Unmanned aerial vehicle flight data abnormity detecting method based on LSTM

Abstract

An unmanned aerial vehicle flight data abnormity detecting method based on LSTM relates to the field of unmanned aerial vehicle detection and system health management. The method aims to settle problems of relatively complicated functions generated in a working process of an unmanned aerial vehicle system and incapability of satisfying a relatively high fitting precision requirement by the approaching capability in performing abnormity detection on flight data of the unmanned aerial vehicle. The method comprises the steps of reconstructing an unmanned aerial vehicle remote measurement data phase space, obtaining an input vector and an output vector, obtaining a training sample set and a testing sample set, building an LSTM basic predication model by means of a TensorFlow deep learning opensource framework and performing parameter optimization, obtaining an optimal LSTM model and furthermore calculating an LSTM predication result; afterwards performing abnormal point detection and abnormal sequence detection, and finally finishing unmanned aerial vehicle flight data abnormity detection.

Description

technical field [0001] The invention belongs to the fields of unmanned aerial vehicle abnormality detection and system health management. Background technique [0002] UAV anomaly detection methods can be mainly divided into three categories, knowledge-based anomaly detection methods, physical model-based anomaly detection methods and data-driven anomaly detection methods. The first two methods need to fully understand the domain knowledge or system structure to achieve better detection results, poor universality, poor anti-interference ability, and generally can only detect known types of anomaly detection. UAV telemetry data has few anomaly labels and complex anomaly types, and it is difficult to establish a unified physical model for its various subsystems. [0003] Data-driven anomaly detection methods belong to the category of multivariate statistical analysis, which can be subdivided into similarity-based methods, statistics-based methods, classification-based methods...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problem that the functions generated during the working process of the UAV system are relatively complex in the abnormal detection of UAV flight data, and its approximation ability cannot meet the requirements of higher fitting accuracy. Now it provides an LSTM-based Anomaly Detection Method of UAV Flight Data

Images