Trajectory tracking fault-tolerant control of quadrotor based on neural network and full-state observer

In this paper, the trajectory tracking problem of a quadrotor is considered, and a control scheme is designed for position loop and attitude loop, respectively. For the position loop, neural networks are adopted to deal with the unknown interconnections and uncertainties of the subsystems, the designed controller guarantees that the quadrotor can follow the preset reference signals. For the attitude loop, Nussbaum functions are incorporated into the backstepping controller design process to eliminate the influences of unknown parameters and ensure the stability and performance of the quadrotor. Simultaneously, a novel state observer design is proposed for the quadrotor to estimate unmeasurable angular velocities, so as to realize the attitude tracking of the quadrotor. The proposed control scheme ensures that all the signals in the quadrotor are bounded and the tracking error of each subsystem converges to a small neighborhood around zero. Finally, the effectiveness of the proposed control scheme is further demonstrated by numerical simulation results.