Accelerometer fault diagnosis method based on convolutional neural network

Abstract

The invention discloses an accelerometer fault diagnosis method based on a convolutional neural network. The accelerometer fault diagnosis method comprises the following steps of 1, performing fault category modeling and feature extraction on an accelerometer; 2, building a CNN convolutional neural network model and performing training; and 3, testing the test sample by adopting the trained convolutional neural network model, and outputting a fault diagnosis result. Compared with the prior art, the method does not have any important original data or feature preprocessing, can eliminate the negative influence caused by the manual extraction of signal features, saves the selection time of a sign extraction method, directly converts the signal features into image pixel features, and is more sufficient in feature extraction. The fault diagnosis accuracy can be improved by utilizing the good image distinguishing capability of the convolutional neural network.

Description

technical field [0001] The invention relates to a fault diagnosis method of an accelerometer in an inertial navigation system. Background technique [0002] Accelerometers are widely used in inertial navigation systems, such as civil aviation aircraft, logistics and transportation drones, and automotive vehicle navigation. The accelerometer is used in the inertial navigation system to measure the acceleration parameters of the moving body, and its working status directly affects the performance of the inertial navigation system. During the operation of the accelerometer, there will be an error that changes slowly over time. When the error accumulates to a certain extent, it is prone to failure. Once a failure occurs, it will cause navigation and positioning errors, and may even lead to huge catastrophic events. The errors generated by the accelerometer include deterministic errors and random errors. Deterministic errors include: installation errors, scale factors, etc., whi...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problems of insufficient signal feature extraction and low accuracy of fault diagnosis in the process of accelerometer fault diagnosis, and provide a method for accelerometer fault diagnosis based on convolutional neural network, which converts the accelerometer output signal into gray Accelerometer image, using the end-to-end feature extraction method, input into the convolutional neural network for training and learning, for fault type identification and diagnosis, which can improve the accuracy of accelerometer fault diagnosis

Images