技术领域technical field
本发明属于误差分析领域，具体涉及到嵌入式声传感器阵列系统在进行远场声源测向时存在角度误差的误差分析装置及方法The invention belongs to the field of error analysis, and in particular relates to an error analysis device and method for an angle error when an embedded acoustic sensor array system performs direction finding of a far-field sound source
背景技术Background technique
声源测向技术因其具有很强的抗电磁干扰能力，在实际应用中有着广泛的研究基础。在智能传感逐步深入实际生活的现如今，从消费电子方面到军事系统方面等，智能声传感器阵列已经应用在诸多领域。受电路微型化的影响，对嵌入式阵列系统的使用越来越多。嵌入式系统不仅包括对传统系统的集成，还包括对于集成后的小型系统的封装。对系统的集成，节约了系统的所占空间，使得声传感器间距相比于传统的声传感器阵列要小的多，可忽略声音幅度的衰减，但增大了测向的角度误差。早在几年前就有国内学者提出了关于MEMS声传感器在封装后测向出现误差的问题(Sound source localization based ondirectivity of MEMS microphones,Solid-State and Integrated CircuitsTechnology,Proceedings.7th International Conference on,Xiaoming Wu；TianlingRen；Litian Liu)，但是文章分析特征单一，物理意义也并不明确。不能反映出问题存在真正原因。Sound source direction finding technology has a broad research basis in practical applications because of its strong anti-electromagnetic interference ability. Now that smart sensing is gradually deepening into real life, from consumer electronics to military systems, smart acoustic sensor arrays have been used in many fields. Influenced by the miniaturization of circuits, the use of embedded array systems is increasing. Embedded systems include not only the integration of traditional systems, but also the packaging of integrated small systems. The integration of the system saves the space occupied by the system, making the distance between the acoustic sensors much smaller than that of the traditional acoustic sensor array, and the attenuation of the sound amplitude can be ignored, but the angular error of the direction finding is increased. As early as a few years ago, domestic scholars raised the issue of direction finding errors in MEMS acoustic sensors after packaging (Sound source localization based on directivity of MEMS microphones, Solid-State and Integrated Circuits Technology, Proceedings. 7th International Conference on, Xiaoming Wu ; TianlingRen; Litian Liu), but the analysis features of the article are single, and the physical meaning is not clear. It cannot reflect the real cause of the problem.
为了防止空气中的杂质、灰尘等对电路的腐蚀而造成系统的性能下降，必须保证系统器件与外界隔离，就需要对系统进行封装。封装后的系统留有各种接口，方便与其他器件的连接。然而嵌入在封装外壳内的声传感器阵列，由于受到微型化系统及封装外壳的影响，对声源方向的估计存在一定的误差。并且声波在传播过程中，受封装外壳的影响，沿外壳爬行，使得延时偏大。In order to prevent the performance of the system from being corroded by impurities, dust, etc. in the air, it is necessary to ensure that the system components are isolated from the outside world, and the system needs to be packaged. The packaged system has various interfaces for easy connection with other devices. However, due to the influence of the miniaturized system and the package shell, the acoustic sensor array embedded in the package shell has certain errors in the estimation of the direction of the sound source. Moreover, during the propagation process, the sound wave is affected by the packaging shell and crawls along the shell, which makes the delay longer.
但是现有技术中尚无一种基于嵌入式阵列的误差分析装置及方法。However, there is no error analysis device and method based on an embedded array in the prior art.
发明内容Contents of the invention
本发明的目的在于提供一种基于嵌入式阵列的误差分析装置及方法。The object of the present invention is to provide an error analysis device and method based on an embedded array.
实现本发明目的的技术解决方案为：一种基于嵌入式阵列的误差分析装置，包括发声装置、声音采集装置和PC机，其中：The technical solution that realizes the object of the present invention is: a kind of error analysis device based on embedded array, comprises sound generating device, sound collection device and PC, wherein:
发声装置用于发出声音；The sound generating device is used to emit sound;
声音采集装置用于采集发声装置发出的声音，该装置与发声装置的高度一致，相距10～15m；The sound collection device is used to collect the sound from the sound emitting device, and the height of the device is the same as that of the sound emitting device, with a distance of 10-15m;
PC机用于接收声音采集装置采集到的音频信号并对其进行后期处理。The PC is used to receive the audio signal collected by the sound collection device and perform post-processing on it.
一种基于上述误差分析装置的误差分析方法，包括以下步骤：An error analysis method based on the above-mentioned error analysis device, comprising the following steps:
步骤1、利用声音采集装置采集发声装置发出的音频信号，在采集过程中转动声音采集装置，从而采集来自不同方向的音频信号；Step 1. Utilize the sound collection device to collect the audio signal sent by the sound generating device, and rotate the sound collection device during the collection process, thereby collecting audio signals from different directions;
步骤2、对采集到的信号进行预处理，具体是对信号进行短时分帧、端点检测处理，从而提炼出所需音频片段；Step 2. Preprocessing the collected signal, specifically performing short-time frame division and endpoint detection processing on the signal, so as to extract the required audio clip;
步骤3、对步骤2提炼出的音频片段进行测向分析，得出声源方向；Step 3. Perform direction finding analysis on the audio clip extracted in step 2 to obtain the direction of the sound source;
步骤4、计算声源测向的角度误差，得出误差分布规律。Step 4. Calculate the angle error of the direction finding of the sound source, and obtain the error distribution law.
本发明与现有技术相比，其显著优点为：1)本发明的方法考虑到各种影响因素，使用了较为全面客观的数据；2)本发明的方法通过具体数据总结误差分布规律得出嵌入式声传感器阵列中存在的声波绕道传输现象，能够适用于其它各种嵌入式声传感器阵列结构；3)本发明的方法数据可靠，易于实施，操作简单；4)本发明的装置结构简单，便于搭建。Compared with the prior art, the present invention has the remarkable advantages of: 1) the method of the present invention takes into account various influencing factors, and uses comparatively comprehensive and objective data; 2) the method of the present invention draws The sound wave detour transmission phenomenon existing in the embedded acoustic sensor array can be applicable to other various embedded acoustic sensor array structures; 3) the method data of the present invention is reliable, easy to implement, simple to operate; 4) the device structure of the present invention is simple, Easy to build.
下面结合附图对本发明作进一步详细描述。The present invention will be described in further detail below in conjunction with the accompanying drawings.
附图说明Description of drawings
图1是本发明使用的嵌入式阵列的装置拓扑图。Fig. 1 is a device topology diagram of an embedded array used in the present invention.
图2是本发明基于嵌入式阵列的误差分析流程图。Fig. 2 is a flow chart of the error analysis based on the embedded array of the present invention.
图3是本发明对角度误差分析绘制的统计图，其中图(a)为角度误差分布直方图，图(b)为角度误差随声源方向改变的误差均值分布图，图(c)为修正后的角度误差随声源方向改变的误差均值分布图，图(d)为延时误差随声源方向改变的分布规律图。Fig. 3 is the statistical chart that the present invention draws to angle error analysis, and wherein figure (a) is the angle error distribution histogram, and figure (b) is the error average value distribution figure that angle error changes with sound source direction, and figure (c) is correction Figure (d) is the distribution pattern of the delay error changing with the direction of the sound source.
图4是本发明对角度误差产生原因进行分析的流程图。Fig. 4 is a flow chart of the present invention analyzing the cause of the angle error.
具体实施方式Detailed ways
结合图1，本发明的一种基于嵌入式阵列的误差分析装置，包括发声装置、声音采集装置1和PC机，其中：In conjunction with Fig. 1, a kind of error analysis device based on embedded array of the present invention comprises sounding device, sound collecting device 1 and PC, wherein:
发声装置用于发出声音；The sound generating device is used to emit sound;
声音采集装置用于采集发声装置发出的声音，该装置与发声装置的高度一致，相距10～15m；The sound collection device is used to collect the sound from the sound emitting device, and the height of the device is the same as that of the sound emitting device, with a distance of 10-15m;
PC机用于接收声音采集装置采集到的音频信号并对其进行后期处理。The PC is used to receive the audio signal collected by the sound collection device and perform post-processing on it.
所述声音采集装置1包括扁平方形外壳、四个微型声传感器、内置电池和可编程处理板，其中四个微型声传感器、内置电池和可编程处理板均位于扁平方形外壳壳体内，所述扁平方形外壳为上下两层结构，分别为上壳体和下壳体，上壳体内设置四个微型声传感器、内置电池，下壳体内设置可编程处理板，该可编程处理板通过通讯接口将数据传送至PC机。The sound collection device 1 includes a flat square housing, four miniature acoustic sensors, built-in batteries and a programmable processing board, wherein the four miniature acoustic sensors, built-in batteries and a programmable processing board are all located in the flat square housing, and the flat The square shell has an upper and lower two-layer structure, which are the upper shell and the lower shell respectively. Four miniature acoustic sensors and built-in batteries are installed in the upper shell, and a programmable processing board is set in the lower shell. The programmable processing board transmits data through the communication interface. Send to PC.
所述四个微型声传感器位于同一高度，构成方形阵列，每个微型声传感器朝向壳体侧壁的外侧。The four miniature acoustic sensors are located at the same height to form a square array, and each miniature acoustic sensor faces the outside of the side wall of the casing.
一种基于上述误差分析装置的误差分析方法，包括以下步骤：An error analysis method based on the above-mentioned error analysis device, comprising the following steps:
步骤1、利用声音采集装置采集发声装置发出的音频信号，在采集过程中转动声音采集装置，从而采集来自不同方向的音频信号；转动声音采集装置的角度为0°～360°。Step 1. Use the sound collection device to collect the audio signal from the sound generating device, and rotate the sound collection device during the collection process to collect audio signals from different directions; the angle of rotation of the sound collection device is 0° to 360°.
步骤2、对采集到的信号进行预处理，具体是对信号进行短时分帧、端点检测处理，从而提炼出所需音频片段；Step 2. Preprocessing the collected signal, specifically performing short-time frame division and endpoint detection processing on the signal, so as to extract the required audio clip;
对信号进行短时分帧、端点检测处理具体为：The short-time framing and endpoint detection processing of the signal is as follows:
步骤2-1、采用帧重叠的方法进行短时分帧处理，每一帧数据作为一个样本；Step 2-1, using the method of frame overlapping to perform short-time frame division processing, and each frame of data is used as a sample;
步骤2-2、计算每一个样本的能量Ek，若能量Ek高于门限值E0，则执行步骤2-3，否则不处理，所述样本的能量Ek的确定公式为：Step 2-2. Calculate the energy Ek of each sample. If the energy Ek is higher than the threshold value E0 , execute step 2-3; otherwise, do not process. The formula for determining the energy Ek of the sample is:
其中，x表示每帧信号，k表示帧数，l表示每帧中的各个样本点序号，N为每帧中样本点总数；Among them, x represents the signal of each frame, k represents the number of frames, l represents the serial number of each sample point in each frame, and N is the total number of sample points in each frame;
步骤2-3、计算短时过零率Z，若Z在门限范围内，表示检测到声源信号，否则不是声源信号，所述短时过零率Z的确定公式为：Step 2-3, calculate the short-term zero-crossing rate Z, if Z is within the threshold range, it means that the sound source signal is detected, otherwise it is not a sound source signal, the determination formula of the short-term zero-crossing rate Z is:
其中，sgn[]表示符号函数。Among them, sgn[] represents the sign function.
步骤3、对步骤2提炼出的音频片段进行测向分析，得出声源方向；对音频片段进行测向分析，具体为：Step 3. Perform direction-finding analysis on the audio segment extracted in step 2 to obtain the direction of the sound source; perform direction-finding analysis on the audio segment, specifically:
步骤3-1、对提炼出所需音频片段xi(n)进行快速傅里叶变换，得到的频域信号记为Xi(ω)，i＝1,2,3...为声传感器的个数，n为整数，表示在信号中的点数序号，ω表示角频率；Step 3-1. Perform fast Fourier transform on the extracted desired audio segment xi (n), and the obtained frequency domain signal is denoted as Xi (ω),i =1,2,3...is the acoustic sensor The number of , n is an integer, indicating the number of points in the signal, ω indicates the angular frequency;
步骤3-2、确定四个微型声传感器中两两的互相关函数所述互相关函数的确定公式为：Step 3-2, determine the cross-correlation function of any two of the four miniature acoustic sensors The cross-correlation function The formula for determining is:
式中，Gij(ω)为声传感器i和声传感器j获取信号的互相关功率谱，其公式为：Gij(ω)＝Xi(ω)Xj*(ω)，Xi(ω)是频域信号，Xj*(ω)是信号在频域内的共轭，Ψij(ω)为加权函数，Ψij(ω)Gij(ω)为广义互相关功率谱；In the formula, Gij (ω) is the cross-correlation power spectrum of the signals acquired by acoustic sensor i and acoustic sensor j, and its formula is: Gij (ω)=Xi (ω)Xj* (ω),Xi (ω ) is the frequency domain signal, Xj* (ω) is the conjugate of the signal in the frequency domain, Ψij (ω) is the weighting function, Ψij (ω)Gij (ω) is the generalized cross-correlation power spectrum;
步骤3-3、对互相关函数做傅里叶逆变换后可得到时间延迟τij，所述时间延迟τij的确定公式为：Step 3-3, pair cross-correlation function The time delay τij can be obtained after inverse Fourier transform, and the formula for determining the time delay τij is:
其中，τij表示使得取最大值时对应的位置；Among them, τij represents such that The corresponding position when taking the maximum value;
步骤3-4、构建时间-角度模型，具体为：Step 3-4, building a time-angle model, specifically:
其中，m为做互相关变换的组数，θ为声源实际所在方向角度，为声传感器到坐标原点的向量，θ[0°,360°]为与x轴正方向之间的夹角，为声源坐标，取声音从声源s到坐标原点所用时间为0时刻，τij为声传感器之间的时间延迟，c为声速；Among them, m is the number of groups for cross-correlation transformation, θ is the actual direction angle of the sound source, is the vector from the acoustic sensor to the coordinate origin, θ[0°,360°] is The included angle with the positive direction of the x-axis, is the coordinate of the sound source, the time taken by the sound from the sound source s to the origin of the coordinate is taken as 0 moment,τij is the time delay between the acoustic sensors, and c is the speed of sound;
步骤3-5、将步骤3-3得到的时间延迟τij带入步骤3-4所构建时间-角度模型，从而得到声源实际所在方向角度θ。Step 3-5, bringing the time delay τij obtained in step 3-3 into the time-angle model constructed in step 3-4, so as to obtain the actual direction angle θ of the sound source.
步骤4、计算声源测向的角度误差，得出误差分布规律。具体为：Step 4. Calculate the angle error of the direction finding of the sound source, and obtain the error distribution law. Specifically:
步骤4-1、确定步骤3得到的声源实际所在方向角度θ与理论声源所在方向角度θ0之间的角度误差Δθ，其确定公式为：Step 4-1, determine the angle error Δθ between the actual direction angle θ of the sound source obtained in step 3 and the direction angle θ0 of the theoretical sound source, the determination formula is:
Δθ＝θ-θ0；Δθ=θ-θ0 ;
步骤4-2、将理论声源所在方向角度θ0带入上述时间-角度模型，得出理论时间延迟τ'ij，之后计算实际时间延迟τij与理论时间延迟τ'ij之间的延时误差Δτij，其确定公式为：Step 4-2. Bring the theoretical sound source direction angle θ0 into the above time-angle model to obtain the theoretical time delay τ'ij , and then calculate the delay between the actual time delay τij and the theoretical time delay τ'ij Error Δτij , its determination formula is:
Δτij＝τij-τ'ij；Δτij =τij -τ'ij;
步骤4-3、转动声音采集装置，重复步骤4-1～步骤4-2，得到角度误差Δθ和延时误差Δτij的数据集，并对其进行统计分析，得出误差分布规律。Step 4-3. Rotate the sound collection device, repeat steps 4-1 to 4-2, obtain the data sets of angle error Δθ and delay error Δτij , and perform statistical analysis on them to obtain the error distribution rule.
下面结合实施例进行更详细的描述：Below in conjunction with embodiment carry out more detailed description:
针对嵌入式阵列系统伴随的误差问题，本发明提出了一种基于嵌入式阵列的误差分析方法。该方法适用于各种嵌入式阵列系统。具体使用了一种嵌入式声传感器阵列系统，在真实声场环境下录取声音信号，使用基于到达时间差的声源定位算法来估计声源方向，并与理论声源方向进行对比。分析系统在各种不同工作环境下测向的角度误差，得出待测声源信号的角度误差的分布规律，分析误差产生原因，通过修正角度误差提高嵌入式声传感器阵列的测向准确度，使系统能在实际环境下实现高准确度测向性能。Aiming at the error problem accompanying the embedded array system, the present invention proposes an error analysis method based on the embedded array. This method is applicable to various embedded array systems. Specifically, an embedded acoustic sensor array system is used to record sound signals in a real sound field environment, and a sound source localization algorithm based on time difference of arrival is used to estimate the sound source direction, and compared with the theoretical sound source direction. Analyze the angle error of the direction finding of the system in various working environments, obtain the distribution law of the angle error of the sound source signal to be measured, analyze the cause of the error, and improve the direction finding accuracy of the embedded acoustic sensor array by correcting the angle error. This enables the system to achieve high-accuracy direction-finding performance in actual environments.
结合图1所示的嵌入式阵列的装置拓扑图，本发明所用的嵌入式阵列系统为一种嵌入式声传感器阵列系统，其构成包括：包括扁平方形外壳、四个微型声传感器、内置电池和可编程处理板，其中四个微型声传感器、内置电池和可编程处理板均位于扁平方形外壳壳体内，所述扁平方形外壳为上下两层结构，分别为上壳体和下壳体，上壳体内设置四个微型声传感器、内置电池，下壳体内设置可编程处理板。通过通讯接口将数据传送至PC端，通过开关控制嵌入式系统的电源供电。In conjunction with the device topology diagram of the embedded array shown in Figure 1, the embedded array system used in the present invention is an embedded acoustic sensor array system, and its composition includes: a flat square shell, four miniature acoustic sensors, a built-in battery and A programmable processing board, wherein four miniature acoustic sensors, a built-in battery and a programmable processing board are all located in a flat square shell. Four miniature acoustic sensors and a built-in battery are arranged in the body, and a programmable processing board is arranged in the lower casing. The data is transmitted to the PC through the communication interface, and the power supply of the embedded system is controlled through the switch.
结合图2，本发明的一种基于嵌入式阵列的误差分析方法，步骤如下：In conjunction with Fig. 2, a kind of error analysis method based on embedded array of the present invention, the steps are as follows:
第一步，声源信号的采集：发声器发出的是高斯白噪声，其频率为48kHz，使用带有四个微型声传感器的嵌入式阵列用来获取声源信号，测得当声源方向来自不同方向的数据，从0度开始，按逆时针方向每隔30°改变声源方向，再进行录取，直至录取一个圆周的声源信号。The first step is the acquisition of the sound source signal: the sound generator emits Gaussian white noise with a frequency of 48kHz. An embedded array with four miniature acoustic sensors is used to obtain the sound source signal. When the sound source direction is measured from different For direction data, start from 0 degrees, change the sound source direction every 30° counterclockwise, and then record until a circular sound source signal is recorded.
第二步，对声源信号的预处理，首先读取用上述专用测量装置录制的含有声源信号的数据，然后将数据分割成样本，并对声源信号进行端点检测，将无声源信号数据滤除，具体内容如下：The second step is the preprocessing of the sound source signal. First, read the data containing the sound source signal recorded by the above-mentioned special measuring device, and then divide the data into samples, and perform endpoint detection on the sound source signal, and convert the data without the sound source signal to Filter out, the specific content is as follows:
(1)将四个声传感器采集的四个通道信号进行短时分帧，采用帧重叠的方法，避免失真。每一帧数据作为一个样本；(1) The four channel signals collected by the four acoustic sensors are divided into frames in a short time, and the method of frame overlapping is adopted to avoid distortion. Each frame of data is used as a sample;
(2)计算短时帧能量若Ek高于门限值E0，则计算短时过零率；(2) Calculate short-time frame energy If Ek is higher than the threshold value E0 , calculate the short-term zero-crossing rate;
(3)计算短时过零率若Z在门限范围内，表示检测到声源信号。(3) Calculate the short-term zero-crossing rate If Z is within the threshold range, it means that a sound source signal is detected.
第三步，对声源信号进行测向分析，采用基于到达时间差的方法，得到声传感器之间的时间延迟，建立声源测向方程组，得出声源方向。具体内容如下：The third step is to analyze the direction of the sound source signal, and use the method based on the arrival time difference to obtain the time delay between the acoustic sensors, establish the sound source direction finding equation group, and obtain the sound source direction. The specific content is as follows:
(1)对获得的四路数字信号分别记为x1(n)，x2(n)，x3(n)，x4(n)，对经过分帧和端点检测的信号做快速傅里叶变换，得到的频域信号分别为X1(ω)，X2(ω)，X3(ω)，X4(ω)，ω表示角频率。(1) Denote the obtained four-way digital signals as x1 (n), x2 (n), x3 (n), x4 (n), and perform fast Fourier on the signals after framing and endpoint detection Leaf transformation, the obtained frequency domain signals are respectively X1 (ω), X2 (ω), X3 (ω), X4 (ω), where ω represents the angular frequency.
(2)利用基于相位谱的广义互相关算法得到声传感器之间的时间延迟，即声音的到达时间差，分别计算由(1)得到的频域信号X1(ω)，X2(ω)，X3(ω)，X4(ω)两两之间的互相关函数，共6组，分别为X1(ω)与X2(ω)、X1(ω)与X3(ω)、X1(ω)与X4(ω)、X2(ω)与X3(ω)、X2(ω)与X4(ω)、X3(ω)与X4(ω)，计算公式为：(2) Use the generalized cross-correlation algorithm based on the phase spectrum to obtain the time delay between the acoustic sensors, that is, the arrival time difference of the sound, and calculate the frequency domain signals X1 (ω), X2 (ω) obtained by (1) respectively, The cross-correlation function between X3 (ω) and X4 (ω), there are 6 groups in total, namely X1 (ω) and X2 (ω), X1 (ω) and X3 (ω), X1 (ω) and X4 (ω), X2 (ω) and X3 (ω), X2 (ω) and X4 (ω), X3 (ω) and X4 (ω), calculation formula for:
其中，Gij(ω)＝Xi(ω)Xj*(ω)，Gij(ω)为声传感器i和声传感器j获取信号的互相关功率谱，Xj*(ω)是信号在频域内的共轭。Ψij(ω)为加权函数，在此，Ψij(ω)Gij(ω)即广义互相关功率谱，将其做傅里叶逆变换后得到时间延迟τij，即使得取最大值时对应的位置，就是声传感器i和声传感器j的声音到达时间差。Among them, Gij (ω)=Xi (ω)Xj* (ω), Gij (ω) is the cross-correlation power spectrum of the signals acquired by acoustic sensor i and acoustic sensor j, Xj* (ω) is the signal at Conjugation in the frequency domain. Ψij (ω) is a weighting function, here, Ψij (ω)Gij (ω) is the generalized cross-correlation power spectrum, which is inversely Fourier transformed to obtain the time delay τij , even if The position corresponding to the maximum value is the sound arrival time difference between the acoustic sensor i and the acoustic sensor j.
(3)以声传感器阵列平面建立平面直角坐标系，四个声传感器处于四个象限内，设声传感器到坐标原点的向量为声源位置到坐标原点的向量为θ[0°,360°]为与x轴正方向之间的夹角，取声音从声源s到坐标原点所用时间为0时刻，TDOAi为声传感器相对于0时刻的时间延迟，τij为声传感器之间的时间延迟，c为声速。为向量点乘，表示向量在上的投影，根据速度公式有：(3) A rectangular coordinate system is established on the plane of the acoustic sensor array, the four acoustic sensors are in four quadrants, and the vector from the acoustic sensor to the origin of coordinates is The vector from the sound source position to the coordinate origin is θ[0°,360°] is The angle between the positive direction of the x-axis and the time taken by the sound from the sound source s to the origin of the coordinates is 0 time, TDOAi is the time delay of the acoustic sensor relative to 0 time, τij is the time delay between the acoustic sensors, c is the speed of sound. is the vector dot product, representing the vector exist The projection on , according to the velocity formula:
联立上述方程式可以得到如下矩阵：Combining the above equations, the following matrix can be obtained:
第四步，对数据做统计分析，计算声传感器之间的延时误差和声源测向的角度误差，进行统计学分析。通过多次试验，每次试验中，当声源方向在θ[0°,360°]间以30°变化时，每个角度采集多组声音，通过上述专利要求得出四个声传感器之间的时延和声源方向的估计值，对延时误差和角度误差进行统计。The fourth step is to perform statistical analysis on the data, calculate the delay error between the acoustic sensors and the angle error of the direction finding of the sound source, and perform statistical analysis. Through multiple tests, in each test, when the direction of the sound source changes at 30° between θ[0°, 360°], multiple groups of sounds are collected at each angle, and the four acoustic sensors are obtained through the above patent requirements. The time delay and the estimated value of the sound source direction are calculated, and the delay error and angle error are counted.
对每次的实验结果，观察角度误差的分布情况；当声源处于某一角度时，对多次实验结果进行角度误差分析，计算角度误差的均值和均方根，得出角度误差与声源的来向的关系及角度误差的波动情况。For each experimental result, observe the distribution of the angle error; when the sound source is at a certain angle, analyze the angle error of multiple experimental results, calculate the mean value and root mean square of the angle error, and obtain the relationship between the angle error and the sound source The relationship between the coming and going and the fluctuation of the angle error.
结合图3、图4，是本发明对延时误差和角度误差分析绘制的统计图及误差原因分析的流程图，可以看出声源测向出现的角度误差分布特征及其与声源所在方向的关系、延时误差分布特征及其与声源所在方向的关系。取逆时针方向为正方向：In conjunction with Fig. 3 and Fig. 4, it is the statistical chart and the flow chart of error cause analysis drawn by the present invention for delay error and angle error analysis, and it can be seen that the angular error distribution characteristics of sound source direction finding and its relationship with the direction of sound source The relationship between the delay error distribution characteristics and its relationship with the direction of the sound source. Take the counterclockwise direction as the positive direction:
从图3d时延误差分布规律，分析得出，误差最大处为声源方向背对声传感器所在平面时。由图3，通过数据计算分析可以验证延时误差存在的原因是声波绕封装壳体爬行，使得实际传播距离大于理论传播距离，导致到达时间差变大，出现测向角度不精准。From the distribution law of time delay error in Figure 3d, it can be concluded from the analysis that the maximum error is when the direction of the sound source faces away from the plane where the acoustic sensor is located. From Figure 3, through data calculation and analysis, it can be verified that the reason for the delay error is that the sound wave crawls around the package shell, making the actual propagation distance greater than the theoretical propagation distance, resulting in a larger arrival time difference and inaccurate direction finding angle.
从图3a、图3b可以看出，当声源方向在0度到360度之间时，角度误差基本上是呈右偏，变化范围基本在0度到5度，其中2度左右居多，取角度误差均值的均方根值约为2度。可以有针对性的对角度误差进行-2度补偿的方式，结果如图3c。从图3c可以看出对误差进行补偿后，角度误差均值变小，提高了测向的准确度。It can be seen from Fig. 3a and Fig. 3b that when the direction of the sound source is between 0° and 360°, the angle error is basically to the right, and the range of variation is basically from 0° to 5°, and most of them are about 2°. The rms value of the angular error mean is about 2 degrees. The angle error can be compensated by -2 degrees in a targeted manner, and the result is shown in Figure 3c. It can be seen from Fig. 3c that after the error is compensated, the mean value of the angle error becomes smaller, which improves the accuracy of direction finding.
本发明的方法以真实声场为背景，以高斯白噪声作为声源，采用基于统计学的分析方法，分析嵌入式阵列系统在测向时存在误差的原因。通过本发明的方法，可以找到误差产生的原因及其规律，并有针对的对嵌入式系统的测向误差进行补偿，实现嵌入式系统的高准确度测向。因此，本发明可用于对嵌入式声传感器阵列中存在的声波绕道传输现象进行定量判断，能够适用于其它各种嵌入式声传感器阵列结构。The method of the invention takes the real sound field as the background, uses Gaussian white noise as the sound source, and adopts an analysis method based on statistics to analyze the cause of errors in the direction finding of the embedded array system. Through the method of the invention, the cause and law of error generation can be found, and the direction-finding error of the embedded system can be compensated in a targeted manner, so as to realize the high-accuracy direction-finding of the embedded system. Therefore, the present invention can be used for quantitatively judging the sound wave detour transmission phenomenon existing in the embedded acoustic sensor array, and can be applied to other various embedded acoustic sensor array structures.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610526960.0ACN105954710B (en) | 2016-07-06 | 2016-07-06 | A kind of error analysis device and method based on embedded Array |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610526960.0ACN105954710B (en) | 2016-07-06 | 2016-07-06 | A kind of error analysis device and method based on embedded Array |
| Publication Number | Publication Date |
|---|---|
| CN105954710A CN105954710A (en) | 2016-09-21 |
| CN105954710Btrue CN105954710B (en) | 2018-11-13 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610526960.0AActiveCN105954710B (en) | 2016-07-06 | 2016-07-06 | A kind of error analysis device and method based on embedded Array |
| Country | Link |
|---|---|
| CN (1) | CN105954710B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110231601A (en)* | 2019-07-01 | 2019-09-13 | 百度在线网络技术（北京）有限公司 | Sensor error compensation method, device, equipment and storage medium |
| CN112190259B (en)* | 2020-09-10 | 2024-06-28 | 北京济声科技有限公司 | Method for testing sound source positioning capability, tester terminal and subject terminal |
| CN115046145A (en)* | 2022-05-13 | 2022-09-13 | 常州大学 | An infrasound localization method for tiny leaks in pipelines |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102103200A (en)* | 2010-11-29 | 2011-06-22 | 清华大学 | Acoustic source spatial positioning method for distributed asynchronous acoustic sensor |
| CN104076331A (en)* | 2014-06-18 | 2014-10-01 | 南京信息工程大学 | Sound source positioning method for seven-element microphone array |
| CN104535965A (en)* | 2014-12-29 | 2015-04-22 | 江苏科技大学 | Parallelized sound source positioning system based on embedded GPU system and method |
| CN104991573A (en)* | 2015-06-25 | 2015-10-21 | 北京品创汇通科技有限公司 | Locating and tracking method and apparatus based on sound source array |
| US20150341719A1 (en)* | 2014-05-20 | 2015-11-26 | Cisco Technology, Inc. | Precise Tracking of Sound Angle of Arrival at a Microphone Array under Air Temperature Variation |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102103200A (en)* | 2010-11-29 | 2011-06-22 | 清华大学 | Acoustic source spatial positioning method for distributed asynchronous acoustic sensor |
| US20150341719A1 (en)* | 2014-05-20 | 2015-11-26 | Cisco Technology, Inc. | Precise Tracking of Sound Angle of Arrival at a Microphone Array under Air Temperature Variation |
| CN104076331A (en)* | 2014-06-18 | 2014-10-01 | 南京信息工程大学 | Sound source positioning method for seven-element microphone array |
| CN104535965A (en)* | 2014-12-29 | 2015-04-22 | 江苏科技大学 | Parallelized sound source positioning system based on embedded GPU system and method |
| CN104991573A (en)* | 2015-06-25 | 2015-10-21 | 北京品创汇通科技有限公司 | Locating and tracking method and apparatus based on sound source array |
| Title |
|---|
| A Low-Power MEMS Microphone Array for Wireless Acoustic Sensors;Geoffrey Ottoy,et al;《IEEE Sensors Applicaitons Symposium》;20160430;全文* |
| Sound Source Localization Based on Directivity of MEMS Microphones;Xiaoming WU，et al;《IEEE》;20041231;第1884-1886页* |
| 基于MEMS麦克风阵列的四通道语音采集系统设计;宫正;《电子技术与软件工程》;20141231;第133-134页* |
| Publication number | Publication date |
|---|---|
| CN105954710A (en) | 2016-09-21 |
| Publication | Publication Date | Title |
|---|---|---|
| CN104898091B (en) | Microphone array self-calibration sound source localization system based on iterative optimization algorithm | |
| CN104360315A (en) | LabVIEW-based (laboratory virtual instrumentation engineering workbench based) microphone array sound source localization method and device | |
| CN109188362B (en) | A kind of microphone array sound source localization signal processing method | |
| CN102305925A (en) | Robot continuous sound source positioning method | |
| CN101957442A (en) | Sound source positioning device | |
| CN103217211B (en) | Transformer substation noise source distribution measurement method based on synthetic aperture principle | |
| CN107942284B (en) | Method and Device for Underwater Direction of Arrival Estimation Based on Two-dimensional Orthogonal Non-Uniform Linear Array | |
| JP7187717B2 (en) | Method and apparatus for characterizing airflow | |
| CN105954710B (en) | A kind of error analysis device and method based on embedded Array | |
| CN109100685A (en) | A kind of passive acoustic direction blending algorithm of two-sided quaternary cross battle array | |
| CN105158734B (en) | A kind of single vector hydrophone Passive Location based on battle array invariant | |
| CN117706482B (en) | A single vector hydrophone target tracking method and system | |
| CN104614709A (en) | Acoustics and electromagnetism-based thunder positioning system and method | |
| CN105550433B (en) | A kind of capacitive micromachined ultrasonic sensor characteristics analysis method | |
| CN110398775A (en) | Method and system for picking up first-arrival signal fluctuations of microseismic events in tunnel water inrush disasters | |
| CN103792513A (en) | Thunder positioning system and method | |
| KR20080070196A (en) | Sound source direction detection system and method by estimating the correlation between sound source location and delay time difference | |
| CN107783135B (en) | A Method of Passive Ranging with Three-element Vector Array | |
| CN116540220A (en) | Target three-dimensional passive positioning method and device | |
| CN106526563A (en) | Quintuple volume array multi-target orientation estimation method based on cross-correlation virtual array | |
| CN105548824B (en) | A kind of electrical equipment partial discharge source positioner and localization method | |
| CN205562027U (en) | Two -dimensional vector field hydrophone | |
| Chen et al. | A sound source localization device based on rectangular pyramid structure for mobile robot | |
| CN104965194B (en) | Indoor multi-sound-source positioning device and method simulating binaural effect | |
| CN111077497A (en) | A device and method for sound source localization |
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information | Inventor after:Xu Zhiyong Inventor after:Wang Xu Inventor after:Zhao Zhao Inventor after:Hao Zhanxue Inventor after:Luo Lei Inventor before:Hao Zhanxue Inventor before:Luo Lei Inventor before:Xu Zhiyong Inventor before:Zhao Zhao Inventor before:Wang Xu | |
| CB03 | Change of inventor or designer information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |