Nested array direction-of-arrival angle estimation method based on variational Bayesian inference

Abstract

The present invention provides a nested array direction-of-arrival angle estimation method based on variational Bayesian inference. A three-layer prior distribution model is employed and variational inference is employed to find approximate distribution of posterior probability distribution to obtain an incident signal direction-of-arrival angle. A nested array formation structure is employed to perform direction-of-arrival angle estimation so as to overcome defects that a distinguishable target number is lower than an array element number when an uniform linear array is employed and improve the distinguishable target number of the array element number in the same condition, based on the sparse reconstruction theory, the posterior probability distribution of the sparse vectors to be estimated tends to be more gentle and better meet the signal sparse features, the direct solution of the posterior probability can be avoided, and therefore, the nested array direction-of-arrival angle estimation method is specially suitable for conditions with complex posterior probability and difficult solution and can reduce the operation complexity.

Description

technical field [0001] The invention relates to the technical field of signal processing, in particular to an array signal direction-of-arrival estimation method, which can be used to accurately estimate various parameters such as the signal direction-of-arrival angle and pitch angle, so as to obtain the position of a signal source. Background technique [0002] DOA estimation of the signal direction of arrival is an important content in the field of array signal processing. It uses an array of sensors arranged in a certain way to sense and receive the spatial acoustic signal, and then uses modern signal processing methods to quickly and accurately estimate the signal. Direction of arrival, pitch angle and other parameters. High-resolution DOA estimation has important application value in radar, sonar, wireless communication and other fields. [0003] At present, super-resolution DOA estimation methods can be divided into subspace algorithms and sparse reconstruction algori...

AI Technical Summary

Problems solved by technology

Among them, the convex optimization algorithm has a large computational complexity and poor estimation accuracy under low signal-to-noise ratio; the traditional DOA estimation algorithm based on the SBL criterion makes a Gauss-Gamma prior two-layer prior probability distribution assumption for the estimated sparse vector, and passes The expected maximum EM criterion obtains the optimal optimal estimate in an iterative manner, the algorithm complexity is high, and the convergence speed is slow

[0004] Most of the existing super-resolution DOA estimation algorithms focus on the use of uniform line arrays. An M-element uniform line array can resolve up to M-1 target signals, which cannot meet the estimation needs when the number of targets is large.

Images