Research On Fault Diagnosis Of Four-rotor Aircraft Sensor

A four-rotor aircraft with a cross-propeller or a quad-rotor helicopter.Each sensor of the aircraft is an important part of the measurement of flight position information.Its operation will affect the safety of the aircraft.Using fault diagnosis knowledge,it is possible to detect faults early,help avoid aircraft accidents,improve the safety and reliability of the aircraft,and the operational capability is of great significance.However,fault diagnosis of flight sensors is a very small typical small sample problem.The occurrence of sensor failures is somewhat random,often difficult to replicate,and the fault information that can be used is very limited.Therefore,this paper studies the fault diagnosis of small quadrotor aircraft sensors.The purpose of this problem is to study a four-rotor aircraft sensor fault diagnosis system based on gray model and Elman neural network.The main research contents of this paper are as follows:Firstly,the background and significance of the fault diagnosis of the four-rotor aircraft sensor are introduced.Then the mathematical model of the flight control system is studied,and the fault modeling of the aircraft attitude sensor is studied.Then the fault prediction of the quadrotor aircraft sensor is carried out by the gray model and the Elman neural network respectively.Secondly,the optimal weight coefficient in the combined model based on the Elman neural network and the gray model is determined by the principle of minimum variance,and finally the combined prediction model is obtained.Finally,the combined model is applied to the fault diagnosis of the sensor of the quadrotor.The simulation results verify that the combined prediction model has higher prediction accuracy for fault diagnosis delay and proves the effectiveness of the diagnosis method.