Electromagnetic signal spectrum sensing method based on deep clustering network

Abstract

The invention discloses an electromagnetic signal spectrum sensing method based on a deep clustering network, and the method comprises the following steps: S1, sorting and positioning samples according to an input signal, and carrying out the regularization processing of a radiation source signal; S2, clustering the input radiation source signals by adopting a deep continuous clustering network; S3, using the clustered radiation source signals, combining an input positioning result, and using a target tracking algorithm to realize front-back association of the signals at multiple moments to form a motion track of the electronic target; S4, according to the motion trails of the electronic targets, analyzing the motion laws among the multiple electronic targets, adopting a set updating algorithm and a platform assignment algorithm to carry out platform aggregation, and exploring a potential target platform. By inputting sorted and positioned signal samples, the system can automatically distinguish signals from different devices, and follow-up electronic target association and platform aggregation work can be carried out according to information of the devices.

Description

Technical field [0001] The present invention relates to the field of electromagnetic signal identification, and more particularly to a spectrum perception method based on a depth clustering network. Background technique [0002] Electromagnetic signal spectrum perception is the focus of electronic counterfeit fields. In modern information war, only the situation of the enemy radar signal can only be taken in time, and the corresponding response measures can be taken to give full play to our weapons and troops. [0003] In devices based on radiation source signals, conventional identification algorithms depends on manually maintained radar knowledge bases, after capturing signals, the radar signal in the knowledge base after capturing signals. However, this type of method depends to dominate manual intervention to a certain extent. And there is no need for new radar signals. Although the classic clustering algorithm can distinguish between unknown radar signals without prior art. ...

AI Technical Summary

Problems solved by technology

However, when the dimensionality of the input signal increases, calculating the distance on the sparse high-dimensional features will lead to the loss of information, so an additional feature selection process or dimensionality reduction is required.

However, the process of classical clustering is discrete and cannot be optimized together with the input coding part, which also leads to poor performance of the algorithm.

Images