Fault diagnosis of an autopilot based on artificial neural network

It is well known that autopilot is one of important equipments of an aircraft in guaranteeing flight security. Its soft fault will arise with the use of autopilot. So it is very important to diagnose the soft fault as well as to make assessment to the future during in its daily maintenance. In this paper, artificial neural network is applied in fault diagnosis system of a helm timing system of an autopilot. A new fault diagnosis system of the helm timing system that includes hardware and software completed by language MATLAB has been designed finally through elaborating reasons causing soft fault of the system and fault effect, selecting BP network model which is trained and tested and adopting the improved BP algorithm so as to improve the network properties effectively. Artificial Neural Network-based technology for fault diagnosis of the helm timing system performs well. Fault diagnosis based on artificial neural network can be feasible and effective for all equipments.