Reverse beamforming method based on Toeplitz improvement of uniform linear array

Abstract

The invention discloses a reverse beamforming method based on the Toeplitz improvement of a uniform linear array. The method comprises the following steps of first establishing an array signal mathematical model, and working out a cross-spectrum matrix for receiving linear array output signals by a cross-spectrum matrix definition; then setting a first array element of a linear receiving array as a reference element, and carrying out relevant operation on signals received by other array elements and a signal received by the reference element to obtain one group of array input relevant values rxx; using r*xx as lines and rxx to form a Toeplitz matrix to obtain a Toeplitz cross-spectrum matrix space distribution RT; and finally carrying out phase compensation summation on all the matrix elements in the cross-spectrum matrix space distribution RT to realize reverse beamforming. The method has the advantages that the great stability and reliability of linear spectrum estimation are kept, disturbances among the signals are effectively eliminated without losing the effective aperture of the array, the position estimation accuracy is high, and the calculation amount is small.

Description

technical field [0001] The invention relates to the technical field of sound source localization, in particular to an inverse beamforming method based on uniform linear array improved Toeplitzization. Background technique [0002] The spatial angular resolution of traditional beamforming algorithms mainly depends on the array aperture and signal-to-noise ratio. Once the array aperture is determined, its angular resolution limit (ie, the Rayleigh limit) is also determined. The spatial spectrum analysis algorithm based on subspace decomposition introduces the concept of characteristic subspace, breaks through the Rayleigh limit of angular resolution, and greatly improves angular resolution and estimation accuracy. However, the performance of this type of algorithm drops sharply under the conditions of strong correlation signals, small snapshots, and low SNR, and the accurate estimation of multi-source DOA needs to pre-estimate the number of spatially distributed sources in adv...

AI Technical Summary

Problems solved by technology

However, when there is a correlation between sources or between sources and noise, the cross-spectrum matrix will no longer be a Toeplitz matrix, and the mutual interference between targets cannot be effectively eliminated. In this case, the target spectrum estimation distortion is still relatively large, and the ideal high-resolution effect cannot be achieved.

[0003] The usual method is to use spatial smoothing technology to decoherent first. The main disadvantages of this method are: (1) multi-subarray spatial smoothing will reduce the effective aperture of the array and reduce the number of sources that can be estimated; (2) spatial smoothing The computational complexity of the technology is large, so the requirements for the realized system parameters are too large

The basic idea of ​​the Toeplitz approximation method is: average the diagonal elements of the covariance matrix, construct a Toeplitz matrix to approximate the array signal correlation matrix, and estimate the target orientation based on the constructed Toeplitz matrix, but the Toeplitz matrix construction is obviously a It is an approximate method, which has a large estimation deviation, and is a non-consistent orientation estimation. At the same time, the construction of the matrix also brings a large amount of calculation.

[0004] The essence of the traditional Toeplitzization method above is to average the elements on the diagonals of the lower triangular part of the covariance matrix of the received signal and replace the corresponding diagonal elements, which has the defect that the energy of the signal space leaks to the noise space, especially in In the case of coherent sources, there will be ambiguity in the spatial separation of signal and noise

Images