Ads-b signal target recognition method based on small sample machine learning model

Abstract

The invention discloses an ads-b signal target recognition method based on a small-sample machine learning model. Firstly, the ads-b signals of different targets are collected as training samples, and the data are preprocessed to construct a training set, a verification set and a test set. ; Secondly, build a small-sample machine learning model based on convolutional neural network and metric learning; then, use the training set to train the small-sample machine learning model, and use the verification set to verify the model during the training process. When the verification training reaches After setting the accuracy rate, solidify the model; finally, test the model, input the real-time ads-b signal, output the corresponding classification label, and finally get the category to which the target belongs. The method provided by the invention overcomes the shortcomings of the traditional machine learning method that requires a large number of labeled samples to train the model, has low implementation cost, has high real-time performance and high efficiency, and can save a lot of manpower and material resources.

Description

technical field [0001] The invention relates to the technical field of ads-b signal processing, in particular to an ads-b signal target recognition method based on a small-sample machine learning model. Background technique [0002] ads-b (ADS-B system is the abbreviation of Automatic Dependent Surveillance-Broadcast System) is a technical means applied to air traffic supervision, which has the advantages of high monitoring accuracy, fast information update rate, low cost of ground equipment construction and maintenance, etc. It can be used to provide application services such as ATC surveillance, airport surface surveillance and future air-to-air surveillance in non-radar coverage areas. Therefore, the signal processing technology of ads-b signal is an important research direction of air traffic control system. [0003] With the development of technology and the continuous improvement of computing power, classification and recognition methods based on artificial intelligenc...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is: the use of traditional machine learning classification and recognition algorithms is limited under the condition that sufficient samples cannot be provided for learning. The present invention provides an ads-b signal based on a small sample machine learning model that solves the above problems Target recognition method, this method can effectively avoid difficulties such as difficult collection of data samples and heavy labeling workload in engineering applications, and is efficient and practical, especially suitable for scenarios where sufficient samples cannot be provided for learning due to environmental constraints

Images