Small space sound source positioning system and method based on three-axis acoustic sensor

Abstract

The invention is suitable for the technical field of sound source localization, and provides a small-space sound source localization system and method based on a triaxial acoustic sensor, and the system comprises a sensor module, an acquisition system, a processing system and an auxiliary module. The sensor module is a single three-axis acoustic vector sensor; the acquisition system comprises a signal conditioning circuit and a high-speed ADC (Analog to Digital Converter) and is used for conditioning and synchronously sampling signals acquired by the sensor module; the processing system comprises a data processing unit; the auxiliary module comprises a precision clock source and a temperature and humidity sensor, the precision clock source is used for ensuring the signal synchronization precision of each channel, and the temperature and humidity sensor is used for collecting environment temperature and humidity data in real time; according to the method, high-precision three-dimensional positioning can be realized only by a single sensor, the problems of complex deployment, high cost, difficult synchronization and the like of a traditional multi-sensor array system are solved, and the method is particularly suitable for industrial equipment fault diagnosis, abnormal sound source tracking and other small-space application scenes.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of sound source positioning technology, in particular to a small space sound source positioning system and method based on a three-axis acoustic sensor. BACKGROUND

[0002] Current small space sound source positioning technology mainly relies on a multi-sensor array system. The traditional three-dimensional space TDOA positioning theoretically requires at least four spatially separated sensors to form multiple independent TDOA measurement values to determine the three-dimensional space coordinates. In actual application, especially in the near field or considering other constraints, the number of sensors may need to be more.

[0003] However, this traditional scheme has the following defects in narrow space applications:

[0004] Firstly, the system deployment complexity is high, and the multi-point array of sensors needs accurate geometric calibration. The installation process is time-consuming and requires high technical requirements for the operator, which is particularly prominent in industrial sites with limited space;

[0005] Secondly, the system economy is poor. Multi-point synchronous acquisition requires independent signal conditioning circuits and high-precision clock synchronization modules, making it difficult to compress the cost of the entire system.

[0006] The above technical defects seriously restrict the application effect in small space scenarios such as device fault diagnosis and abnormal sound source tracking that require accurate position information. Especially in industrial sites with complex vibration noise environment and strict installation space, the existing scheme often fails to meet the comprehensive requirements of real-time, accuracy and economy.

[0007] Therefore, in view of the above status, it is urgent to provide a small space sound source positioning system and method based on a three-axis acoustic sensor to overcome the deficiencies in current practical applications. SUMMARY

[0008] The purpose of the present application is to provide a small space sound source positioning system and method based on a three-axis acoustic sensor, which effectively solves the problems in the background art.

[0009] The present application is implemented as follows: a small space sound source positioning system based on a three-axis acoustic sensor, comprising:

[0010] a sensor module, a collection system, a processing system and an auxiliary module;

[0011] The sensor module is a single three-axis acoustic vector sensor, which is provided with one sound pressure collection channel and three orthogonal vibration velocity collection channels for synchronous collection of sound pressure signals and three-dimensional particle vibration velocity components.

[0012] The acquisition system comprises a signal conditioning circuit and a high-speed ADC, which are used for conditioning and synchronous sampling of signals acquired by the sensor module.

[0013] The processing system comprises a data processing unit, which is used for real-time execution of an algorithm process of three-axial synchronous acquisition of sound pressure and vibration velocity signals, sound intensity vector calculation, sound source direction solution, radial vibration velocity projection, sound intensity distance analysis and three-dimensional coordinate generation.

[0014] The auxiliary module comprises a precise clock source and a temperature and humidity sensor, the precise clock source is used for ensuring the synchronization accuracy of the channels, and the temperature and humidity sensor is used for real-time acquisition of environmental temperature and humidity data to correct the atmospheric density and sound velocity parameters.

[0015] As a further scheme of the application, the three orthogonal vibration velocity acquisition channels of the three-axial acoustic vector sensor correspond to X, Y and Z three axes respectively, and the amplitude response and phase consistency error of each channel are controlled within a preset range.

[0016] The high-speed ADC transmits the signals acquired by the sensor module to the processing system through a low-jitter SPI interface.

[0017] As a further scheme of the application, the data processing unit adopts any one of FPGA, DSP, MCU, ASIC or CPU+GPU to meet the computing power requirement of sound intensity vector real-time calculation and frequency domain relationship analysis.

[0018] The application further provides a small-space sound source positioning method based on a three-axial acoustic sensor, which is applied to the system and comprises the following steps.

[0019] S1, three-axial synchronous acquisition of sound pressure and vibration velocity signals: the three-axial acoustic vector sensor synchronously acquires sound pressure signals p(t) and three-dimensional particle vibration velocity components v x (t), v y (t) and v z (t), and transmits the signals to the processing system after synchronous sampling by the high-speed ADC, and simultaneously performs time delay calibration by a known sound source in the system initialization stage, so that the time alignment error between the channels is controlled within 1 mu s;

[0020] S2, sound intensity vector calculation: the processing system calculates the sound intensity vector according to the sound pressure signals and the three-dimensional particle vibration velocity components according to the formula , wherein contains I x , I y and I z three components, and in the formula is the sound intensity vector, I x , I y and I zare the sound intensity components in X, Y, Z directions respectively, p(t) is the sound pressure signal, is the particle velocity vector, v x (t) is the particle velocity component in X direction, y (t) is the particle velocity component in Y direction, z (t) is the particle velocity component in Z direction;

[0021] S3, sound source direction calculation: based on the sound intensity vector components obtained in S2, the real-time sound source direction angles are calculated by azimuth angle calculation formula and pitch angle calculation formula, and then the sliding time window is used to process the mean value of the direction angles to obtain stable sound source direction (θ d ,φ d );

[0022] The azimuth angle is calculated as:

[0023] θ(t) = atan2(I y (t), I x (t));

[0024] where I x (t), I y (t) are the sound intensity components in X, Y directions at t moment;

[0025] The pitch angle is calculated as:

[0026]

[0027] where I x (t), I y (t), I z (t) are the sound intensity components in X, Y, Z directions at t moment;

[0028] The length of the sliding time window is set to 100 ms, corresponding to 4800 sampling points, and the mean value calculation formula is:

[0029]

[0030] where N is the window number, θ(t i ), φ(t i ) are the azimuth angle and pitch angle calculated by the i-th window respectively;

[0031] S4, radial velocity projection: according to the sound source direction (θ d ,φ d ) obtained in step S3, the sound source direction unit vector is constructed as:

[0032]

[0033] Then the three-dimensional particle velocity components are projected to the direction of the unit vector to obtain the radial velocity component:

[0034]

[0035] wherein, is the original three-dimensional velocity signal, is the unit vector of the sound source direction;

[0036] S5, sound intensity distance analysis: 4096-point FFT transform is performed on the sound pressure signal p(t) in step S1 and the radial velocity component v r (t) in step S4, to obtain the frequency spectrum P(f) and V r (f);

[0037] The distance is calculated according to the frequency point in the effective frequency band:

[0038]

[0039] wherein p is the air density, c is the sound velocity, both of which are based on standard environmental conditions, and the values of p and c can be corrected in real time through the environmental data collected by the temperature and humidity sensor;

[0040] The distances calculated for the full frequency band are weighted and averaged to obtain the final sound source distance r:

[0041]

[0042] S6, three-dimensional coordinate generation: according to the sound source direction (θ d , φ d ) obtained in step S3 and the sound source distance r obtained in step S5, the spherical coordinates are converted to Cartesian coordinates:

[0043]

[0044] As a further scheme of the present application: in S1, the sampling rate of the high-speed ADC is adjusted according to the sound source frequency range; when the sound source frequency range is wide, the sampling rate is increased to ensure the signal acquisition accuracy; when the sound source frequency range is narrow, the sampling rate is reduced to reduce the data processing amount.

[0045] As a further scheme of the present application: in S3, any one of the CORDIC algorithm, the table lookup method or the polynomial approximation method is used to solve the trigonometric functions in the azimuth angle and the pitch angle calculation.

[0046] As a further scheme of the present application: in S3, the length of the sliding time window is dynamically adjusted according to the environmental noise level; when the noise level is high, the length of the sliding time window is lengthened to improve the stability of the direction calculation; when the noise level is low, the length of the sliding time window is shortened to improve the positioning real-time performance.

[0047] As a further scheme of the present application: in S4, the radial velocity component v r (t) time, the fixed-point number in the Q15 or Q31 format is used to replace the decimal number to perform operation, and three multiply-add operations are completed in one clock cycle by using the single instruction multiple data stream technology, so that the calculation efficiency is improved.

[0048] As a further scheme of the present application: in S5, the effective frequency band is divided by using the 1 / 3 octave sub-band division or the equal bandwidth sub-band division, and the distance is calculated independently in each sub-band and then weighted and averaged.

[0049] As a further scheme of the present application: in S5, the FFT point number is selected according to the processor computing power and the accuracy requirement, when the processor computing power is strong and the distance resolution accuracy requirement is high, a larger FFT point number is selected, and when the processor computing power is limited and the real-time requirement is high, a smaller FFT point number is selected.

[0050] Compared with the prior art, the present application has the following beneficial effects:

[0051] In terms of hardware architecture, the traditional scheme needs at least four independent sensors to form an array, each sensor needs to be separately wired, powered and signal collected, the system cost is high, the installation process needs to be precisely geometrically calibrated, and the installation process is time-consuming. The present application only needs a single three-axis sensor, the hardware cost is greatly compressed, and the installation process is simplified.

[0052] In terms of physical model, the traditional array technology is based on geometric acoustics principle and depends on the time difference of sound wave arrival to calculate the position, and error accumulation is easy to occur in a complex reflection environment. The present application innovatively uses the sound intensity vector model to analyze the distance through the physical relationship between sound pressure and velocity, to realize the three-dimensional positioning of a single sensor, and there is no time delay error between the traditional multiple sensors, and the present application has potential advantages in overcoming the multi-path propagation interference and shows stronger adaptability in a small closed space.

[0053] In summary, the present application realizes the function of the traditional array system by using a single sensor architecture, and provides an innovative solution for small space sound source positioning. BRIEF DESCRIPTION OF DRAWINGS

[0054] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings needed to be used in the specific embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0055] Figure 1 The framework diagram of the small space sound source positioning system based on the three-axis acoustic sensor provided by the present application is shown.

[0056] Figure 2 The flow chart of the small-space sound source positioning method based on a three-axis acoustic sensor provided by the present application. DETAILED DESCRIPTION

[0057] The technical solutions of the present application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0058] The present application will be further explained and described below with reference to specific embodiments.

[0059] Please refer to Figure 1 The small-space sound source positioning system based on a three-axis acoustic sensor provided by the embodiments of the present application comprises:

[0060] a sensor module, a collection system, a processing system and an auxiliary module;

[0061] The sensor module is a single three-axis acoustic vector sensor, which is provided with one sound pressure collection channel and three orthogonal vibration velocity collection channels, and is used for synchronously collecting sound pressure signals and three-dimensional particle vibration velocity components.

[0062] The collection system comprises a signal conditioning circuit and a high-speed ADC, and is used for conditioning and synchronously sampling the signals collected by the sensor module.

[0063] The processing system comprises a data processing unit, which is used for executing the algorithm processes of three-axis synchronous collection of sound pressure and vibration velocity signals, sound intensity vector calculation, sound source direction solution, radial vibration velocity projection, sound intensity distance analysis and three-dimensional coordinate generation in real time.

[0064] The auxiliary module comprises a precision clock source and a temperature and humidity sensor, the precision clock source is used for ensuring the synchronization accuracy of each channel signal, and the temperature and humidity sensor is used for collecting environmental temperature and humidity data in real time to correct the atmospheric density and sound velocity parameters.

[0065] In the embodiments, the three orthogonal vibration velocity collection channels of the three-axis acoustic vector sensor correspond to X, Y and Z three axes respectively, and the amplitude response and phase consistency error of each channel are controlled within a preset range.

[0066] The high-speed ADC transmits the signals collected by the sensor module to the processing system through a low-jitter SPI interface.

[0067] The data processing unit adopts any one of FPGA, DSP, MCU, ASIC or CPU+GPU to meet the computing power requirement of sound intensity vector real-time calculation and frequency domain relationship analysis.

[0068] Please refer to Figure 2 The embodiment of the application also provides a small-space sound source positioning method based on a three-axis acoustic sensor, which comprises the following steps:

[0069] S1, three-axis synchronous acquisition of sound pressure and vibration velocity signals;

[0070] S2, sound intensity vector calculation;

[0071] S3, sound source direction solving;

[0072] S4, radial vibration velocity projection;

[0073] S5, sound intensity distance analysis;

[0074] S6, three-dimensional coordinate generation.

[0075] The method realizes complete conversion from original acoustic signals to three-dimensional space coordinates through the combination of physical models and mathematical transformations.

[0076] Further,

[0077] S1: three-axis synchronous acquisition of sound pressure and vibration velocity signals

[0078] This step is the data acquisition and preprocessing stage of the system, and high-precision three-axis acoustic vector sensors are used to realize synchronous acquisition and preprocessing of sound pressure signals p(t) and three-dimensional particle vibration velocity components v x (t), v y (t), v z (t). The sensor hardware includes one sound pressure acquisition channel and three orthogonal vibration velocity acquisition channels (X / Y / Z), which ensures the amplitude response and phase consistency of each channel. During data acquisition, the three-channel signals are synchronously sampled by high-speed ADC and transmitted to the processing system through a low-jitter SPI interface, thereby eliminating the synchronization error problem of the multi-sensor system from the hardware level. In order to ensure the strict synchronization of the three-axis sound pressure signals, the system performs time delay calibration through a known sound source in the initialization stage, so that the time alignment error between channels is controlled within 1 mu s.

[0079] The innovation of this step lies in the integration of the logarithmic sampling chip and the data processing chip and the design of the precise clock, which eliminates the synchronization error problem of the traditional multi-sensor system.

[0080] This step provides a high-consistency and low-noise raw data basis for the entire sound source positioning system, which is a key guarantee for subsequent high-precision positioning.

[0081] S2: Sound intensity vector calculation

[0082] The sound intensity vector is the physical basis of this technology, and it is defined as the product of the sound pressure scalar and the particle velocity vector:

[0083]

[0084] In the formula Let p(t) be the sound intensity vector, and p(t) be the sound pressure signal. v is the velocity vector of the particle. x (t), v y (x), v z (t) represents the particle velocity components along the X, Y, and Z axes, respectively;

[0085] In engineering implementation, the system calculates each component in real time at a high sampling rate (e.g., 48kHz in this example). The physical meaning of the sound intensity vector lies in the fact that its direction can approximately characterize the sound energy propagation path, while its amplitude reflects the sound power density. In specific implementation, multiplication operations are performed through a data processing unit (which can be an FPGA, DSP, etc.), with three multiplication operations completed for each sampling point (calculating I...). x ,I y ,I z The directional information of the sound intensity vector provides crucial input for subsequent sound source localization, while its amplitude variation implies the distance information to the sound source. This step transforms the raw sensor data into a vector field description with clear physical meaning, laying the theoretical foundation for the entire localization system.

[0086] S3: Sound source direction calculation

[0087] Based on sound intensity vector components, the system calculates the spatial direction of the sound source in real time:

[0088] Azimuth calculation:

[0089] θ(t)=atan2(I y (t),I x (t))

[0090] In the formula I x (t), I y (t) represents the sound intensity components along the X and Y axes at time t;

[0091] Pitch angle calculation:

[0092]

[0093] In the formula I x (t), I y (t), I z (t) represents the sound intensity components along the X, Y, and Z axes at time t;

[0094] In engineering implementation, CORDIC algorithm can be used to solve the trigonometric functions. To improve the stability of the direction estimation, the system uses a sliding time window processing: the average of the direction angle is calculated in a 100 ms time window (corresponding to 4800 sampling points):

[0095]

[0096] where N is the number of windows, θ(t i , φ(t i ) are the azimuth and elevation angles calculated in the ith window, respectively.

[0097] This average processing effectively suppresses the influence of instantaneous measurement noise and ensures the stability of the direction output. The output (θ d , φ d ) of the direction solving module accurately describes the spatial orientation of the sound source relative to the sensor.

[0098] S4: Radial velocity projection

[0099] After obtaining the direction of the sound source, the system projects the three-dimensional velocity to the direction:

[0100] The unit vector of the sound source direction is constructed:

[0101]

[0102] The radial velocity component is calculated:

[0103]

[0104] where v is the original three-dimensional velocity signal, is the unit vector of the sound source direction;

[0105] This step is one of the key steps in the entire positioning algorithm. It combines the sound source direction (θ d , φ d ) obtained in the previous steps and the original three-dimensional velocity signal v to calculate a scalar signal (v r (t)) closely related to the distance of the sound source. Without this projection, the subsequent distance analysis (step S5) cannot be performed.

[0106] Considering that the sound source positioning needs to output results in real time, the amount of original data is huge (such as 48000 points per second at a sampling rate of 48 kHz), and calculating v r (t) requires performing a three-dimensional dot product operation (3 times multiplication + 2 times addition) for each sampling point. This is a very intensive computing task. In order to quickly calculate v r(t), two optimizations are needed when implementing the algorithm:

[0107] 1. Replace the decimal numbers with integers: The three numbers representing the direction of the sound source (-1 to 1 decimal) are converted into Q15 format (16-bit signed fixed-point number, 1-bit integer + 15-bit decimal), so that the floating-point multiplication and addition can be converted into faster integer multiplication and addition. It should be noted that fixed-point quantization (such as Q15 format) will introduce quantization error, and in practical applications, it is necessary to evaluate whether the quantization bits are sufficient according to the system positioning accuracy requirements, or to consider using higher precision fixed-point numbers (such as Q31) or floating-point operations.

[0108] 2. Make full use of the parallelism of the processing system: Through Single Instruction Multiple Data (SIMD, Single Instruction Multiple Data) technology, three multiplication and addition operations are completed in one clock cycle, and the processing system simultaneously performs three multiplications v x ·u dx 、v y ·u dy 、v z ·u dz , and adds up the results.

[0109] After the above two optimizations, the entire positioning algorithm can calculate the position of the sound source in real time as the sound signal source continuously flows in.

[0110] The projected v r (t) physically represents the component of the vibration velocity in the direction of the sound source-sensor connection, and is a direct manifestation of sound energy propagation. The key significance of this step is to convert three-dimensional vibration velocity information into a scalar signal directly related to distance, providing a dedicated input channel for subsequent distance analysis.

[0111] S5: Sound intensity distance analysis

[0112] This step analyzes the distance of the sound source through the frequency domain relationship between sound pressure and radial vibration velocity:

[0113] 1. Perform 4096-point FFT transformation on p(t) and v r (t) to obtain the frequency spectrum P(f) and V r (f);

[0114] 2. Calculate the distance according to the frequency points in the effective frequency band:

[0115]

[0116] This relationship holds under free-field conditions and is the result of plane wave or spherical wave approximation.

[0117] where p is the air density and c is the sound speed. Both values are based on standard ambient conditions (atmospheric pressure 101.325 kPa, temperature 20 °C, dry air) to ensure the universality and engineering applicability of the scheme. In actual deployment, the values of p and c can be corrected in real time by temperature and humidity sensors to adapt to different working conditions.

[0118] 3. Full-band weighted average to get the final distance r:

[0119]

[0120] The physical principle is based on the characteristics of sound wave propagation in free field: the amplitude ratio of sound pressure to radial vibration velocity is related to distance and frequency. High-frequency components are more sensitive to distance changes, while low-frequency components provide stability. In engineering implementation, 1 / 3 octave sub-band division is used, and each sub-band is calculated independently and then averaged to balance accuracy and computational complexity.

[0121] S6: Three-dimensional coordinate generation

[0122] The final stage converts the spherical coordinates to Cartesian coordinates:

[0123]

[0124] This conversion mathematically establishes the mapping relationship from direction-distance to spatial position. The system output is a three-dimensional rectangular coordinate (x, y, z), which can be directly used for spatial positioning of the sound source.

[0125] In summary, the core innovation of the present application is to realize three-dimensional positioning based on the sound intensity physical model of single three-axis acoustic vector sensor. Its technical parameters can be flexibly adjusted according to actual application scenarios, which is specifically manifested as:

[0126] 1. Parameter scalability

[0127] The specific numerical values involved in the examples (such as 48 kHz sampling rate, 100 ms time window, 4096-point FFT transform, Q15 fixed-point format, etc.) are only exemplary descriptions of the specific implementation, and are not a limitation of the present application. In actual application, the sampling rate can be adjusted according to the frequency range of the sound source; the time window length can be dynamically adjusted according to the environmental noise level; the FFT point number can be selected according to the processing system's computing power and precision requirements; the fixed-point quantization format can be adapted to multiple bit widths.

[0128] The adjustment of the above parameters is based on the same technical principle and still belongs to the protection scope of the present patent.

[0129] 2. Algorithm equivalent replacement

[0130] The operation module in the core algorithm can be implemented using mathematical equivalent methods, for example:

[0131] The atan2 function in azimuth / elevation angle calculation can be replaced by lookup table method or polynomial approximation;

[0132] The moving average window can be replaced by Kalman filter or other noise suppression algorithm;

[0133] The 1 / 3 octave subband division in distance analysis can be replaced by equal bandwidth subband or other frequency domain grouping method.

[0134] 3. Hardware implementation compatibility

[0135] The data processing unit is not limited to FPGA / DSP, and can also be implemented by MCU, ASIC or CPU+GPU architecture to realize the above-mentioned process, as long as the core steps of sound intensity vector real-time calculation and frequency domain relationship analysis are met.

[0136] Finally, it should be pointed out that: the above embodiments are only used to illustrate the technical solutions of the present application, and are not limited thereto; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A small space sound source positioning system based on a three-axis acoustic sensor, characterized in that, The application relates to a three-axial acoustic vector sensor system and a three-axial acoustic vector sensor system positioning method. The sensor module is a single three-axial acoustic vector sensor, the three-axial acoustic vector sensor is provided with one sound pressure acquisition channel and three orthogonal vibration velocity acquisition channels, and is used for synchronously acquiring sound pressure signals and three-dimensional particle vibration velocity components. The acquisition system comprises a signal conditioning circuit and a high-speed ADC, and is used for conditioning and synchronously sampling the signals acquired by the sensor module. The processing system comprises a data processing unit, which is used for executing the algorithm processes of three-axial synchronous acquisition of sound pressure and vibration velocity signals, sound intensity vector calculation, sound source direction solving, radial vibration velocity projection, sound intensity distance analysis and three-dimensional coordinate generation in real time. The auxiliary module comprises a precision clock source and a temperature and humidity sensor, the precision clock source is used for ensuring the synchronization accuracy of the channels, and the temperature and humidity sensor is used for acquiring environmental temperature and humidity data in real time to correct the atmospheric density and sound velocity parameters. The three orthogonal vibration velocity acquisition channels of the three-axial acoustic vector sensor correspond to X, Y and Z three axes respectively, and the amplitude response and phase consistency error of each channel are controlled within a preset range.

2. The system of claim 1, wherein, The high-speed ADC transmits the signals acquired by the sensor module to the processing system through a low-jitter SPI interface. The data processing unit adopts any one of FPGA, DSP, MCU, ASIC or CPU+GPU to meet the computing power requirement of sound intensity vector real-time calculation and frequency domain relationship analysis.

3. The system of claim 1, wherein, The method comprises the following steps:

4. A small space sound source positioning method based on a three-axis acoustic sensor, applied to the system of any one of claims 1-3, characterized in that, The azimuth angle is calculated as follows: S1, three-axis synchronous acquisition of sound pressure and vibration velocity signals: the sound pressure signal p(t) and three-dimensional particle vibration velocity components v x (t) are synchronously acquired by a three-axis acoustic vector sensor y (t) are synchronously acquired by a three-axis acoustic vector sensor z (t) are synchronously acquired by a three-axis acoustic vector sensor, and are transmitted to a processing system after being synchronously sampled by a high-speed ADC; meanwhile, time delay calibration is performed by a known sound source during system initialization, so that the time alignment error between channels is controlled within 1 μs; S2, sound intensity vector calculation: the processing system calculates the sound intensity vector according to the sound pressure signal and the three-dimensional particle velocity components according to the formula x y z x y z x y z ​​​​​​​​​​​​​ S3, sound source direction solving: based on the sound intensity vector component obtained in S2, the real-time sound source direction angle is solved through the azimuth angle calculation formula and the pitch angle calculation formula, and then the sliding time window is used to average process the direction angle to obtain the stable sound source direction (θ d ,φ d ); The pitch angle is calculated as follows: θ(t) = atan2(I y (t),I x (t)) In the formula I x (t), I y (t) are the sound intensity components in the X and Y directions at time t, respectively. The length of the sliding time window is set to 100 ms, corresponding to 4800 sampling points, and the mean value calculation formula is as follows: wherein I x (t), I y (t), I z (t) are the sound intensity components of the X, Y, Z axes at time t, respectively. The radial vibration velocity component is obtained by projecting the three-dimensional particle vibration velocity component to the unit vector direction: where N is the number of windows, θ(t i ), φ(t i ) are the azimuth and elevation angles calculated for the i-th window, respectively. S4, radial velocity projection: according to the sound source direction (θ d ,φ d ) obtained in step S3, a sound source direction unit vector is constructed: The distance is calculated according to the frequency point in the effective frequency band: wherein is the original three-dimensional velocity signal, is the unit vector of the sound source direction; S5, sound intensity distance analysis: 4096-point FFT transform is performed on the sound pressure signal p(t) in step S1 and the radial velocity component v r (t) in step S4 to obtain the frequency spectrum P(f) and V r (f); In the formula, rho is the air density, c is the sound velocity, both of which are based on the standard environmental conditions and can be corrected in real time by the environmental data acquired by the temperature and humidity sensor; and The final sound source distance r is obtained by weightedly averaging the distances calculated in the full frequency band: In S1, the sampling rate of the high-speed ADC is adjusted according to the sound source frequency range, when the sound source frequency range is wide, the sampling rate is increased to ensure the signal acquisition accuracy; and when the sound source frequency range is narrow, the sampling rate is reduced to reduce the data processing amount. S6, three-dimensional coordinate generation: according to the sound source direction (θ d ,φ d ) obtained in step S3 and the sound source distance r obtained in step S5, the spherical coordinates are converted into Cartesian coordinates:

5. The method of claim 4, wherein, In S3, any one of the CORDIC algorithm, the table lookup method or the polynomial approximation method is adopted to realize the trigonometric function solving in the calculation of the azimuth angle and the pitch angle.

6. The method of claim 4, wherein, In S3, the length of the sliding time window is dynamically adjusted according to the environmental noise level, when the noise level is high, the length of the sliding time window is lengthened to improve the direction solving stability; and when the noise level is low, the length of the sliding time window is shortened to improve the positioning real-time performance.

7. The method of claim 4, wherein, In S5, the effective frequency band is divided by adopting the 1 / 3 octave subband division or the equal bandwidth subband division mode, the distance is calculated in each subband, and then the weighted average is performed.

8. The method of claim 4, wherein, In S4, the radial velocity component v r At (t), the fixed-point number in Q15 or Q31 format is used to replace the decimal number for operation, and the single instruction multiple data stream technology is used to complete three multiplication and addition operations in one clock cycle, so as to improve the calculation efficiency.

9. The method of claim 4, wherein, In S5, the FFT point number is selected according to the computing power and accuracy requirement of the processor, when the computing power of the processor is strong and the distance analysis accuracy requirement is high, a larger FFT point number is selected; and when the computing power of the processor is limited and the real-time requirement is high, a smaller FFT point number is selected.

10. The method of claim 4, wherein, ​

Images