WO2018119910A1 - 一种运动声源的定位方法和定位装置 - Google Patents

一种运动声源的定位方法和定位装置 Download PDF

Info

Publication number
WO2018119910A1
WO2018119910A1 PCT/CN2016/113108 CN2016113108W WO2018119910A1 WO 2018119910 A1 WO2018119910 A1 WO 2018119910A1 CN 2016113108 W CN2016113108 W CN 2016113108W WO 2018119910 A1 WO2018119910 A1 WO 2018119910A1
Authority
WO
WIPO (PCT)
Prior art keywords
sound source
search
sound
motion
search point
Prior art date
Application number
PCT/CN2016/113108
Other languages
English (en)
French (fr)
Inventor
杨殿阁
连小珉
张凯
苗丰
Original Assignee
清华大学
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 清华大学 filed Critical 清华大学
Priority to PCT/CN2016/113108 priority Critical patent/WO2018119910A1/zh
Publication of WO2018119910A1 publication Critical patent/WO2018119910A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
    • G01H11/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves by detecting changes in electric or magnetic properties
    • 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
    • G01S5/22—Position of source determined by co-ordinating a plurality of position lines defined by path-difference measurements

Definitions

  • Embodiments of the present application relate to, but are not limited to, the field of sound source localization technology, and in particular, a positioning method and a positioning device for a motion sound source based on a time difference of sound.
  • the sound source localization technique is a method of receiving a sound signal through a microphone array, performing calculation processing, and finally determining the position of the sound source in space.
  • the sound source localization method based on Time Difference of Arrival (TDOA) uses a generalized cross-correlation and other time delay estimation algorithm to calculate the delay between each microphone, and estimates the sound source position according to the spatial positional relationship of the microphone. Only four are needed.
  • the microphone can locate the three-dimensional position of the sound source, and has a small amount of calculation and good real-time performance, and has been widely used.
  • the static TDOA method is not suitable for motion sound sources, such as short-term motion sound sources such as gunshots, gunshots, and sudden fault sounds on fast moving platforms, as well as steady-state moving sound sources that continuously emit sound.
  • the embodiment of the present application provides a positioning method and a positioning device for a moving sound source.
  • the positioning method of the moving sound source is a method for quickly locating a moving sound source based on the TDOA method, and includes the following steps:
  • the sound source localization method using the sound time difference is used to obtain the estimated position of the motion sound source, and the deviation between the estimated position of the motion sound source and the corresponding search point is calculated. Difference
  • the position of the search point corresponding to the minimum deviation value is obtained, and the position of the search point is taken as the position of the motion sound source.
  • the method provided by the embodiment of the present invention uses a microphone array to collect sound pressure signals of a moving sound source, establishes a sound source search surface, determines a search point on the sound source search surface, and then determines a sound source estimated position by using a TDOA method, and calculates The deviation between the estimated position and the search point.
  • the location of the search point is the actual position of the motion sound source, the deviation value is zero. Therefore, in the method of the embodiment, the search point that minimizes the deviation value is searched for on the sound source search surface, and the search point is searched at this time.
  • the deviation between the position and the actual position of the moving sound source is the smallest, so the search point is approximated to the actual position where the moving sound source is located, and the positioning of the moving sound source is realized.
  • the positioning device of the moving sound source provided by the embodiment of the present application is based on the positioning method described above, and the positioning device includes:
  • At least four microphones arranged as an array of microphones, the microphone array being arranged to collect sound pressure signals of the moving sound source;
  • Processing device including:
  • a receiving module for receiving the sound pressure signal
  • a display device is provided for displaying an estimated position of the displayed motion sound source for output.
  • FIG. 1 is a schematic flow chart of a method for locating a moving sound source according to an embodiment of the present application
  • FIG. 2 is a schematic flow chart of a method for locating a moving sound source according to another embodiment of the present application
  • FIG. 3 is a schematic diagram of a sensor array and a sound source search surface when positioning by using a positioning method of a moving sound source according to an embodiment of the present application;
  • FIG. 4 is a schematic diagram of a model for eliminating Doppler effect when positioning by using a positioning method of a moving sound source according to an embodiment of the present application;
  • FIG. 5 is a schematic diagram of a simplex method search on a sound source search surface
  • FIG. 6 is a schematic flow chart of a method for locating a moving sound source according to still another embodiment of the present application.
  • FIG. 7 is a schematic diagram of determining a search point by using a grid method on a sound source search surface
  • FIG 9 is a schematic view of the processing device of the positioning device shown in Figure 8.
  • FIGS. 10a-10d are first examples of positioning by using the positioning method of the embodiment of the present application, wherein FIG. 10a is a schematic diagram of a microphone array, and FIGS. 10b-10d are schematic diagrams of results of positioning by using different positioning methods;
  • FIGS. 11a-11d are second examples of positioning by using the positioning method of the embodiment of the present application, wherein FIG. 11a is a schematic diagram of a microphone array, and FIGS. 11b-11d are schematic diagrams of results obtained by using different positioning methods.
  • An embodiment of the present application provides a method for positioning a motion sound source based on TDOA. As shown in FIG. 1 , the positioning method includes the following steps:
  • the sound source search surface 1) Discretize the sound source search surface to obtain a series of grid points, which are the search points on the sound source search surface.
  • the estimated position of the motion sound source can be obtained by using the TDOA method and the deviation value between the estimated position of the motion sound source and the corresponding search point can be calculated; among all the deviation values, the deviation value is the most The position of the search point corresponding to the hour is the position of the moving sound source.
  • MTDOA Moving Time Difference Of Arrival
  • the process of determining the position of the moving sound source by the simplex method is as follows: the point on the search surface of the selected sound source is the search point, and the estimated position of the moving sound source is obtained by using the TDOA method and the estimated position and corresponding position of the moving sound source are calculated.
  • the deviation value between the search points determines whether the deviation value satisfies the iteration termination condition; if satisfied, the estimated position of the motion sound source at this time is the position of the motion sound source, and if not, the determination is performed according to the simplex method.
  • the above-described motion TDOA positioning method based on the simplex method is called FMTDOA (Fast Moving Time Difference Of Arrival) positioning method.
  • the search point is assumed to be the position of the moving sound source, and the collected sound pressure signal is subjected to Doppler effect canceling processing.
  • the step of arranging the array of microphones at least four microphones are arranged on the same plane.
  • the microphone array is a planar quaternary array comprising four microphones.
  • the sound source search surface is parallel to the plane in which the microphone array is located.
  • the embodiment of the present application adopts a planar quaternary array, that is, four microphones are arranged on the plane Pm.
  • step S202 The sound pressure signal is acquired by the microphone array (step S202), and the sound source search surface Ps is established (step S203).
  • the sound source search surface Ps should be parallel to the xOy plane, and the range is x ⁇ (x min , x max ), y ⁇ (y min , y max ) (where, usually x
  • the range is ⁇ 2m
  • the range of y is ⁇ 2m.
  • the range of x and y can be changed according to the application.
  • step S204 selecting a point on the sound source search surface as a search point (step S204), and then performing a Doppler effect canceling process on the collected sound pressure signal by using a time domain non-simplified model-based Doppler effect canceling method (step S205) .
  • the Doppler effect cancellation method is as follows: the sound source is located at s(t) at time t, the distance between the i-th microphone m i and s(t) is r i (t), and the microphone m at time t
  • the signal received by i is the sound source emitted at the s(t- ⁇ (t)) position, where c is the speed of sound, and R i (t) is the distance between mi and s(t- ⁇ (t)).
  • the sound pressure signal received by the microphone from Morse's theory of motion and acoustics can be expressed as:
  • This signal is the actual signal received by the microphone with the amplitude and phase offset caused by the Doppler effect. Since the position of the real sound source is unknown, the Doppler effect can only be eliminated for the search point on the sound source search surface S (ie, the assumed sound source point). For g k (t), after correcting the amplitude and phase, the sound pressure signal that eliminates the Doppler effect is
  • the position vector For the search point g k (t) on the sound source search plane Ps, the position vector is The corresponding delay is recorded as ⁇ t i1,k , and the equations for solving the sound source coordinates can be obtained:
  • step S207 Calculating the positioning deviation (step S207), defining the search point
  • the position deviation of the sound source is
  • the D function is a binary function about (x, y) defined in the plane Ps. If the search point is located at the actual sound source point, the sound source estimation position calculated in step 4) Same as r s , ie otherwise Therefore, the sound source point is the zero point of D.
  • step S209 determining whether the positioning deviation satisfies the iteration termination condition (step S209), and if not, determining the next search point by the simplex method (step S208), and then repeating the above steps 4), 5) (ie, steps S205-S209), Until the positioning deviation satisfies the iterative termination condition, the search point position at this time is output as the coordinates of the motion sound source position (step S210).
  • the minimum point coordinate of the sound source position deviation function D can be obtained. which is Then the point can be approximated as the sound source position.
  • ⁇ is the allowable error, which determines the accuracy of the algorithm and the number of iterations.
  • the initial point will converge to the minimum point. This point can be approximated as the sound source point.
  • the embodiment of the present application has the following advantages: 1. Based on the TDOA method, the positioning of the moving sound source is realized; 2. The TDOA positioning method using the Doppler effect is eliminated, and the signal amplitude caused by the Doppler effect is eliminated. The phase offset solves the problem that TDOA can not be used for positioning the moving sound source. 3. Only 4 microphones can achieve better positioning effect, which greatly reduces the size of the array space and makes the measurement and calculation process simpler. Using the simplex search algorithm to find the real moving sound source on the sound source search surface, the initial point of the moving sound source position deviation function D can be obtained by several iterations, and then the position of the moving sound source can be obtained to avoid traversing the whole sound. The source search surface finds the real motion sound source, which greatly reduces the amount of calculation and significantly improves the positioning timeliness.
  • Another embodiment of the present application also provides a moving sound source using the MTDOA positioning method.
  • the process of positioning includes: arranging a microphone array (step S301), collecting a sound pressure signal of the moving sound source with the microphone array (step S302) and establishing a sound source search surface (step S303).
  • the aforementioned three steps S301-S303 are the same as steps S201-S203 in the positioning method shown in FIG. 2.
  • the method shown in FIG. 6 differs from the positioning method shown in FIG. 2 in that:
  • the sound source search surface Ps is discretized to obtain M ⁇ N grid points (step S304).
  • the degree of discretization and the number of grid points obtained can be set according to the positioning accuracy requirements;
  • step S305 using each grid point as a search point (ie, a hypothetical sound source point), the Doppler effect is eliminated (step S305);
  • step S306 After the Doppler effect is eliminated, the TDOA calculation is performed (step S306);
  • the magnitudes of the values of all the sound source position deviations are compared, wherein the position of the search point (mesh point) corresponding to the deviation value is the position of the motion sound source, and the coordinates of the grid point are output (step S308).
  • the positioning method provided in this embodiment needs to traverse all the search points of the sound source search surface, and then find the position of the search point when the deviation value is the smallest.
  • step 5 for the kth search point g k (t) on the sound source search surface, an estimate of two or more motion sound sources can be obtained. Need to be right now Take the average and then use the averaged The calculation of the sound source position deviation D function is performed.
  • the embodiment of the present application further provides a positioning device for a moving sound source.
  • the positioning device operates according to the above positioning method, including four microphones m 1 and m arranged in a planar quaternary array. 2 , m 3 and m 4 , the processing device 1 and the display device 2, wherein the four microphones, the processing device 1 and the display device 2 can be in communication communication by wire or wireless.
  • four microphones m 1 , m 2 , m 3 and m 4 can be used to collect the sound pressure signals of the moving sound source.
  • the processing device 1 includes: a module M10 for receiving the sound pressure signal; a module M11 for establishing a sound source search surface; and for using the sound based on the search point on the sound source search surface a sound source localization method of time difference to obtain an estimated position of the motion sound source, and a module M12 for calculating a deviation value between the estimated position of the motion sound source and the corresponding search point; and a method for obtaining the deviation value
  • the estimated position of the moving sound source corresponding to the hour, and the estimated position of the moving sound source is outputted by the module M13.
  • the processing device 1 can receive the sound pressure signal collected by the microphone, and establish a sound source search surface.
  • the estimated position of the motion sound source is obtained by using the TDOA method, and the estimated position of the motion sound source is calculated.
  • the deviation value between the corresponding search points is obtained, and the position of the search point corresponding to the minimum deviation value is obtained, and the position of the search point is output.
  • the display device 2 can be used to receive the position of the output search point and display it.
  • the above embodiment includes only four microphones, and in other embodiments, five or more microphones disposed on the same plane may also be included.
  • the following is an example of performing simulation using the positioning method and the positioning device provided by the embodiment of the present application.
  • the microphone array is shown in Figure 10a and its coordinates:
  • Figure 10b shows the positioning result of the static TDOA method without Doppler effect
  • Figure 10c shows the motion of TDOA (MTDOA) based on the grid scan after eliminating the Doppler effect
  • Figure 10d is based on the simple elimination of the Doppler effect. Shape search for the motion TDOA (FMTDOA) positioning results.
  • Table 1 shows the results of positioning using different positioning methods.
  • the grid scan (0.1x0.1) indicates that the point on the search surface of the sound source with the grid size of 0.1mx0.1m is scanned.
  • the grid scan (0.05x0.05) indicates that the grid size is 0.05.
  • a point on the sound source search surface of mx0.05m is scanned.
  • Sound source 1000Hz single-frequency sine wave, moving speed 50km/h;
  • the microphone array is shown in Figure 11a;
  • Figure 11b shows the positioning results of the static TDOA method without the Doppler effect
  • Figure 11c shows the motion of the TDOA based on the grid scan after eliminating the Doppler effect
  • Figure 11d is based on the simple elimination of the Doppler effect. Shape search for the motion TDOA (FMTDOA) positioning results.
  • Table 2 shows the results of positioning using the positioning methods in Figs. 11c and 11d.
  • Positioning method Positioning result (m) Positioning error (m) Calculation time (s) Grid scan (0.105, 0.003, 3) 0.024 68.8 Simplex search (0.090, -0.014, 3) 0.023 9.2
  • each module/unit in the above embodiment may be implemented in the form of hardware, for example, by implementing an integrated circuit to implement its corresponding function, or may be implemented in the form of a software function module, for example, executing a program stored in the memory by a processor. / instruction to achieve its corresponding function.
  • Embodiments of the invention are not limited to any specific form of combination of hardware and software.
  • the embodiment of the present application utilizes fewer microphones, can eliminate the influence of the Doppler effect on the moving sound source, and quickly and effectively locate the sound source. Breaking through the dependence of the traditional sound source identification on the large microphone array, and breaking the sonic time difference method can only be used for the positioning of the stationary sound source, providing a new method for the rapid identification of various short-time sound sources. .

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)

Abstract

一种运动声源的定位方法和定位装置，其中，该运动声源的定位方法包括：将至少四个传声器布置成传声器阵列(S102)；用传声器阵列采集运动声源的声压信号，并建立声源搜索面(S104)；基于声源搜索面上的搜索点，利用声达时间差的声源定位方法得出运动声源的估计位置，并计算运动声源的估计位置与对应的搜索点之间的偏差值(S106)；得出偏差值最小时所对应的搜索点的位置，将该搜索点的位置作为运动声源的位置(S108)。

Description

一种运动声源的定位方法和定位装置 技术领域
本申请实施例涉及但不限于声源定位技术领域，特别是一种基于声达时间差的运动声源的定位方法和定位装置。
背景技术
声源定位技术是一种通过传声器阵列接收声音信号，进行计算处理，最终确定声源在空间中的位置的方法。基于声达时间差(Time Difference of Arrival，TDOA)的声源定位方法利用广义互相关等时延估计算法计算各传声器间的时延，并根据传声器的空间位置关系估计声源位置，只需要四个传声器即可定位声源的三维位置，计算量小、实时性好，得到了广泛应用。但是该静态TDOA方法不适用于运动声源，例如快速移动平台上的枪声、炮声及车辆的突发故障声等短时运动声源以及连续发声的稳态运动声源。
研究如何利用少量传声器，简便、高效地识别定位，建立运动声源的快速定位方法是非常有意义的。
发明内容
以下是对本文详细描述的主题的概述。本概述并非是为了限制权利要求的保护范围。
本申请实施例提供了一种运动声源的定位方法和定位装置。
本申请实施例提供的运动声源的定位方法是基于TDOA方法的运动声源快速定位方法，包括以下步骤：
将至少四个传声器布置传声器阵列；
用传声器阵列采集运动声源的声压信号，并建立声源搜索面；
基于声源搜索面上的搜索点，利用声达时间差的声源定位方法得出运动声源的估计位置，并计算运动声源的估计位置与对应的搜索点之间的偏 差值；
得出偏差值最小时所对应的搜索点的位置，将该搜索点的位置作为运动声源的位置。
本申请实施例提供的方法，利用传声器阵列采集运动声源的声压信号，并建立声源搜索面，在声源搜索面上确定搜索点，然后利用TDOA方法确定声源的估计位置，并计算出估计位置与搜索点之间的偏差值。理论上，若搜索点所在的位置为运动声源的实际位置，则偏差值为零，因此，在本实施例的方法在声源搜索面上寻找使偏差值最小的搜索点，此时搜索点的位置与运动声源的实际位置之间的偏差最小，故而该搜索点近似为运动声源所在的实际位置，实现了对运动声源的定位。
本申请实施例提供的运动声源的定位装置是基于上述的定位方法进行定位，该定位装置包括：
至少四个传声器，布置成传声器阵列，所述传声器阵列设置成用于采集运动声源的声压信号；
处理装置，包括：
用于接收所述声压信号的接收模块，
用于建立声源搜索面的模块，
用于基于所述声源搜索面上的搜索点，利用声达时间差的声源定位方法得出运动声源的估计位置，并计算所述运动声源的估计位置与对应的搜索点之间的偏差值的模块，和
用于得出所述偏差值最小时所对应的运动声源的估计位置，并将该运动声源的估计位置输出的模块；以及
显示装置，设置成用于输出的显示运动声源的估计位置。
在阅读并理解了附图和详细描述后，可以明白其他方面。
附图概述
当结合附图考虑时，通过参照下面的详细描述，能够更完整更好地理 解本申请实施例以及容易得知其中许多伴随的优点，但此处所说明的附图用来提供对本申请实施例的进一步理解，构成本申请实施例的一部分，本申请的示意性实施例及其说明用于解释本申请，并不构成对本申请的不当限定，如图其中：
图1是本申请实施例的运动声源的定位方法的流程示意图；
图2是本申请另一实施例的运动声源的定位方法的流程示意图；
图3是利用本申请实施例的运动声源的定位方法定位时传感器阵列与声源搜索面的示意图；
图4为利用本申请实施例的运动声源的定位方法定位时消除多普勒效应的模型示意图；
图5为在声源搜索面上利用单纯形法搜索时的示意图；
图6是本申请又一实施例的运动声源的定位方法的流程示意图；
图7为在声源搜索面上利用网格法确定搜索点的示意图；
图8是本申请实施例的运动声源的定位装置的结构示意图，其中四传声器阵列平面位于xOy平面，运动声源在z＝z0平面内以一定速度v与x轴平行的方向移动；
图9是图8所示的定位装置的处理装置的示意图；
图10a-图10d为利用本申请实施例的定位方法进行定位的第一示例，其中图10a是传声器阵列的示意图，图10b-图10d为利用不同定位方法进行定位的结果示意图；以及
图11a-图11d为利用本申请实施例的定位方法进行定位的第二示例，其中图11a是传声器阵列的示意图，图11b-图11d为利用不同定位方法进行定位的结果示意图。
本发明的实施方式
显然，本领域技术人员基于本申请的宗旨所做的许多修改和变化属于本申请的保护范围。
本申请的实施例提供了一种基于TDOA的运动声源的定位方法，如图1所示，该定位方法包括以下步骤：
S102，将至少四个传声器布置传声器阵列；
S104，用传声器阵列采集运动声源的声压信号，并建立声源搜索面；
S106，基于声源搜索面上的搜索点，利用声达时间差的声源定位方法得出运动声源的估计位置，并计算运动声源的估计位置与对应的搜索点之间的偏差值；
S108，得出偏差值最小时所对应的搜索点的位置，将该搜索点的位置作为运动声源的位置。
其中，对于声源搜索面上的搜索点的确定，可采用两种方法：
1)对声源搜索面进行离散化，得到一系列的网格点，该网格点即为声源搜索面上的搜索点。此时，可基于所有搜索点，利用TDOA方法得出运动声源的估计位置并计算出运动声源的估计位置与对应的搜索点之间的偏差值；在所有的偏差值中，偏差值最小时所对应的搜索点的位置即为运动声源的位置。上述基于网格扫描的运动TDOA定位方法称为MTDOA(Moving Time Difference Of Arrival)定位方法。
2)利用单纯形法，确定声源搜索面上的搜索点。利用单纯形法确定运动声源的位置的过程为：选定声源搜索面上的一点为搜索点，并利用TDOA方法得出运动声源的估计位置并计算出运动声源的估计位置与对应的搜索点之间的偏差值，判断偏差值是否满足迭代终止条件；若满足，则此时的运动声源的估计位置即为运动声源的位置，若不满足，则根据单纯形法确定下一搜索点，然后重复利用TDOA方法得出运动声源的估计位置并计算出运动声源的估计位置与对应的搜索点之间的偏差值以及判断偏差值是否满足迭代终止条件的步骤，直到偏差值满足迭代终止条件，则此时的运动声源的估计位置即为运动声源的位置。上述基于单纯形法的运动TDOA定位方法称为FMTDOA(Fast Moving Time Difference Of Arrival)定位方法。
可选地，在基于搜索点进行偏差值计算的步骤中，在基于搜索点得出 运动声源的估计位置之前，假定该搜索点为运动声源的位置，对采集到的声压信号进行多普勒效应消除处理。
可选地，在布置传声器阵列的步骤中，将至少四个传声器布置在同一平面上。
可选地，传声器阵列为包括四个传声器的平面四元阵列。
可选地，声源搜索面与传声器阵列所在的平面平行。
下面，结合图2的流程图说明采用FMTDOA定位方法对运动声源进行定位的过程。
1)布置传声器阵列(步骤S201)，将传声器阵列布置成如图3所示，本申请实施例采用平面四元阵列，即将四个传声器布置在平面Pm上。首先建立空间坐标系，面向运动声源，水平向右为x轴正方向，竖直向下为y轴正方向，z轴水平指向声源，坐标原点O为传声器阵列的中心(如图8中所示)。则第i(i＝1,2,3,4)个传声器mi的坐标矢量为
Figure PCTCN2016113108-appb-000001
2)用传声器阵列采集声压信号(步骤S202)，并建立声源搜索面Ps(步骤S203)。如图8和图3所示，在平面z＝z0处运动声源以一定速度v沿与x轴平行的方向运动。设t时刻声源点s(t)的位置矢量为rs(t)＝(xs(t),ys,z0)。以运动声源所在平面为声源搜索面Ps，声源搜索面Ps应平行于xOy平面，范围为x∈(xmin,xmax),y∈(ymin,ymax)(其中，通常x的范围为±2m，y的范围为±2m。当然，x、y的范围可根据应用情况进行变化)。
3)选择声源搜索面上的一点为搜索点(步骤S204)，然后采用基于时域非简化模型的多普勒效应消除方法对采集的声压信号进行多普勒效应消除处理(步骤S205)。
如图4所示，多普勒效应消除的方法为：设t时刻声源位于s(t)，第i个传声器mi与s(t)的距离为ri(t)，t时刻传声器mi接收到的信号是声源在s(t-δ(t))位置发出的，其中
Figure PCTCN2016113108-appb-000002
c为声速，Ri(t)是mi与s(t-δ(t))的距离。由莫尔斯运动声学理论可得传声器接收到的声压信号可表示为：
Figure PCTCN2016113108-appb-000003
其中
Figure PCTCN2016113108-appb-000004
是马赫数。
该信号为传声器接收到的实际信号，带有多普勒效应引起的幅值和相位的偏移。由于真实声源位置未知，所以只能对声源搜索面S上的搜索点(即假定的声源点)进行多普勒效应的消除。对于gk(t)，经幅值和相位修正后，消除多普勒效应的声压信号为
Figure PCTCN2016113108-appb-000005
其中，o(ξ)是时延算子有o(ξ)·f(t)＝f(t)·o(ξ)＝f(t+ξ)，
Figure PCTCN2016113108-appb-000006
是gk(t0)与mi的距离，
Figure PCTCN2016113108-appb-000007
与
Figure PCTCN2016113108-appb-000008
可以写为
Figure PCTCN2016113108-appb-000009
Figure PCTCN2016113108-appb-000010
4)经多普勒效应消除后，进行TDOA计算(步骤S206)，计算各个传声器mi(i＝2,3，4)相对参考传声器m1的时延Δti1。对于声源搜索平面Ps上的搜索点gk(t)，其位置矢量为
Figure PCTCN2016113108-appb-000011
对应的时延记为Δti1,k，可得到用于求解声源坐标的方程组：
Figure PCTCN2016113108-appb-000012
其中，c为声速，
Figure PCTCN2016113108-appb-000013
是由搜索点
Figure PCTCN2016113108-appb-000014
推导出的运动声源的估计值。
对于四传声器平面阵列，可以求得
Figure PCTCN2016113108-appb-000015
的三个坐标：
Figure PCTCN2016113108-appb-000016
Figure PCTCN2016113108-appb-000017
Figure PCTCN2016113108-appb-000018
其中，
Figure PCTCN2016113108-appb-000019
Figure PCTCN2016113108-appb-000020
Figure PCTCN2016113108-appb-000021
Figure PCTCN2016113108-appb-000022
Figure PCTCN2016113108-appb-000023
5)计算定位偏差(步骤S207)，定义搜索点
Figure PCTCN2016113108-appb-000024
处的声源位置偏差为
Figure PCTCN2016113108-appb-000025
则D函数为定义在平面Ps内的关于(x,y)的二元函数，如果搜索点就位于实际声源点，经步骤4)计算后的声源估计位置
Figure PCTCN2016113108-appb-000026
就与rs相同，即
Figure PCTCN2016113108-appb-000027
否则
Figure PCTCN2016113108-appb-000028
因此声源点就为D的零点即
Figure PCTCN2016113108-appb-000029
6)判断定位偏差是否满足迭代终止条件(步骤S209)，若否，则利用单纯形法确定下一搜索点(步骤S208)，然后重复上述步骤4)、5)(即步骤S205-S209)，直到定位偏差满足迭代终止条件，输出此时的搜索点位置，作为运动声源位置的坐标(步骤S210)。
其中，利用单纯形优化搜索算法按照步骤4)、5)对平面Ps进行最小值搜索(搜索过程如图5所示)，可以求得声源位置偏差函数D的最小值点坐标
Figure PCTCN2016113108-appb-000030
即
Figure PCTCN2016113108-appb-000031
则该点就可近似为声源位置。
求解过程如下：
①在平面Ps内取不共线3点作为单纯形的顶点，记为
Figure PCTCN2016113108-appb-000032
Figure PCTCN2016113108-appb-000033
并使
Figure PCTCN2016113108-appb-000034
令
Figure PCTCN2016113108-appb-000035
则
Figure PCTCN2016113108-appb-000036
关于
Figure PCTCN2016113108-appb-000037
的反射点为
Figure PCTCN2016113108-appb-000038
其中α为反射系数，通常α取1。
②令
Figure PCTCN2016113108-appb-000039
则
Figure PCTCN2016113108-appb-000040
关于
Figure PCTCN2016113108-appb-000041
的收缩点为
Figure PCTCN2016113108-appb-000042
其中β为收缩系数，根据二分法的思想，β通常取0.5。
③根据以下情况对
Figure PCTCN2016113108-appb-000043
或者
Figure PCTCN2016113108-appb-000044
进行替换
Figure PCTCN2016113108-appb-000045
或者
Figure PCTCN2016113108-appb-000046
④迭代收敛条件(即迭代终止条件)为
其中ε为允许误差，决定了算法的精度和迭代次数。
经若干次迭代后初始点即会收敛于最小值点
Figure PCTCN2016113108-appb-000048
该点就可以近似为声源点。
本申请的实施例具有以下优点：1、基于TDOA方法，实现了对于运动声源的定位；2、采用了消除多普勒效应的TDOA定位方法，消除了多普勒效应引起的信号幅值和相位的偏移，解决了TDOA不能用于运动声源定位的问题；3、仅采用4个传声器就可以取得较理想的定位效果，大大减少了阵列空间尺寸，使测量、计算过程更加简单；4、采用单纯形搜索算法寻找声源搜索面上的真实运动声源，进行若干次迭代就可求出运动声源位置偏差函数D的最小值点，进而获得运动声源的位置，避免遍历整个声源搜索面寻找真实运动声源，大大减少了计算量，显著地提高了定位时效性。
本申请的另一实施例还提供了一种采用MTDOA定位方法对运动声源 进行定位的过程，如图6所示，包括：布置传声器阵列(步骤S301)，用传声器阵列采集运动声源的声压信号(步骤S302)和建立声源搜索面(步骤S303)。前述三个步骤S301-S303与图2所示的定位方法中的步骤S201-S203相同。图6所示的方法与图2所示的定位方法的不同之处在于：
如图7所示，建立声源搜索面Ps后，将声源搜索面Ps离散化，得到M×N个网格点(步骤S304)。其中离散化的程度及得到的网格点的数目可根据定位精度要求设定；
然后以每个网格点为搜索点(即假定的声源点)，进行多普勒效应的消除(步骤S305)；
经多普勒效应消除后，进行TDOA计算(步骤S306)；
计算出对应于每个搜索点的声源位置偏差(步骤S307)；
比较所有的声源位置偏差的值的大小，其中偏差值最小时所对应的搜索点(网格点)的位置即为运动声源的位置，输出该网格点的坐标(步骤S308)。
本实施例提供的定位方法，需要对声源搜索面的所有搜索点进行遍历，然后找出其中偏差值最小时的搜索点的位置。
需要说明的是，在上述的两个实施例中，当运动声源的移动速度足够小，多普勒效应不太明显，并且对定位精度要求不高的情况下，可省略消除多普勒效应的步骤，简化定位过程。
上述两实施例中，均采用四个传声器。当然在其他实施例中，也可以采用5个或更多个传声器，其中四个传声器可以确定一个运动声源的估计值
Figure PCTCN2016113108-appb-000049
采用5个或更多个传声器时，在步骤5)中，对于声源搜索面上第k个搜索点gk(t)，可获得两个或更多个运动声源的估计值
Figure PCTCN2016113108-appb-000050
此时需对
Figure PCTCN2016113108-appb-000051
取平均，然后利用取平均后的
Figure PCTCN2016113108-appb-000052
进行声源位置偏差D函数的计算。
本申请的实施例还提供了一种运动声源的定位装置，如图8所示，该定位装置按照上述的定位方法进行工作，包括按平面四元阵列排布的四个传声器m1、m2、m3和m4、处理装置1和显示装置2，其中四个传声器、处理装置1和显示装置2可通过有线方式或无线方式进行通信连通。
其中，四个传声器m1、m2、m3和m4可用于采集运动声源的声压信号。
如图9所示，处理装置1包括：用于接收所述声压信号的模块M10；用于建立声源搜索面的模块M11；用于基于所述声源搜索面上的搜索点，利用声达时间差的声源定位方法得出运动声源的估计位置，并计算所述运动声源的估计位置与对应的搜索点之间的偏差值的模块M12；和用于得出所述偏差值最小时所对应的运动声源的估计位置，并将该运动声源的估计位置输出的模块M13。处理装置1可接收传声器采集的声压信号，并建立声源搜索面，基于声源搜索面上的搜索点，利用TDOA方法得出运动声源的估计位置，并计算运动声源的估计位置与对应的搜索点之间的偏差值，得出偏差值最小时所对应的搜索点的位置，并将该搜索点的位置输出。
显示装置2可用于接收输出的搜索点的位置并进行显示。
上述实施例中仅包含四个传声器，在其他实施例中，还可以包括设置在同一平面上的五个或更多个传声器。
下面为采用本申请实施例提供的定位方法和定位装置进行仿真的示例。
示例一：
声源：1000Hz单频正弦波，运动速度：30m/s，t＝0s时声源位于(0.06,0.06,5)m处；
传声器阵列如图10a所示，其坐标：
Figure PCTCN2016113108-appb-000053
采样频率：20000Hz；
0.1s信号用于定位处理；
定位误差：
Figure PCTCN2016113108-appb-000054
图10b为未消多普勒效应的静态TDOA方法的定位结果，图10c为消除多普勒效应后基于网格扫描的运动TDOA(MTDOA)定位结果，图10d为消除多普勒效应后基于单纯形搜索的运动TDOA(FMTDOA)定位结果。
表1为利用不同定位方法进行定位的结果。
表1
Figure PCTCN2016113108-appb-000055
表1中，网格扫描(0.1x0.1)表示对网格大小为0.1mx0.1m的声源搜索面上的点进行扫描，网格扫描(0.05x0.05)表示对网格大小为0.05mx0.05m的声源搜索面上的点进行扫描。
从表1及图10b-图10d可以看出，消除多普勒效应后对运动声源的定位精度提高，且单纯形搜索的定位方法比网格扫描的定位方法的用时短，定位精度高，实现了快速有效的定位。
示例二：
声源：1000Hz单频正弦波，移动速度50km/h；
麦克风阵列如图11a所示；
采样频率：20000Hz；
0.1s信号用于定位；
定位误差：
Figure PCTCN2016113108-appb-000056
图11b为未消除多普勒效应的静态TDOA方法的定位结果，图11c为消除多普勒效应后基于网格扫描的运动TDOA(MTDOA)定位结果，图11d为消除多普勒效应后基于单纯形搜索的运动TDOA(FMTDOA)定位结果。
表2为用图11c和图11d中的定位方法进行定位的结果。
表2
定位方法 定位结果(m) 定位误差(m) 计算时间(s)
网格扫描 (0.105,0.003,3) 0.024 68.8
单纯形搜索 (0.090,-0.014,3) 0.023 9.2
从表2及图11b和图11c可以看出，消除多普勒效应后的定位精度较高， 且单纯形搜索的定位方法比网格扫描的定位方法的用时短，定位精度高，实现了快速有效的定位。
本领域普通技术人员可以理解，上述方法中的全部或部分步骤可通过程序来指令相关硬件(例如处理器)完成，所述程序可以存储于计算机可读存储介质中，如只读存储器、磁盘或光盘等。可选地，上述实施例的全部或部分步骤也可以使用一个或多个集成电路来实现。相应地，上述实施例中的各模块/单元可以采用硬件的形式实现，例如通过集成电路来实现其相应功能，也可以采用软件功能模块的形式实现，例如通过处理器执行存储于存储器中的程序/指令来实现其相应功能。本发明实施例不限制于任何特定形式的硬件和软件的结合。
以上公开内容规定为说明性的而不是穷尽性的。对于本领域技术人员来说，本说明书将暗示许多变化和可选择方案。所有这些可选择方案和变化规定为被包括在本权利要求的范围内。本领域技术人员应认识到此处所述的实施方案的其它等效变换，这些等效变换也规定为由本权利要求所包括。
在此完成了对本申请可选择的实施方案的描述。本领域技术人员可认识到此处所述的实施例仅用于说明本申请，其中方法的各实施步骤等都是可以有所变化的，凡是在本申请技术方案的基础上进行的等同变换和改进，均不应排除在本申请的保护范围之外。
工业实用性
本申请实施例利用较少的传声器，能够对运动声源消除多普勒效应的影响，对声源进行快速有效的定位。突破传统运动声源识别中对大传声器阵列的依赖，同时突破声达时差法只能用于静止声源定位的限制，为各种运动型短时发声声源的快速识别提供一种新的方法。

Claims (10)

  1. 一种运动声源的定位方法，包括：
    将至少四个传声器布置成传声器阵列；
    用所述传声器阵列采集运动声源的声压信号，并建立声源搜索面；
    基于所述声源搜索面上的搜索点，利用声达时间差的声源定位方法得出运动声源的估计位置，并计算所述运动声源的估计位置与对应的搜索点之间的偏差值；
    得出所述偏差值最小时所对应的搜索点的位置，将所述搜索点的位置作为运动声源的位置。
  2. 如权利要求1所述的定位方法，其中所述基于所述声源搜索面上的搜索点，利用声达时间差的声源定位方法得出运动声源的估计位置，并计算所述运动声源的估计位置与对应的搜索点之间的偏差值的步骤包括：
    对所述声源搜索面进行离散化，得到一系列的网格点，所述网格点即为所述声源搜索面上的搜索点，并基于所有搜索点进行偏差值计算。
  3. 如权利要求1所述的定位方法，其中在所述基于所述声源搜索面上的搜索点，利用声达时间差的声源定位方法得出运动声源的估计位置，并计算所述运动声源的估计位置与对应的搜索点之间的偏差值的步骤中，利用单纯形法，确定所述声源搜索面上的搜索点。
  4. 如权利要求1-3中任一项所述的定位方法，其中在所述基于所述声源搜索面上的搜索点，利用声达时间差的声源定位方法得出运动声源的估计位置，并计算所述运动声源的估计位置与对应的搜索点之间的偏差值的步骤中，
    在所述利用声达时间差的声源定位方法得出运动声源的估计位置之前，对所述声压信号进行多普勒效应消除处理。
  5. 如权利要求1-4中任一项所述的定位方法，其中在所述将至少四个传声器布置成传声器阵列的步骤中，将所述至少四个传声器布置在同一平面上。
  6. 如权利要求5所述的定位方法，其中所述传声器阵列为包括四个传声器的平面四元阵列。
  7. 如权利要求5或6所述的定位方法，其中所述传声器阵列所在的平面与所述声源搜索面平行。
  8. 一种运动声源的定位装置，包括：
    至少四个传声器，布置成传声器阵列，所述传声器阵列设置成用于采集运动声源的声压信号；
    处理装置，包括：
    用于接收所述声压信号的接收模块，
    用于建立声源搜索面的模块，
    用于基于所述声源搜索面上的搜索点，利用声达时间差的声源定位方法得出运动声源的估计位置，并计算所述运动声源的估计位置与对应的搜索点之间的偏差值的模块，和
    用于得出所述偏差值最小时所对应的搜索点的位置，并将该搜索点的位置输出的模块；以及
    显示装置，设置成用于输出的显示搜索点的位置。
  9. 如权利要求8所述的定位装置，其中所述处理装置还包括：
    对所述声压信号进行多普勒效应消除处理的模块。
  10. 如权利要求8或9所述的定位装置，其中所述至少四个传声器布置在同一平面上。
PCT/CN2016/113108 2016-12-29 2016-12-29 一种运动声源的定位方法和定位装置 WO2018119910A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2016/113108 WO2018119910A1 (zh) 2016-12-29 2016-12-29 一种运动声源的定位方法和定位装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2016/113108 WO2018119910A1 (zh) 2016-12-29 2016-12-29 一种运动声源的定位方法和定位装置

Publications (1)

Publication Number Publication Date
WO2018119910A1 true WO2018119910A1 (zh) 2018-07-05

Family

ID=62710841

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/113108 WO2018119910A1 (zh) 2016-12-29 2016-12-29 一种运动声源的定位方法和定位装置

Country Status (1)

Country Link
WO (1) WO2018119910A1 (zh)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109471145A (zh) * 2018-10-17 2019-03-15 中北大学 一种基于四元声阵列平台的双声定位定姿法
CN109525929A (zh) * 2018-10-29 2019-03-26 中国传媒大学 一种录音定位方法及装置
CN109633554A (zh) * 2019-01-18 2019-04-16 浙江大学 基于概率数据关联的移动声源到达时延估计方法
CN110095755A (zh) * 2019-04-01 2019-08-06 北京云知声信息技术有限公司 一种声源定位方法
CN110673097A (zh) * 2019-10-24 2020-01-10 浙江工商大学 利用音频到达时间差对空间中发声物体的定位方法
CN113702910A (zh) * 2021-08-31 2021-11-26 冠捷显示科技（厦门）有限公司 一种基于双麦克风的声音定位方法及其装置
CN115278449A (zh) * 2022-09-26 2022-11-01 中国飞行试验研究院 确定麦克风阵列单元坐标的方法、装置、设备及存储介质

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101413824A (zh) * 2008-12-04 2009-04-22 清华大学 一种基于随机传声器阵列的运动物体声场测量方法

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101413824A (zh) * 2008-12-04 2009-04-22 清华大学 一种基于随机传声器阵列的运动物体声场测量方法

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
MIAO , FENG: "Moving sound source localization based on triangulation method", JOURNAL OF SOUND AND VIBRATION, vol. 385, 20 September 2016 (2016-09-20) - 22 December 2016 (2016-12-22), pages 94 - 99, XP029758319, ISSN: 0022-460X, DOI: 10.1016/j.jsv.2016.09.001 *
MIAO, FENG: "A moving sound source localization method based on TDOA", INTER. NOISE 2014, 16 November 2014 (2014-11-16) - 19 November 2014 (2014-11-19), Melbourne AUSTRALIA, XP055605663 *
ZHANG, KAI ET AL.: "SOUND SOURCE DETECTION USING A FAST MOVING TIME DIFFERENCE OF ARRIVAL METHOD", 23RD INTERNATIONAL CONGRESS ON SOUND & VIBRATION, 14 July 2016 (2016-07-14) *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109471145A (zh) * 2018-10-17 2019-03-15 中北大学 一种基于四元声阵列平台的双声定位定姿法
CN109525929A (zh) * 2018-10-29 2019-03-26 中国传媒大学 一种录音定位方法及装置
CN109633554A (zh) * 2019-01-18 2019-04-16 浙江大学 基于概率数据关联的移动声源到达时延估计方法
CN110095755A (zh) * 2019-04-01 2019-08-06 北京云知声信息技术有限公司 一种声源定位方法
CN110095755B (zh) * 2019-04-01 2021-03-12 云知声智能科技股份有限公司 一种声源定位方法
CN110673097A (zh) * 2019-10-24 2020-01-10 浙江工商大学 利用音频到达时间差对空间中发声物体的定位方法
CN110673097B (zh) * 2019-10-24 2022-01-18 浙江工商大学 利用音频到达时间差对空间中发声物体的定位方法
CN113702910A (zh) * 2021-08-31 2021-11-26 冠捷显示科技（厦门）有限公司 一种基于双麦克风的声音定位方法及其装置
CN115278449A (zh) * 2022-09-26 2022-11-01 中国飞行试验研究院 确定麦克风阵列单元坐标的方法、装置、设备及存储介质
CN115278449B (zh) * 2022-09-26 2023-03-10 中国飞行试验研究院 确定麦克风阵列单元坐标的方法、装置、设备及存储介质

Similar Documents

Publication Publication Date Title
WO2018119910A1 (zh) 一种运动声源的定位方法和定位装置
Nunes et al. A steered-response power algorithm employing hierarchical search for acoustic source localization using microphone arrays
US8363848B2 (en) Method, computer readable storage medium and system for localizing acoustic source
Dorfan et al. Tree-based recursive expectation-maximization algorithm for localization of acoustic sources
Ribeiro et al. Geometrically constrained room modeling with compact microphone arrays
US9069065B1 (en) Audio source localization
Ba et al. L1 regularized room modeling with compact microphone arrays
Pertilä et al. Passive self-localization of microphones using ambient sounds
KR102097641B1 (ko) 구형 마이크로폰 어레이를 이용한 음원의 입사 방향 추정방법
Dang et al. A feature-based data association method for multiple acoustic source localization in a distributed microphone array
Saqib et al. Sound-based distance estimation for indoor navigation in the presence of ego noise
Spencer Closed-form analytical solutions of the time difference of arrival source location problem for minimal element monitoring arrays
Åström et al. Extension of time-difference-of-arrival self calibration solutions using robust multilateration
KR20090128221A (ko) 음원 위치 추정 방법 및 그 방법에 따른 시스템
Tuna et al. 3D room geometry inference using a linear loudspeaker array and a single microphone
CN114167356A (zh) 一种基于多面体麦克风阵列的声源定位方法和系统
Padois et al. On the use of geometric and harmonic means with the generalized cross-correlation in the time domain to improve noise source maps
JP2005049153A (ja) 音声方向推定装置及びその方法
CN108872939B (zh) 基于声学镜像模型的室内空间几何轮廓重构方法
El Baba et al. Reflector localization based on multiple reflection points
Larsson et al. Fast and robust stratified self-calibration using time-difference-of-arrival measurements
JP6650245B2 (ja) インパルス応答生成装置及びプログラム
US20230244244A1 (en) Information processing apparatus, information processing method, and program
WO2022219558A9 (en) System and method for estimating direction of arrival and delays of early room reflections
Dang et al. Multiple sound source localization based on a multi-dimensional assignment model

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16925604

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16925604

Country of ref document: EP

Kind code of ref document: A1