control method based on dynamic inversion. In order to compensate the dynamic inversion error, the paper designs an adaptive dynamic inversion controller with neural networks, which have excellent performance in nonlinear functional approximation. The pseudo-control error caused by actuator character is feedback to reference model through pseudo-control hedge, which eliminating the influences each other between the adaptive unit and actuators. The simulation result shows the improvement of the gradual convergence of tracking error and the robustness of system through adaptively compensating inversion error by neural networks.