Radar waveform design method based on deep neural network

Abstract

The invention relates to a radar waveform design method based on a deep neural network. The method comprises the steps of setting four types of environment variables: targets, interference, clutters and noises; respectively designing radar waveforms through utilization of an MI (Mutual Information) criterion and an SINR (Signal to Interference plus Noise Ratio) criterion and obtaining frequency domain energy distribution of the radar waveforms; forming a training set and a testing set; designing a DNNs; training the DNNs through utilization of the training set; and generating the radar waveforms through utilization of the trained DNNs. Compared with an existing radar waveform design method, the method has the advantages that modes of generating signals through utilization of the MI criterion and the SINR criterion can be taken into consideration, radar resolution and frequency domain energy distribution precision are improved, and a great deal of formula derivation resulting from combination of different criteria is avoided.

Description

technical field [0001] The invention belongs to the field of radar signal processing, and in particular relates to a radar waveform design method based on a deep neural network, which can improve the resolution of the radar and the accuracy of energy distribution in the frequency domain. Background technique [0002] Radar obtains target information in space based on receiving reflected electromagnetic waves. The radar waveform is transmitted by the transmitter, and the receiver receives the waveform reflected by the target, and then discriminates and analyzes the characteristics of the target. Therefore, the radar waveform has an important impact on the performance of the radar. In particular, the cognitive radar that has emerged in recent years can flexibly design radar waveforms according to environmental characteristics, which improves the recognition and detection performance of radar targets. There are various radar waveform design methods, among which, based on signa...

AI Technical Summary

Problems solved by technology

However, so far, deep learning has been mostly used in automatic target recognition in the radar field, and has not been widely used in other fields

[0004] The radar waveform design method based on a single criterion will affect other indicators while improving a certain indicator, but cannot guarantee the overall performance; if two or more criteria are combined, it will be difficult to establish an objective function

Images