CN-114624651-B - Array attitude angle compensation method for multi-microphone array sound source localization

CN114624651BCN 114624651 BCN114624651 BCN 114624651BCN-114624651-B

Abstract

The invention discloses an array attitude angle compensation method for multi-microphone array sound source positioning, which comprises the steps of firstly realizing the estimation of a sound source azimuth angle and a high-low angle by utilizing a Shan Maike wind array sound source direction finding method based on time delay estimation to obtain sound source direction vectors of a sound source to be tested relative to each microphone array, then measuring the array attitude angle of the array by utilizing attitude angle measuring equipment, correcting the array attitude angle by utilizing a direction cosine matrix formed by the array attitude angles, thereby converting the sound source direction vectors estimated by each microphone array into the same common horizontal coordinate system, and finally realizing the positioning of the sound source to be tested by utilizing an overall least square algorithm. The invention effectively solves the problem that when the sound source positioning is carried out based on the multi-microphone arrays, the sound source positioning error is increased because each microphone array cannot be erected completely horizontally, reduces the adverse effect of the array attitude angle on the sound source positioning, and improves the sound source positioning precision of the multi-microphone arrays.

Inventors

ZHAO ZHAO
ZHANG TENG

Assignees

南京理工大学

Dates

Publication Date: 20260505
Application Date: 20220314

Claims (4)

1. An array attitude angle compensation method for multi-microphone sound source localization is characterized by comprising the following steps: Step 1, a single microphone array sound source is used for direction finding, and a time delay estimation method is adopted for estimating and obtaining azimuth angles and high-low angles of a sound source to be detected, so that a sound source azimuth vector of the sound source to be detected relative to each microphone array is obtained; the method for estimating the azimuth vector of the sound source to be measured based on the single microphone array comprises the following steps: Step 1-1, detecting sound signals recorded by each channel of a single microphone array, and extracting a data part with a target sound signal; step 1-2, obtaining the arrival time difference between the arrival of the target sound source at each array element channel of the array by using a generalized cross-correlation mode, wherein the specific steps are as follows; Step 1-2-1, wherein signals received by two array elements of the array are x 1 (t) and x 2 (t) respectively, and the corresponding cross power spectral density functions are as follows ; Wherein "×" number represents complex conjugate; step 1-2-2, setting the weighting function adopted by generalized cross correlation as H (f), and setting the cross power spectral density function after weighting treatment as ; Step 1-2-3, according to wiener-Xin Qin theorem, knowing that the cross-correlation function of the signal and the cross-power spectral density thereof are fourier transform pairs, the generalized cross-correlation functions of x 1 (t) and x 2 (t) ; Steps 1-2-4, available from the cross-correlation principle, The corresponding τ value of the peak of (a) is the estimated value of the delay τ 12 between signal times x 1 (t) and x 2 (t); Step 1-2-5, repeating the steps 1-2-1 to 1-2-4 to obtain time delay estimation among all channels of the microphone array; Step 1-2-6, for a 5-element microphone array, obtaining a TDOA vector of the sound source according to the result of step 1-2-5 ; Step 1-3, solving a structural coefficient matrix of a microphone array, wherein the specific steps are as follows; Step 1-3-1, for a 5-element microphone array, the position coordinates of each array element are respectively represented by r i , i=1, 2,3,4,5; step 1-3-2, obtaining a structural coefficient matrix corresponding to the array ; Step 1-4, obtaining estimated values of azimuth angle and high-low angle of a sound source by using a sound source direction finding method based on time delay estimation; step 1-4-1, let the unit vector of the target sound source S with respect to the microphone array be expressed as ; Wherein θ is the azimuth angle of the sound source, and phi is the high-low angle of the sound source; step 1-4-2, obtaining a linear equation set according to a direction finding algorithm based on time delay estimation ; ; Wherein W is obtained in step 1-3, and τ is obtained in step 1-2; step 1-4-3 due to For overdetermined linear equation set, there is a unique linear least squares solution The method is obtained by the following formula, wherein W + is Moore-Penrose inverse matrix of array coefficient matrix W; ; Step 1-4-4, let W + = [u 1 ,u 2 ,u 3 ] T , wherein u 1 、u 2 、u 3 is a 10×1 column vector, thereby obtaining estimates of the azimuth and the altitude of the sound source respectively ; ; Step 1-5, obtaining an estimated value of the azimuth vector of the sound source by using the solving result of step 1-4; Step 2, correcting an array attitude angle, namely measuring the attitude angle of an array by using an array attitude angle measuring device, wherein the attitude angle comprises a yaw angle, a pitch angle and a roll angle, and then correcting the array attitude angle by using a direction cosine matrix, and converting the sound source azimuth vector estimated by each microphone array in the step1 into the same common horizontal coordinate system; for the sound source azimuth vector of the sound source obtained in the step 1 relative to each microphone array The array attitude angle correction is carried out, and specifically comprises the following steps: Step 2-1, measuring an array attitude angle by utilizing an attitude angle sensor, and respectively recording a yaw angle alpha, a pitch angle beta and a roll angle psi; Step 2-2, obtaining a direction cosine matrix corresponding to the array attitude angle The specific solving steps are as follows; step 2-2-1, according to the Euler rotation theorem, the directional cosine matrices corresponding to the yaw angle α, pitch angle β and roll angle ψ may be expressed as , , ; Step 2-2-2, namely completely rotating the array horizontally to a direction matrix corresponding to the non-horizontal state of the array Can be expressed as ; Step 2-2-3, because the array is always in rectangular coordinate system in the rotation process, the three direction cosine matrixes are unit orthogonal matrixes, so the direction matrixes Also in the form of unitary orthogonal matrices, i.e. satisfying the relationship ; Step 2-2-4, direction cosine matrix from carrier coordinate system b system to global horizontal coordinate system n system Can be expressed as ; Step 2-3, using the obtained direction cosine matrix For the sound source azimuth vector of the sound source obtained in the step 1 relative to each microphone array Correcting an array attitude angle, which specifically comprises the following steps of; step 2-3-1, setting the azimuth vector of the corresponding target sound source in the array completely horizontal state as ; Step 2-3-2, then And estimated in step 1 Satisfy the relation of ; By using the formula, the sound source azimuth vector estimated by each node of the system can be converted into a global public horizontal coordinate system; Step 3, positioning a multi-microphone array sound source, namely firstly measuring the position coordinates of each microphone array in a global coordinate system, combining the sound source azimuth vectors in the step 2, and realizing the positioning of the sound source to be measured by using a total least square method, wherein the specific implementation steps are as follows: Step 3-1, calculating central coordinates of each array by using a measuring device, wherein the central coordinates are respectively recorded as (x j ,y j ,z j ), j=1, 2..n, n is the number of the microphone arrays used, and the estimated sound source azimuth angles and the estimated altitude angles are respectively recorded as theta j and phi j , and j=1, 2..n; Step 3-2, assuming that the coordinates of the target sound source S are (x, y, z), the following relation can be established ; Abbreviated ms=q; Step 3-3, using the total least square method to obtain s an estimated value ; Wherein λ is the minimum eigenvalue of matrix C T C, C is the augmentation matrix [ M q ]; step 3-4, finally obtaining the estimated value of the three-dimensional coordinates of the sound source S 。
2. The method for compensating array attitude angle for multi-microphone sound source localization according to claim 1, wherein the solution in steps 1-4 is used to obtain an estimated value of azimuth vector of the sound source: 。
3. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the multi-microphone sound source localization oriented array attitude angle compensation method of any one of claims 1-2 when the program is executed by the processor.
4. A computer readable storage medium having stored thereon a computer program, which when executed by a processor implements the multi-microphone sound source localization oriented array attitude angle compensation method according to any one of claims 1-2.

Description

Array attitude angle compensation method for multi-microphone array sound source localization Technical Field The invention belongs to the technical field of sound source positioning, and particularly relates to an array attitude angle compensation method for multi-microphone array sound source positioning. Background The sound source positioning technology based on the microphone array is a research hot spot at home and abroad in recent years. The main principle is that a microphone array with a certain geometric topological structure is used for collecting sound source signals, and the sound signals are processed and analyzed through an array signal processing technology, so that the sound source position is determined. Because the passive sound positioning device has the characteristics of 360-degree detection, low cost, simple and convenient arrangement, no influence of terrain, electromagnetic interference and the like, the current sound source positioning technology is widely applied to various fields of army and civilian and the like, such as gun azimuth investigation, anti-sniper, audio/video conference systems, intelligent robots and the like. Most of traditional sound source localization technologies adopt a mode of single microphone sounding node networking or single microphone array, are affected by factors such as microphone spacing, actual sound velocity and the like, have limited system detection range and lower precision. The sound source positioning based on the multi-microphone array node is one of the research hot spots in recent years, and the characteristic of accurate angle measurement of the single-microphone array node is utilized to treat the single-microphone array node as an angle measurement sensor, so that the sound source positioning problem is converted into the positioning problem based on a plurality of arrival angles, and the detection distance and the positioning precision of the sound detection system are improved. The method comprises the steps of providing a multi-array combined processing algorithm based on data level fusion, improving the positioning accuracy of an indoor sound source, providing a multi-array combined processing algorithm based on data level fusion by using sound positioning research [ Yang Yichun, li Xiaodong, tian Jing, teng Pengxiao ] of multi-array data fusion, providing an air explosion point positioning system by using three-dimensional five-element arrays in the literature [ Wang Wei ], providing a three-dimensional explosion point positioning system by using a three-dimensional five-element array, and providing a positioning accuracy higher than that of a single node, providing a three-array fusion explosion point sound source testing system by using a multi-array fusion mode in the literature [ Lei, western An industrial university, 2021 ]. However, the existing research on realizing sound source positioning based on multiple microphones assumes that each microphone array is completely horizontal when sound source positioning is performed, and the arrays are difficult to be completely horizontal in the actual erection process, namely, the pitch angle and the roll angle of the arrays are non-zero, which leads to larger error of the final sound source positioning result, so that the attitude angle of the arrays must be corrected, and adverse effects of the array attitude on sound source positioning are reduced. Disclosure of Invention The invention aims to provide an array attitude angle compensation method in sound source positioning for multi-microphone array nodes, which solves the problem that the array nodes cannot be completely horizontal in the erection process, reduces the adverse effect of the array attitude angle on the sound source positioning and improves the sound source positioning precision. The technical scheme for realizing the aim of the invention is that the array attitude angle compensation method for multi-microphone array sound source positioning comprises the following steps: Step 1, a single microphone array sound source is used for direction finding, and a time delay estimation method is adopted for estimating and obtaining azimuth angles and high-low angles of a sound source to be detected, so that a sound source azimuth vector of the sound source to be detected relative to each microphone array is obtained; Step 2, correcting an array attitude angle, namely measuring the attitude angle of an array by using an array attitude angle measuring device, wherein the attitude angle comprises a yaw angle, a pitch angle and a roll angle, and then correcting the array attitude angle by using a direction cosine matrix, and converting the sound source azimuth vector estimated by each microphone array in the step1 into the same common horizontal coordinate system; And 3, positioning the sound source of the multi-microphone array, namely firstly measuring the position coordinates of each microphone array in a global coordinate syste