CN111273231A - Indoor sound source positioning method based on different microphone array topological structure analysis - Google Patents

Indoor sound source positioning method based on different microphone array topological structure analysis Download PDF

Info

Publication number
CN111273231A
CN111273231A CN202010206270.3A CN202010206270A CN111273231A CN 111273231 A CN111273231 A CN 111273231A CN 202010206270 A CN202010206270 A CN 202010206270A CN 111273231 A CN111273231 A CN 111273231A
Authority
CN
China
Prior art keywords
sound source
microphone array
positioning
microphone
array
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202010206270.3A
Other languages
Chinese (zh)
Inventor
孙昊彬
王玫
宋浠瑜
罗丽燕
周陬
仇洪冰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guilin University of Electronic Technology
Original Assignee
Guilin University of Electronic Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guilin University of Electronic Technology filed Critical Guilin University of Electronic Technology
Priority to CN202010206270.3A priority Critical patent/CN111273231A/en
Publication of CN111273231A publication Critical patent/CN111273231A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/18—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using ultrasonic, sonic, or infrasonic waves
    • G01S5/20—Position of source determined by a plurality of spaced direction-finders

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Circuit For Audible Band Transducer (AREA)

Abstract

The invention discloses an indoor sound source positioning method based on different microphone array topological structure analysis. The method can be used for positioning the sound source when the topological structure of the microphone array is changed, and analyzing the error of the sound source and comparing the error with other types of arrays. Meanwhile, a phase transformation weighting controllable response power positioning algorithm based on random region contraction is used, and a positioning result can be well obtained under the indoor high reverberation condition. A user can select a corresponding microphone array topological structure for analysis according to own requirements. After the microphone array meeting the self requirement is selected, the amplitude-frequency characteristic of the received signal can be calibrated by using an array calibration scheme combining multi-channel low-pass filtering and multi-channel adaptive filtering, and the positioning precision is improved.

Description

Indoor sound source positioning method based on different microphone array topological structure analysis
Technical Field
The invention relates to the field of indoor position service, in particular to an indoor sound source positioning method based on different microphone array topological structure analysis.
Background
Voice is one of the most effective ways for people to communicate information. In a communication system using digital audio technology, people use a microphone to collect voice signals, process or store the voice signals, and apply the signals to man-machine interaction, video conferencing, remote transmission, and the like. Is provided with
Figure 100002_DEST_PATH_IMAGE002
Is the distance of the sound source from the microphone array,
Figure 100002_DEST_PATH_IMAGE004
is the aperture of the microphone array and,
Figure 100002_DEST_PATH_IMAGE006
is the operating wavelength of the sound source, under near field conditions of the microphone, i.e. when
Figure 100002_DEST_PATH_IMAGE008
When the voice signal collected by the microphone is established, the voice signal can be considered as a signal without noise interference, and the voice quality is higher. However, in many cases this condition is not met, as in human-computer interaction, video conferencing, etc., the speaking person is typically in the far field of the array. Therefore, in the case of the far field of the array, strong environmental noise, reflected sound, directional interference, etc. are inevitably mixed in the microphone received signal, so that the quality of the picked-up voice signal is greatly reduced.
The single-channel speech cannot achieve accurate sound source localization and tracking, so the algorithm for sound source localization and tracking is generally directed to multi-channel speech. There are three types of commonly used multi-channel sound source localization techniques:
1. the method based on the time difference of arrival is characterized in that under the conditions of low signal-to-noise ratio and strong reverberation, the existing time delay estimation method has large errors, and in addition, the positioning method is only suitable for a single sound source and is difficult to be used for positioning multiple sound sources;
2. the method based on high-resolution spectrum estimation can realize orientation, has poor precision on precise positioning, and has strict requirements on received acoustic signals, so the method is difficult to be used in an actual voice sound source positioning system;
3. the method of controllable beam forming based on maximum output power becomes one of the most popular sound source positioning algorithms at present, and the algorithm has good robustness under high reverberation and high positioning accuracy. In addition, the single-channel speech enhancement method is difficult to suppress directional interference and perform noise reduction processing, so that a far-field beam forming method is required to be adopted for multi-channel speech enhancement and processing, different microphone array topologies are considered, and the spatial filtering effect of the array is improved.
According to the array signal processing theory, the optimal arrangement of the array elements has important influence on the performance of the array processing system. Microphone array topologies can be divided into three categories: one-dimensional arrays (such as linear arrays such as nested linear arrays and equidistant linear arrays), two-dimensional arrays (such as planar arrays such as circular arrays and square arrays), and three-dimensional arrays (such as three-dimensional arrays such as star arrays and spherical arrays). When the array topology is different, for example, the dimension of the array, the number of the array elements, and the array element spacing all affect the positioning accuracy and the operation speed of the microphone array positioning algorithm. In the actual space positioning process, the one-dimensional and two-dimensional array positioning effects are not good, so that the research on a reasonable three-dimensional array topological structure has practical significance.
Currently, indoor mobile sound source localization studies based on microphone arrays are all performed under the assumption that the frequency response of the received signals of the microphone arrays maintains high consistency. However, in an actual test, due to the fact that the manufacturing of the microphone has tolerance, and is influenced by a plurality of composite factors such as the using time and the indoor complex environment, the frequency response characteristic of the received signal of the microphone array has large deviation from a theoretical value, and the positioning accuracy is reduced due to the deviation, so that the calibration of the frequency response of the microphone array is of great significance for further improving the positioning accuracy of the indoor mobile sound source.
Disclosure of Invention
The invention aims to solve the defects and provide an indoor sound source positioning method based on different microphone array topological structure analyses. The method utilizes microphone arrays with different topological structures to obtain voice signals, carries out a positioning algorithm based on phase transformation weighting and controllable corresponding power, and adds an optimization algorithm of random area contraction, analyzes the performance of each microphone array from a positioning result, and provides an array calibration scheme combining multi-channel low-pass filtering and multi-channel adaptive filtering aiming at the actual problem that the sound source positioning capability is insufficient under the condition that interference exists between the microphone arrays and speakers (sound sources) in a real indoor environment, so that the positioning result has higher precision.
The technical scheme for realizing the purpose of the invention is as follows:
an indoor sound source positioning method based on different microphone array topological structure analysis comprises the following steps:
(1) the method comprises the steps that an indoor sound source positioning system of a microphone array is arranged, and the system consists of a microphone array topological structure analysis module, an array self-adaptive filtering correction module, a sound source positioning algorithm and an analysis module which are sequentially connected;
(2) extracting a voice signal: arranging proper microphone arrays indoors, enabling speakers to sound, recording voices of the speakers, and extracting audio signals corresponding to each microphone
Figure 100002_DEST_PATH_IMAGE010
、
Figure 100002_DEST_PATH_IMAGE012
……
Figure 100002_DEST_PATH_IMAGE014
；
(3) Dividing sound source space into multiple grids, and calculating power of each point on the grids in turn
Figure 100002_DEST_PATH_IMAGE016
(
Figure 100002_DEST_PATH_IMAGE018
The point of maximum power being soundPoints of source localization
Figure 100002_DEST_PATH_IMAGE020
=
Figure 100002_DEST_PATH_IMAGE022
(
Figure 510418DEST_PATH_IMAGE018
；
(4) Total power of any one point
Figure 427559DEST_PATH_IMAGE016
(
Figure 100002_DEST_PATH_IMAGE024
) Performing generalized cross-correlation based on phase transformation on signals of all microphone pairs on a microphone array pairwise and summing:
Figure 275472DEST_PATH_IMAGE016
(
Figure 732998DEST_PATH_IMAGE024
)=
Figure 100002_DEST_PATH_IMAGE026
wherein k and l represent the kth and l microphones,
Figure 100002_DEST_PATH_IMAGE028
weights representing phase transformations, τ, (
Figure 100002_DEST_PATH_IMAGE030
) Represents the time from the sound arriving at the kth microphone from position X; in the formula (II)
Figure 609687DEST_PATH_IMAGE028
Defined as a combined weighting function
Figure 100002_DEST_PATH_IMAGE032
：
Figure 100002_DEST_PATH_IMAGE034
Taking into account the calculation
Figure 556784DEST_PATH_IMAGE016
(
Figure 823817DEST_PATH_IMAGE024
) Symmetry involved, and removing some fixed energy terms, then
Figure 22717DEST_PATH_IMAGE016
(
Figure 753913DEST_PATH_IMAGE024
) Moiety varying with x
Figure 100002_DEST_PATH_IMAGE036
Comprises the following steps:
Figure 809593DEST_PATH_IMAGE036
=
Figure 100002_DEST_PATH_IMAGE038
(5) global search is carried out in the whole room, and a coordinate point Y with the maximum energy is obtained by utilizing a random region contraction algorithm (SRC); randomly finding out an N-dimensional matrix in a given initial value, gradually reducing the range in the sequential process until reaching a small enough range, and finding out a peak value; thereby calculating a positioning coordinate point.
In step (4), to simplify the calculation
Figure 563923DEST_PATH_IMAGE020
Can be replaced by:
Figure 100002_DEST_PATH_IMAGE040
=
Figure 100002_DEST_PATH_IMAGE042
in step (5), the process of the random region shrinking algorithm is as follows:
1) i is defined as the number of iterations,
Figure 100002_DEST_PATH_IMAGE044
representing the number of points randomly drawn at the ith iteration,
Figure 100002_DEST_PATH_IMAGE046
representing the number of points contained in the sub-search space of the next generation,
Figure 100002_DEST_PATH_IMAGE048
representing the next generation sub-search space. Define once per calculation
Figure 425568DEST_PATH_IMAGE036
Is just recorded as one
Figure 100002_DEST_PATH_IMAGE050
，
Figure 100002_DEST_PATH_IMAGE052
After the ith iteration
Figure 11270DEST_PATH_IMAGE050
The number of times of the operation of the motor,
Figure 100002_DEST_PATH_IMAGE054
denotes the stop value,. phi.
Figure 503432DEST_PATH_IMAGE050
The maximum number of allowed calculations.
Figure 100002_DEST_PATH_IMAGE056
Representing a new sub-search space
Figure 807374DEST_PATH_IMAGE048
The boundary of (2);
2) the number of initialization iterations i = 0;
3) setting initial parameters:
Figure 100002_DEST_PATH_IMAGE058
、
Figure 100002_DEST_PATH_IMAGE060
，
Figure 100002_DEST_PATH_IMAGE062
；
4) computing
Figure 144814DEST_PATH_IMAGE044
All of
Figure 850602DEST_PATH_IMAGE036
A value of (d);
5) arrange out
Figure 185769DEST_PATH_IMAGE046
So that
Figure 977007DEST_PATH_IMAGE046
≪
Figure 55822DEST_PATH_IMAGE044
；
6) According to
Figure 553799DEST_PATH_IMAGE046
Shrinking the current search space and updating the search space
Figure 387763DEST_PATH_IMAGE048
And new region boundaries
Figure 338401DEST_PATH_IMAGE056
；
7) If it is not
Figure 100002_DEST_PATH_IMAGE064
Or is or
Figure 100002_DEST_PATH_IMAGE066
And is
Figure 100002_DEST_PATH_IMAGE068
Determining the coordinate position of the point, storing the result and outputting the result;
8) if only
Figure 363119DEST_PATH_IMAGE066
If yes, the result is discarded;
9) in that
Figure 715603DEST_PATH_IMAGE046
To find a subset
Figure 100002_DEST_PATH_IMAGE070
So that
Figure 720468DEST_PATH_IMAGE070
Is greater than
Figure 158402DEST_PATH_IMAGE046
Average value of (d);
10) repeating the step 3) and the step 4) in the current search space
Figure 579019DEST_PATH_IMAGE048
In the random selection
Figure DEST_PATH_IMAGE072
Points, calculating their corresponding
Figure 113906DEST_PATH_IMAGE036
A value of (d);
11) will be provided with
Figure 227355DEST_PATH_IMAGE070
Put the points in the subset
Figure DEST_PATH_IMAGE074
And selecting
Figure 214903DEST_PATH_IMAGE072
In
Figure 439211DEST_PATH_IMAGE036
Of greatest value
Figure DEST_PATH_IMAGE076
Point put in subset
Figure 563025DEST_PATH_IMAGE074
In and store
Figure 909692DEST_PATH_IMAGE074
Collecting and putting the data into the next iteration for use;
12) order to
Figure DEST_PATH_IMAGE078
And performing the next iteration and returning to the step 5).
The invention has the advantages that: the invention provides a set of indoor sound source positioning method based on different microphone array topological structure analysis and an array calibration scheme integrating multi-channel low-pass filtering and multi-channel adaptive filtering. The method can be used for positioning the sound source when the topological structure of the microphone array is changed, and analyzing the error of the sound source and comparing the error with other types of arrays. Meanwhile, a phase transformation weighting controllable response power positioning algorithm based on random region contraction is used, and a positioning result can be well obtained under the indoor high reverberation condition. A user can select a corresponding microphone array topological structure for analysis according to own requirements. After the microphone array meeting the self requirement is selected, the amplitude-frequency characteristic of the received signal can be calibrated by using an array calibration scheme combining multi-channel low-pass filtering and multi-channel adaptive filtering, and the positioning precision is improved.
Drawings
FIG. 1 is a flow chart of microphone array indoor speaker localization according to an embodiment of the present invention;
FIG. 2 is a diagram of the positioning effect of different microphone array types according to the embodiment of the present invention;
FIG. 3 is a diagram illustrating the positioning effect of different array element spacings of an array according to an embodiment of the present invention;
FIG. 4 shows the positioning error and the calculated amount when the three-dimensional orthogonal array element spacing is 10cm according to the embodiment of the present invention;
FIG. 5 is a schematic diagram of an SRP-PHAT positioning system based on the fusion of multi-channel low-pass filtering and multi-channel adaptive filtering according to an embodiment of the present invention;
FIG. 6 is a graph comparing microphone frequency responses before filtering according to an embodiment of the present invention;
FIG. 7 is a graph comparing the frequency response of a microphone after filtering according to an embodiment of the present invention.
Detailed Description
The invention is further illustrated with reference to the following figures and examples.
Example (b):
as shown in fig. 1, an indoor sound source localization method based on different microphone array topological structure analysis is to set a microphone array indoor speaker localization system, which is composed of three modules: the system comprises a microphone array topological structure analysis module, an array self-adaptive filtering correction module and a speaker positioning algorithm module.
(1) Microphone array topology analysis module:
in order to explore the influence of different array topologies on the positioning result, the present example adopts a controlled variable method to the microphone array: and carrying out variable adjustment on array dimension, array element interval and array element number to form microphone arrays with different topological structures. The three different topological structure arrays of the one-dimensional linear array, the two-dimensional T-shaped array and the three-dimensional orthogonal array are subjected to development analysis, and error analysis shown in figure 2 shows that the topological structure of the three-dimensional orthogonal array has better positioning performance than the other two arrays, and the optimal selection of the number of array elements under the array is shown. Fig. 3 is an analysis of the array element spacing with a determined number of array elements in the array dimension, the graphical result reflecting the optimal choice of array element spacing. Fig. 4 reflects the accuracy and calculated quantity curve of positioning with the determined array dimension, array element spacing and array element number after the above analysis.
(2) The array adaptive filtering correction module:
the array calibration scheme (as shown in fig. 5) combining the multi-channel low-pass filtering and the multi-channel adaptive filtering, which is provided by the present example, is used as an intermediate module connecting the microphone array topology analysis module and the speaker positioning algorithm module, and can correct the array elements on the determined array type, thereby improving the positioning accuracy.
(3) Speaker location algorithm module:
the module calculates the controllable response power of the received signal using phase transformation weighting. And searching a coordinate enabling the controllable response power to reach the maximum in a preset sound source space, and obtaining the position estimation of the real sound source. The voice signals are directly obtained by the microphone array and then separated to obtain multi-path single microphone voice signals. Because the process of searching the maximum value of the power is too large in calculation amount, the system searches the peak value by using a random region contraction optimization algorithm. And comparing the obtained positioning coordinates with the real coordinates, and analyzing the performances of different microphone arrays through the comparison of the errors. The method comprises the following specific steps:
1. extracting voice signals, arranging a proper microphone array indoors, enabling a speaker to sound, recording the voice of the speaker, and extracting audio signals corresponding to each microphone
Figure 322219DEST_PATH_IMAGE010
、
Figure 412535DEST_PATH_IMAGE012
……
Figure 594118DEST_PATH_IMAGE014
。
2. The principle of the controllable response power positioning algorithm is to divide the sound source space into a plurality of grids and calculate the power of each point on the grids in turn
Figure 783790DEST_PATH_IMAGE016
(
Figure 745930DEST_PATH_IMAGE018
The point of maximum power being the sound source locationDot
Figure 312041DEST_PATH_IMAGE020
=
Figure 613709DEST_PATH_IMAGE022
(
Figure 36600DEST_PATH_IMAGE018
。
3. Total power of any one point
Figure 423719DEST_PATH_IMAGE016
(
Figure 855838DEST_PATH_IMAGE024
) It can be considered that signals of all microphone pairs in the microphone array are subjected to generalized cross-correlation based on phase transformation and summed two by two:
Figure 746433DEST_PATH_IMAGE016
(
Figure 543488DEST_PATH_IMAGE024
)=
Figure 214641DEST_PATH_IMAGE026
wherein k and l represent the kth and l microphones,
Figure 388133DEST_PATH_IMAGE028
weights representing phase transformations, τ, (
Figure 398814DEST_PATH_IMAGE030
) Representing the time from the sound arriving at the kth microphone from position X. In the formula (II)
Figure 429087DEST_PATH_IMAGE028
Defined as a combined weighting function
Figure 525219DEST_PATH_IMAGE032
：
Figure DEST_PATH_IMAGE034A
Taking into account the calculation
Figure 299140DEST_PATH_IMAGE016
(
Figure 226645DEST_PATH_IMAGE024
) Symmetry involved, and removing some fixed energy terms, then
Figure 365502DEST_PATH_IMAGE016
(
Figure 948930DEST_PATH_IMAGE024
) Moiety varying with x
Figure 260963DEST_PATH_IMAGE036
Comprises the following steps:
Figure 246237DEST_PATH_IMAGE036
=
Figure 624171DEST_PATH_IMAGE038
thus, to simplify the calculation
Figure 694896DEST_PATH_IMAGE020
Can be replaced by:
Figure 748302DEST_PATH_IMAGE040
=
Figure DEST_PATH_IMAGE080
4. and carrying out global search in the whole room, and obtaining a coordinate point Y with the maximum energy by utilizing a random region contraction algorithm (SRC). The basic idea of the random area shrinkage algorithm is to find an N-dimensional matrix at random in a given initial value, and gradually narrow the range in a sequential process until a sufficiently small range is reached to find the peak. Thereby calculating a positioning coordinate point.
The process of the random area shrink algorithm is as follows:
1) i is defined as the number of iterations,
Figure 650399DEST_PATH_IMAGE044
representing the number of points randomly drawn at the ith iteration,
Figure 131059DEST_PATH_IMAGE046
representing the number of points contained in the sub-search space of the next generation,
Figure 485817DEST_PATH_IMAGE048
representing the next generation sub-search space. Define once per calculation
Figure 342915DEST_PATH_IMAGE036
Is just recorded as one
Figure 37201DEST_PATH_IMAGE050
，
Figure 751079DEST_PATH_IMAGE052
After the ith iteration
Figure 796396DEST_PATH_IMAGE050
The number of times of the operation of the motor,
Figure 253922DEST_PATH_IMAGE054
denotes the stop value,. phi.
Figure 802715DEST_PATH_IMAGE050
The maximum number of allowed calculations.
Figure 890756DEST_PATH_IMAGE056
Representing a new sub-search space
Figure 220107DEST_PATH_IMAGE048
The boundary of (2);
2) the number of initialization iterations i = 0;
3) setting initial parameters:
Figure 419007DEST_PATH_IMAGE058
、
Figure 822306DEST_PATH_IMAGE060
，
Figure 143566DEST_PATH_IMAGE062
；
4) computing
Figure 897896DEST_PATH_IMAGE044
All of
Figure 634907DEST_PATH_IMAGE036
A value of (d);
5) arrange out
Figure 220610DEST_PATH_IMAGE046
So that
Figure 650454DEST_PATH_IMAGE046
≪
Figure 688817DEST_PATH_IMAGE044
；
6) According to
Figure 229520DEST_PATH_IMAGE046
Shrinking the current search space and updating the search space
Figure 872991DEST_PATH_IMAGE048
And new region boundaries
Figure 270474DEST_PATH_IMAGE056
；
7) If it is not
Figure 999396DEST_PATH_IMAGE064
Or is or
Figure 78210DEST_PATH_IMAGE066
And is
Figure 638504DEST_PATH_IMAGE068
Determining the coordinate position of the point, storing the result and outputting the result;
8) if only
Figure 144572DEST_PATH_IMAGE066
If yes, the result is discarded;
9) in that
Figure 360790DEST_PATH_IMAGE046
To find a subset
Figure 305612DEST_PATH_IMAGE070
So that
Figure 658096DEST_PATH_IMAGE070
Is greater than
Figure 662961DEST_PATH_IMAGE046
Average value of (d);
10) similar to steps 3) and 4), in the current search space
Figure 100896DEST_PATH_IMAGE048
In the random selection
Figure 521513DEST_PATH_IMAGE072
Points, calculating their corresponding
Figure 56399DEST_PATH_IMAGE036
A value of (d);
11) will be provided with
Figure 904270DEST_PATH_IMAGE070
Put the points in the subset
Figure 829500DEST_PATH_IMAGE074
And selecting
Figure 116125DEST_PATH_IMAGE072
In
Figure 443201DEST_PATH_IMAGE036
Of greatest value
Figure 727552DEST_PATH_IMAGE076
Point put in subset
Figure 202396DEST_PATH_IMAGE074
In and store
Figure 964815DEST_PATH_IMAGE074
Collecting and putting the data into the next iteration for use;
12) order to
Figure 202856DEST_PATH_IMAGE078
And performing the next iteration and returning to the step 5).
By combining fig. 2, 3 and 4, a microphone array suitable for the actual scene of the user can be selected according to different positioning accuracy requirements, different microphone number requirements and different array sizes. As shown in fig. 5, when the voice of the speaker is transmitted through the acoustic channel generated in the indoor environment, the voice is received by the preamplifier of the microphone array, the received microphone signals are subjected to fusion filtering based on multi-channel low-pass filtering and multi-channel adaptive filtering, noise except the voice signal of the speaker is filtered by the low-pass filter, the amplitude-frequency characteristic of the received signal is calibrated by the adaptive filter, and the amplitude-frequency characteristic before and after calibration is shown in fig. 6 and 7, so that the positioning effect is more accurate.

Claims (3)

1. An indoor sound source positioning method based on different microphone array topological structure analysis is characterized in that: the method comprises the following steps:
(1) the method comprises the steps that an indoor sound source positioning system of a microphone array is arranged, and the system consists of a microphone array topological structure analysis module, an array self-adaptive filtering correction module, a sound source positioning algorithm and an analysis module which are sequentially connected;
(2) extracting a voice signal: arranging proper microphone arrays indoors, enabling speakers to sound, recording voices of the speakers, and extracting audio signals corresponding to each microphone
Figure DEST_PATH_IMAGE002
、
Figure DEST_PATH_IMAGE004
……
Figure DEST_PATH_IMAGE006
；
(3) Dividing sound source space into multiple grids, and calculating power of each point on the grids in turn
Figure DEST_PATH_IMAGE008
(
Figure DEST_PATH_IMAGE010
The point with the highest power is the sound source localization point
Figure DEST_PATH_IMAGE012
=
Figure DEST_PATH_IMAGE014
(
Figure 635158DEST_PATH_IMAGE010
；
(4) Total power of any one point
Figure 534981DEST_PATH_IMAGE008
(
Figure DEST_PATH_IMAGE016
) Performing generalized cross-correlation based on phase transformation on signals of all microphone pairs on a microphone array pairwise and summing:
Figure 163409DEST_PATH_IMAGE008
(
Figure 261815DEST_PATH_IMAGE016
)=
Figure DEST_PATH_IMAGE018
wherein k and l represent the kth and l microphones,
Figure DEST_PATH_IMAGE020
weights representing phase transformations, τ, (
Figure DEST_PATH_IMAGE022
) Represents the time from the sound arriving at the kth microphone from position X; in the formula (II)
Figure 12602DEST_PATH_IMAGE020
Defined as a combined weighting function
Figure DEST_PATH_IMAGE024
：
Figure DEST_PATH_IMAGE026
Taking into account the calculation
Figure 524355DEST_PATH_IMAGE008
(
Figure 628577DEST_PATH_IMAGE016
) Symmetry involved, and removing some fixed energy terms, then
Figure 847069DEST_PATH_IMAGE008
(
Figure 644123DEST_PATH_IMAGE016
) Moiety varying with x
Figure DEST_PATH_IMAGE028
Comprises the following steps:
Figure 321136DEST_PATH_IMAGE028
=
Figure DEST_PATH_IMAGE030
(5) global search is carried out in the whole room, and a coordinate point Y with the maximum energy is obtained by utilizing a random region contraction algorithm (SRC); randomly finding out an N-dimensional matrix in a given initial value, gradually reducing the range in the sequential process until reaching a small enough range, and finding out a peak value; thereby calculating a positioning coordinate point.
2. The method of claim 1, further comprising: in the step (4), calculation is performed
Figure 556945DEST_PATH_IMAGE012
Can be replaced by:
Figure DEST_PATH_IMAGE032
=
Figure DEST_PATH_IMAGE034
。
3. the method of claim 1, further comprising: in step (5), the process of the random region shrinking algorithm is as follows:
1) i is defined as the number of iterations,
Figure DEST_PATH_IMAGE036
representing the number of points randomly drawn at the ith iteration,
Figure DEST_PATH_IMAGE038
in a sub-search space representing the next generationThe number of points to be included in the content,
Figure DEST_PATH_IMAGE040
representing a next generation sub-search space; define once per calculation
Figure 754577DEST_PATH_IMAGE028
Is just recorded as one
Figure DEST_PATH_IMAGE042
，
Figure DEST_PATH_IMAGE044
After the ith iteration
Figure 784850DEST_PATH_IMAGE042
The number of times of the operation of the motor,
Figure DEST_PATH_IMAGE046
denotes the stop value,. phi.
Figure 943299DEST_PATH_IMAGE042
The maximum number of allowed calculations;
Figure DEST_PATH_IMAGE048
representing a new sub-search space
Figure 451640DEST_PATH_IMAGE040
The boundary of (2);
2) the number of initialization iterations i = 0;
3) setting initial parameters:
Figure DEST_PATH_IMAGE050
、
Figure DEST_PATH_IMAGE052
，
Figure DEST_PATH_IMAGE054
；
4) computing
Figure 707041DEST_PATH_IMAGE036
All of
Figure 908215DEST_PATH_IMAGE028
A value of (d);
5) arrange out
Figure 491643DEST_PATH_IMAGE038
So that
Figure 803676DEST_PATH_IMAGE038
≪
Figure 788950DEST_PATH_IMAGE036
；
6) According to
Figure 98708DEST_PATH_IMAGE038
Shrinking the current search space and updating the search space
Figure 231749DEST_PATH_IMAGE040
And new region boundaries
Figure 285156DEST_PATH_IMAGE048
；
7) If it is not
Figure DEST_PATH_IMAGE056
Or is or
Figure DEST_PATH_IMAGE058
And is
Figure DEST_PATH_IMAGE060
Determining the coordinate position of the point, storing the result and outputting the result;
8) if only
Figure 249570DEST_PATH_IMAGE058
If yes, the result is discarded;
9) in that
Figure 792547DEST_PATH_IMAGE038
To find a subset
Figure DEST_PATH_IMAGE062
So that
Figure 141445DEST_PATH_IMAGE062
Is greater than
Figure 998543DEST_PATH_IMAGE038
Average value of (d);
10) repeating the step 3) and the step 4) in the current search space
Figure 755146DEST_PATH_IMAGE040
In the random selection
Figure DEST_PATH_IMAGE064
Points, calculating their corresponding
Figure 469024DEST_PATH_IMAGE028
A value of (d);
11) will be provided with
Figure 514341DEST_PATH_IMAGE062
Put the points in the subset
Figure DEST_PATH_IMAGE066
And selecting
Figure 971867DEST_PATH_IMAGE064
In
Figure 582977DEST_PATH_IMAGE028
Of greatest value
Figure DEST_PATH_IMAGE068
Point put in subset
Figure 671018DEST_PATH_IMAGE066
In and store
Figure 369DEST_PATH_IMAGE066
Collecting and putting the data into the next iteration for use;
12) order to
Figure DEST_PATH_IMAGE070
And performing the next iteration and returning to the step 5).
CN202010206270.3A 2020-03-23 2020-03-23 Indoor sound source positioning method based on different microphone array topological structure analysis Pending CN111273231A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010206270.3A CN111273231A (en) 2020-03-23 2020-03-23 Indoor sound source positioning method based on different microphone array topological structure analysis

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010206270.3A CN111273231A (en) 2020-03-23 2020-03-23 Indoor sound source positioning method based on different microphone array topological structure analysis

Publications (1)

Publication Number Publication Date
CN111273231A true CN111273231A (en) 2020-06-12

Family

ID=70998417

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010206270.3A Pending CN111273231A (en) 2020-03-23 2020-03-23 Indoor sound source positioning method based on different microphone array topological structure analysis

Country Status (1)

Country Link
CN (1) CN111273231A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113312985A (en) * 2021-05-10 2021-08-27 中国地质大学（武汉） Audio-visual dual-mode 360-degree omnibearing speaker positioning method
CN113589229A (en) * 2021-09-29 2021-11-02 山东世纪矿山机电有限公司 Target positioning device for multi-array noise positioning and using method
CN114720943A (en) * 2022-06-06 2022-07-08 深圳市景创科技电子股份有限公司 Multi-channel sound source positioning method and system

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106093864A (en) * 2016-06-03 2016-11-09 清华大学 A kind of microphone array sound source space real-time location method
CN107102296A (en) * 2017-04-27 2017-08-29 大连理工大学 A kind of sonic location system based on distributed microphone array
CN107621625A (en) * 2017-06-23 2018-01-23 桂林电子科技大学 Sound localization method based on double micro-microphone battle arrays

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106093864A (en) * 2016-06-03 2016-11-09 清华大学 A kind of microphone array sound source space real-time location method
CN107102296A (en) * 2017-04-27 2017-08-29 大连理工大学 A kind of sonic location system based on distributed microphone array
CN107621625A (en) * 2017-06-23 2018-01-23 桂林电子科技大学 Sound localization method based on double micro-microphone battle arrays

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
WEIPING CAI等: "Localization of Multiple Speech Sources Based on Sub-band Steered Response Power", 《2010 INTERNATIONAL CONFERENCE ON ELECTRICAL AND CONTROL ENGINEERING》 *
WEIPING CAI等: "Localization of Multiple Speech Sources Based on Sub-band Steered Response Power", 《2010 INTERNATIONAL CONFERENCE ON ELECTRICAL AND CONTROL ENGINEERING》, 11 November 2010 (2010-11-11), pages 1246 - 1249 *
张铁成: "基于麦克风阵列的声音识别与定位算法研究", 《中国优秀博硕士学位论文全文数据可（硕士） 信息科技辑》 *
张铁成: "基于麦克风阵列的声音识别与定位算法研究", 《中国优秀博硕士学位论文全文数据可（硕士） 信息科技辑》, 15 February 2017 (2017-02-15), pages 1 - 2 *
袁安富等: "一种改进的联合SRP-PHAT语音定位算法", 《南京信息工程大学学报(自然科学版)》 *
袁安富等: "一种改进的联合SRP-PHAT语音定位算法", 《南京信息工程大学学报(自然科学版)》, no. 06, 28 December 2012 (2012-12-28), pages 57 - 62 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113312985A (en) * 2021-05-10 2021-08-27 中国地质大学（武汉） Audio-visual dual-mode 360-degree omnibearing speaker positioning method
CN113589229A (en) * 2021-09-29 2021-11-02 山东世纪矿山机电有限公司 Target positioning device for multi-array noise positioning and using method
CN113589229B (en) * 2021-09-29 2022-01-21 山东世纪矿山机电有限公司 Target positioning device for multi-array noise positioning and using method
CN114720943A (en) * 2022-06-06 2022-07-08 深圳市景创科技电子股份有限公司 Multi-channel sound source positioning method and system

Similar Documents

Publication Publication Date Title
CN111273231A (en) Indoor sound source positioning method based on different microphone array topological structure analysis
CN106093864B (en) A kind of microphone array sound source space real-time location method
CN102324237B (en) Microphone-array speech-beam forming method as well as speech-signal processing device and system
EP1429581A2 (en) Method of broadband constant directivity beamforming for non linear and non axi-symmetric sensor arrays embedded in a obstacle
MX2014006499A (en) Apparatus and method for microphone positioning based on a spatial power density.
CN109188362B (en) Microphone array sound source positioning signal processing method
CN113192486B (en) Chorus audio processing method, chorus audio processing equipment and storage medium
CN111856402B (en) Signal processing method and device, storage medium and electronic device
CN115902776B (en) Sound source positioning method based on passive sound signal
CN112255629A (en) Sequential ESPRIT two-dimensional incoherent distribution source parameter estimation method based on combined UCA array
CN110736976A (en) sonar beam former performance estimation method of arbitrary array
CN113593596B (en) Robust self-adaptive beam forming directional pickup method based on subarray division
CN114994608B (en) Multi-device self-organizing microphone array sound source positioning method based on deep learning
CN113491137B (en) Flexible differential microphone array with fractional order
CN104008287A (en) Sound field reconstructing and ghost image restraining method based on PSO-MVDR
Sun et al. Robust spherical microphone array beamforming with multi-beam-multi-null steering, and sidelobe control
Lima et al. Efficient steered-response power methods for sound source localization using microphone arrays
Nakano et al. Automatic estimation of position and orientation of an acoustic source by a microphone array network
Lan et al. Underwater Acoustic DOA Estimation of incoherent signal based on improved GA-MUSIC
Pertilä Acoustic source localization in a room environment and at moderate distances
CN109239665B (en) Multi-sound-source continuous positioning method and device based on signal subspace similarity spectrum and particle filter
CN113591537B (en) Double-iteration non-orthogonal joint block diagonalization convolution blind source separation method
CN111983553B (en) Gridless DOA estimation method based on cross-prime multi-carrier-frequency sparse array
Mazur et al. Robust room equalization using sparse sound-field reconstruction
CN106950530B (en) Acoustic vector-sensor array column decorrelation LMS method for parameter estimation based on cross covariance

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20200612