Adaptive Incremental Nonlinear Dynamic Inversion Control Based on Neural Network for UAV Maneuver

Abstract

In this paper, a flight control strategy based on incremental nonlinear dynamic inversion (INDI) using properties of general flight control systems and nonlinear dynamic inversion by feeding back angular accelerations is presented. The INDI control law is based on a linearized approximation of incremental plant dynamics and reduces dependence on modeling accuracy. However, there is still an un-negligible nonlinear character of unmanned aerial vehicle (UAV) during high angle maneuvers. The main contributions of this article are 1) proposing an adaptive neural network compensation method : INDI with neural network(INDI_NN) to correctly consider the model uncertainties; 2) a quaternion based reference model which is suitable for vehicles to experience a high angle maneuver. The simulation results support the proposed control scheme in getting better tracking performance during large range of attitudes. Hence, the proposed control method makes INDI controller more practical and more suitable for high angle maneuver like Immelman Turn.