Intelligence analytical redundancy-based fault-tolerance design for flight control system

To restrain modeling uncertainties and disturbance, a method for sensor failure diagnosis and accommodation is proposed, which is based on intelligence analytical redundancy. The analytical models of sensors are set up by using the Radio Basis Function neural network which is designed applying efficient algorithm of on-line learning and parameter optimization. According to the relationship between the sensors, residual is produced and analyzed for sensors fault isolation and signal reconfiguration. By the method, model uncertainties, noises and disturbance are restrained. The scheme is illustrated through simulations applying the flight control system of a fighter. Results show sensor online fault isolation and reconfiguration is achieved.