DOA estimation method and system based on self-selection neural network, storage medium and equipment

A self-selecting neural and convolutional neural network technology, applied to neural learning methods, biological neural network models, neural architectures, etc., can solve problems such as long time, slow convergence speed of DOA estimation, and low estimation accuracy, so as to save computing costs , Save computer performance, save labor time

Pending Publication Date: 2022-02-15
JIANGSU UNIV OF SCI & TECH
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Purpose of the invention: The present invention provides a DOA estimation method and system based on a self-selecting neural network, aiming to solve the problem that the neural network structure is limited by manual experience design when using a convolutional neural network for DOA estimation in the prior art. The convergence speed is slow, and the estimation accuracy is low and the time is long in the environment of a small number of array elements or a low signal-to-noise ratio

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • DOA estimation method and system based on self-selection neural network, storage medium and equipment
  • DOA estimation method and system based on self-selection neural network, storage medium and equipment
  • DOA estimation method and system based on self-selection neural network, storage medium and equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0067] like figure 1 , the invention discloses a DOA estimation method based on a self-selected neural network, comprising:

[0068] S1. Collect array received data when narrowband far-field signals are incident from different directions, calculate the covariance matrix of the array received data, convert the real part and imaginary part of the covariance matrix into two-dimensional matrices, and combine them into two-channel An array reception matrix; constructing a sample set with the two-channel array reception matrix and the incident direction as samples;

[0069] When the antenna array is a uniform linear array, the number of array elements is M, the distance between array elements is d, the number of snapshots is Z, the speed of the signal in the propagation medium is v, and the frequency of the signal source is ω 0 . The signal source of K narrow-band far-field signals sends a signal S k (t), direction is θ k , k=1,2,...,K, then the received data of the mth array el...

Embodiment 2

[0116] Table 1

[0117]

[0118] In this embodiment, the DOA estimation method disclosed in the present invention is verified through a test set. The experiment is based on the Tensorflow framework of python, and the SGD optimizer is used for training. The learning rate lr is 0.01, the decay is 0.00001, and the momentum is 0.9. When the number of array elements is M=10 and M=20, DOA estimation is carried out under the conditions of SNR 20dB, 10dB, 0dB and -10dB. When M=10, the input size of the convolutional neural network is 10*10*2, when M=20, the input size of the convolutional neural network is 20*20*2, and the batch_size is 32. Algorithm parameters are shown in Table 1.

[0119] The comparison of the algorithm effect under the condition of different array elements and different signal-to-noise ratios is shown in Table 2. In Table 2, Loss is the error function value of the test set samples, and Accuracy is the accuracy of DOA estimation results.

[0120] Table 2 Compa...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a DOA estimation method and system based on a self-selection neural network. The method comprises the steps: 1, collecting array receiving data when a narrowband far-field signal enters from different directions, calculating a covariance matrix, and generating a dual-channel array receiving matrix according to the covariance matrix; and constructing a sample set by taking the dual-channel array receiving matrix and the incident direction as samples; 2, generating N initial convolutional neural networks, wherein the input of each convolutional neural network is a dual-channel array receiving matrix, and the output of each convolutional neural network is a probability vector in the incident direction; 3, optimizing the N convolutional neural networks by adopting a particle swarm algorithm, and determining an optimal convolutional neural network; and 4, generating a receiving matrix according to the data received by the antenna array, inputting the matrix into the optimal convolutional neural network, and outputting a DOA estimation result. According to the method, the convolutional neural network does not need to be designed according to experience, and the convergence speed and the precision are high during DOA estimation.

Description

technical field [0001] The invention belongs to the technical field of antenna array signal processing, in particular to a method, system, storage medium and equipment for estimating the direction of arrival based on a self-selecting network. Background technique [0002] For decades, direction of arrival (DOA) estimation has been an important direction of array signal processing research, and it has been widely used in target detection, positioning and tracking, communication navigation, measurement and other fields. [0003] The original DOA estimation method cannot exceed the constraints of the Rayleigh limit, and the resolution is inherently limited. For example, the beamforming method represented by Bartlett; the maximum entropy method proposed by Burg et al.; the minimum variance method proposed by Capon; and the generalized cross-correlation time delay estimation method (TDOA) proposed by Carter and Knapp, etc. With the development of the times, DOA estimation has en...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06F30/27G06F30/25G06N3/04G06N3/08
CPCG06F30/27G06F30/25G06N3/08G06N3/045
Inventor 李鹏飞田雨波
Owner JIANGSU UNIV OF SCI & TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap