Research On Fault Diagnosis And Fault Tolerant Technology Of UAV Sensor

As the eyes of the UAV,the sensor provides comprehensive,reliable and accurate data information for the UAV flight control system.When the sensor fails,the UAV will not be able to guarantee the correct flight trajectory and attitude,so the fault diagnosis and fault-tolerant control technology of the UAV sensor has become an important guarantee for its safe flight.Based on the use of NAGA-BP network to establish the UAV attitude sensor observer model,this paper has carried out research on sensor fault diagnosis and fault-tolerant control technology.The main work of the article is as follows:(1)Introduced the attitude control system of small fixed-wing UAV and the fault types of sensors.Through the introduction of UAV flight attitude control system,the important role of gyroscope and accelerometer in attitude control process is discussed.After analyzing the working principle and common fault types of MEMS three-axis accelerometer and gyroscope,the common faults of the two sensors are modeled and classified,which lays the foundation for the fault diagnosis of the attitude sensor.(2)A fault diagnosis system based on NAGA-BP network observer is designed for nonlinear system faults.Because the traditional genetic algorithm has the problem that the effect is not obvious at the later stage of training,the optimization effect of BP neural network is not obvious enough,so the genetic algorithm crossover,mutation function and selection mechanism are improved,and the improved genetic algorithm is used to optimize the initial weight and threshold of the neural network.It avoids the problem of local optima and improves the convergence speed of the network.At the same time,NAGA-BP network is used to establish an observer model of the attitude sensor for fault diagnosis of the sensor.Simulation experiments prove that NAGA-BP network has better convergence ability,and the fault diagnosis system based on NAGA-BP network observer can accurately find fault signals.(3)Aiming at various faults of the attitude sensor,a soft closed-loop fault-tolerant control system based on NAGA-BP network observer is designed.When the sensor fails and is detected,the estimated value of the established observer model is used to replace the actual value of the sensor to participate in the system control to achieve the purpose of fault tolerance.At the same time,the SPRT method is used to improve the fault detection module,reduce the detection delay,improve the detectionaccuracy,and ensure that the fault-tolerant system can participate in control in a timely and effective manner.The effectiveness of the fault-tolerant control method is proved through simulation experiment.