Near-field source positioning method based on auto-encoder and parallel network

Abstract

The invention provides a near-field source positioning method based on an auto-encoder and a parallel network. The method comprises the steps of: generating near-field source data received by an array under the condition of a single information source, constructing an auto-encoding network and a parallel full-connection network, inputting the near-field source data into a combined network of the auto-encoding network and the parallel full-connection network, and when angle information is obtained, calculating out a single spectrum peak search mode. According to the method, the signal received by the array is subjected to subspace segmentation, and direct output of the DOA spatial spectrum under the condition of unknown signal source number is realized through the parallel full-connection network, so that the efficiency of the algorithm is greatly improved. The capability of multi-source positioning can be obtained only by training single-source data, so that the data volume of training is greatly reduced, and meanwhile, the training difficulty of the neural network is reduced to a certain extent. And an offline training and online testing process is adopted, so that the algorithm complexity in an actual use process is greatly reduced.

Description

technical field [0001] The invention relates to the field of signal processing, in particular to a parameter estimation method for array signal processing. Background technique [0002] Most of the traditional spatial spectrum estimation algorithms are proposed for the far-field signal model. In recent years, with the rise of near-field communication, array signal processing technology has been widely applied to the near-field signal model. Spectrum estimation algorithms cannot directly deal with near-field signal problems. Therefore, research on parameter estimation algorithms for near-field signal models has high practical value. [0003] However, traditional near-field algorithms are usually not well adapted to the defects of actual arrays, and there are still many problems to be solved in terms of algorithm complexity, estimation accuracy, and parameter matching. In recent years, some scholars have also proposed to use deep learning models to solve the problem of near-f...

AI Technical Summary

Problems solved by technology

[0002] Most of the traditional spatial spectrum estimation algorithms are proposed for the far-field signal model. In recent years, with the rise of near-field communication, array signal processing technology has been widely applied to the near-field signal model. Spectrum estimation algorithms cannot directly deal with near-field signal problems, therefore, research on parameter estimation algorithms for near-field signal models has high practical value

[0003] However, traditional near-field algorithms are usually not well adapted to the defects of actual arrays, and there are still many problems to be solved in terms of algorithm complexity, estimation accuracy, and parameter matching.

In recent years, some scholars have also proposed to use deep learning models to solve the problem of near-field source location. However, these algorithms require a huge amount of training data under the condition of multiple sources, and network training is difficult.

At the same time, most algorithms need to know the number of information sources a priori, but the number of information sources is often not available in practical application scenarios, which greatly limits the practical application of the algorithm

Images