Active interference identification method based on bilinear efficient neural network

Abstract

The invention belongs to the technical field of radar interference signal identification, and particularly relates to an active interference identification method based on a bilinear efficient neural network. The invention designs a more intelligent interference identification method to solve problems that the existing interference signal is difficult to identify under a low interference-to-noise ratio and depends on priori knowledge. According to the method, modeling analysis is carried out on various interference signals, the bilinear efficient neural network is adopted for recognition from the perspective of signal time-frequency images, and very high accuracy can still be obtained under the low interference-to-noise ratio. Simulation experiments prove that the bilinear efficient neural network is used for identifying the effectiveness of interference signals, and compared with a traditional mode of manually extracting features, the method is higher in precision and simpler and more convenient. According to the method, priori knowledge of interference signal characteristics is not needed, certain robustness is achieved under the low interference-to-noise ratio, and the application limitation of an existing radar active interference recognition method is broken through.

Description

technical field [0001] The invention belongs to the technical field of radar interference signal identification, and in particular relates to an active interference identification method based on a bilinear high-efficiency neural network. Background technique [0002] With the help of digital radio frequency memory (Digital Radio Frequency Memory, DRFM) to accurately copy and reproduce radar signals, modern active interference is more manifested in the form of coherent interference, which has a strong confusing effect on the identification of echo signals, significantly reducing It improves the working performance and detection efficiency of radar, and brings new challenges to the development of radar anti-jamming technology. [0003] The selection of anti-interference measures needs to be carried out on the basis of determining the type of interference. Traditional DRFM-based identification of radar active jamming types relies on accumulated work experience in advance, and...

AI Technical Summary

Problems solved by technology

[0004] Through the search of the existing technical literature, it was found that the "multi-type radar active interference perception new method" published in the "Journal of Xi'an Jiaotong University" (2019, 53(10): 103-108+121) by Liu Mingqian et al. The interference signal is subjected to variational mode decomposition, and the corresponding rectangular integral bispectrum and Renyi entropy of the calculated natural mode components are obtained. The interference type classification is obtained by means of random forest, but only the case where the interference-to-noise ratio is above 0dB is discussed. Insufficient analysis of model performance under low interference-to-noise ratio; Ruan Huailin et al. published in "Journal of Detection and Control" (2018,40(04):62-67) "Active deception jamming recognition based on stacked sparse autoencoder In ", the time-frequency analysis of the radar receiving signal is carried out, and the feature is reduced and sent to the autoencoder for feature extraction. Although there is a high recognition rate, the types of interference discussed are limited, and there is no deep mining of the network.

[0005] The existing literature search results show that radar active jamming is mainly identified by artificially extracting shallow features. Although some scopes can achieve certain results, the separability of most shallow features is poor, and researchers need rich prior knowledge. However, only a small amount of literature uses deep learning theory, but most of them are limited to the same type of interference signal, and the model has not been studied in depth

Therefore, a method for identifying active interference in a low-interference-to-noise ratio environment is proposed. The specific method is to convert the interference signal into a two-dimensional time-frequency image, and perform feature extraction, classification and identification of the interference signal through a bilinear high-efficiency neural network. Solved the technical problems of insufficient intelligence in the identification of existing radar jamming signals and low accuracy at low interference-to-noise ratios

Images