Fault-tolerant control of fixed wing UAVs with multiple control surfaces

An adaptive filtered backstepping method is proposed to solve the expected torque problem caused by various internal unmodeled dynamics, system uncertain parameters, and unknown input gains generated when the control surface of a multi-control surface UAV malfunctions. The fault-tolerant control method utilizes quadratic programming to redistribute control on the remaining control surfaces based on rudder effectiveness. In order to improve control accuracy, an RBF neural network is used to identify and compensate for unmodeled disturbances to solve the problem of high complexity of fault-tolerant control technology for existing UAV attitudes. Fault-tolerant control of UAV attitude is realized to enhance the survival rate when the control surface fails during UAV flight.