Large-scale array fast adaptive anti-interference method based on convolutional neural network

Abstract

The invention provides a large-scale array fast adaptive anti-interference method based on a convolutional neural network, mainly solving the problems of large calculation amount and poor beam conformality in large-scale phased array adaptive beam forming in the prior art. The method in the implementation scheme comprises the following steps of constructing a phased array radar receiving signal model, a training sample set and a test sample set, establishing an array element weight prediction sub-network W to generate an array element weight, generating a directional diagram Q by using a directional diagram generator P according to the array element weight, and cascading W and P to form a directional diagram zeroing conformal network H, setting a directional diagram zeroing conformal network loss function L composed of a zeroing loss function L1 and a conformal loss function L2, training H by using the training sample set and adopting a gradient descent method, and inputting the test sample set into the trained weight prediction network to obtain a large-scale array anti-interference result. Due to the fact that the influence of interference null on the directional diagram is avoided, the calculation amount of the self-adaptive anti-interference algorithm is reduced, and the method can be used for target detection.

Description

technical field [0001] The invention belongs to the technical field of radar signal processing, in particular to a large-scale array fast self-adaptive anti-jamming method, which can be used for target detection. Background technique [0002] Radar signal processing is the process of analyzing and processing the radar echo signal to finally obtain useful information such as the distance and azimuth of the target. At present, phased array radar technology has been widely used in various types of radar, such as weather radar, bird detection radar, traffic control radar. [0003] Beamforming, a commonly used technology in phased array radar, refers to the method of forming a spatially directional beam by weighting and summing the output power of the array elements. Use the beamformed pattern for subsequent processing to obtain information such as target distance and azimuth. Therefore, the use of appropriate beamforming technology is the premise of obtaining the target range ...

AI Technical Summary

Problems solved by technology

In 2004, Suksmono introduced the complex neural network into the field of adaptive beamforming in the document Intelligent beamforming by using a complex-valued neural network, which improved the convergence speed of adaptive beamforming, but this method still needs iterative solution, adaptive weight output still takes too long

Many of the above documents use the radial basis neural network for beamforming, but this kind of neural network is not only complicated in the training process, but also requires additional clustering and other operations, and in order to avoid multi-layer networks, it also needs to know a lot of prior knowledge

In 2019, Li Jiaxin proposed the use of feed-forward neural network BP neural network for adaptive beamforming in the document Neural Network-based Main Lobe Interference Suppression Technology, and generated a pattern with deep nulls near the interference angle and other azimuth angles. However, this method requires prior knowledge of the number of interferers and the angle of interference before beamforming

[0006]First, sub-array division will affect the degree of freedom of the system

Too many sub-arrays will cause too few array elements in each sub-array, so the amount of data in each sub-array is too small, and the calculated anti-interference weight is inaccurate, resulting in widening of the main lobe beam and raising the side lobe. The shape of the pattern is not good;

[0007] Second, if the number of sub-arrays is too small, the number of array elements in each sub-array will be too large. When calculating the data covariance matrix, the calculation amount is too large to be able to Quickly calculate the adaptive anti-jamming weights in each sub-array

Images