On-site adaptive planar microphone array calibration method

Abstract

A field-level adaptive planar microphone array calibration method is disclosed. A coordinate system is established based on a reference microphone. A standard sound source emits the same single-frequency signal at different locations, which is received by the reference microphone, and the phase difference between the signal and other microphones is measured. The frequency band of interest and the sound source location are set, and the sound source is set to emit sound. The theoretical values ​​of the output signals of the other microphones are derived based on the output signal of the reference microphone. An adaptive algorithm is used to iteratively calculate the optimized filter corresponding to each microphone. This invention uses the geometric position of the microphones on the microphone array board to infer the phase difference between each microphone and calibrates the microphone phase difference in a more accurate manner using adaptive filters. It enables automated online calibration for phase differences caused by circuit installation and testing environment factors, significantly improving the accuracy of microphone arrays in practical applications while reducing operational difficulty and lowering requirements. There are no strict requirements on the positional relationship of the microphones or the selection of the reference microphone.

Description

Technical Field

[0001] This invention relates to a technology in the field of noise localization, specifically a field-level adaptive planar microphone array calibration method. Background Technology

[0002] Sound source localization technology has wide applications in fault detection and far-field sound pickup. Localization algorithms are often designed using the phase difference acquired by different microphones in a microphone array. The phase difference reflects the time difference of sound waves arriving at different microphones in the array and is crucial for deriving the azimuth angle of the sound source. Therefore, the accuracy of the phase difference is critical to the accuracy of localization. It is essential to ensure that the phase difference of microphones in the same array is caused only by the time difference of sound wave arrival. Phase difference calibration is a necessary step before the microphone array can operate. Currently, most methods for calibrating the phase difference of microphone arrays are performed in a laboratory before installation, using offline methods to calibrate the performance differences of different microphones. However, this type of method ignores the phase difference introduced by various electrical components and circuits after the microphones are mounted on the circuit board. Furthermore, changes in the working environment, such as temperature and humidity, can affect the phase difference of the microphone array, making previous calibration results unreliable. Offline calibration methods also increase the difficulty of recalibration. Summary of the Invention

[0003] To address the aforementioned shortcomings of existing technologies, this invention proposes a field-adaptive planar microphone array calibration method that enables field-adaptive adjustment and is simple and universal.

[0004] This invention is achieved through the following technical solution:

[0005] This invention relates to a field-level adaptive planar microphone array calibration method. A coordinate system is established based on a reference microphone. The same single-frequency signal is emitted from a standard sound source at different locations and received by the reference microphone. The phase difference between the signal and other microphones is measured. The frequency band of interest and the sound source location are set, and the sound source is set to emit sound. The theoretical values ​​of the output signals of the other microphones are derived based on the output signal of the reference microphone. The optimized filter corresponding to each microphone is obtained by iterative calculation using an adaptive algorithm.

[0006] Technical effect

[0007] This invention is based on an improved phase difference calibration technique. It infers the phase difference between microphones by utilizing the geometric position of the microphones on the microphone array board and calibrates the microphone phase difference in a more accurate manner using an adaptive filter. Compared with existing technologies, this invention enables automated online calibration of phase differences caused by circuit installation and testing environment factors. This significantly improves the accuracy of microphone arrays in practical applications while reducing operational difficulty and lowering requirements. It also eliminates strict requirements on the positional relationship of the microphones and the selection of the reference microphone. Attached Figure Description

[0008] Figure 1 This is a flowchart of the present invention;

[0009] Figure 2 This is a schematic diagram of the present invention;

[0010] Figure 3 A signal flow graph for a microphone array;

[0011] Figure 4 A plan view of an 8-microphone array arranged in a circle;

[0012] In the picture: 11 Microphone, 12 Circuit Board. Detailed Implementation

[0013] like Figure 2 As shown, this embodiment relates to a field-level adaptive planar microphone array calibration method, including:

[0014] S1) For the board as shown Figure 1 Given a planar N-microphone array with a known geometric arrangement, a reference microphone is selected, and its output signal is used as a reference for the output signals of the other microphones. A three-dimensional coordinate system is established with the acoustic center of the reference microphone as the origin of the planar coordinate system, and the coordinates of this microphone are (0, 0, 0). Further, the coordinates (x, y) of the acoustic centers of all microphones in the array are determined. i y i ,0), where i = 1, 2, ..., N-1.

[0015] Since the positioning algorithm uses the relative difference between phases rather than the absolute difference, it is sufficient to specify a microphone located in the array to achieve the reference purpose, without the need to introduce an additional reference signal.

[0016] like Figure 1 As shown, one of the microphones is arbitrarily designated as the reference microphone, such as the microphone in the lower left corner of the figure. The acoustic center of this microphone is taken as the origin, the diameter passing through this point is the x-axis, the straight line perpendicular to the x-axis in the plane passing through this point is the y-axis, and the straight line passing through the origin and perpendicular to the xy-plane is the z-axis, thus establishing a three-dimensional spatial coordinate system. The sound source S is located on the z-axis of the coordinate system. A suitable distance is chosen between S and the microphone array to make it a near-field model. Therefore, the sound signal emitted by the sound source has a distance-dependent phase difference when it reaches different microphones.

[0017] In other situations, such as Figure 4 The circular circuit board shown has a diameter of 60mm. The microphones are located on the eight equally spaced lines of the board, and the acoustic center of each microphone is 26mm from the center of the circle. The acoustic center coordinates of the eight microphones on the board can be obtained. The reference microphone coordinates are (0, 0, 0), and the coordinates of the other seven microphones are (x...).i y i ,0),i=1,2,...,N-1.

[0018] S2) To determine the speed of sound c in the environment, select two positions in the coordinate system that are at different distances from the acoustic center of the reference microphone, with a distance difference of d. The sound source emits the same sound signal with the same operating frequency f0 at the two positions, which is received by the reference microphone. The phase difference between the two received signals is then obtained. The speed of sound was calculated.

[0019] For example, select two points (0, 0, 1000) and (0, 0, 1300) at different distances from the reference microphone, with a distance difference of 300mm. Place a standard sound source S at these two points respectively, emitting the same 500Hz sound signal, which is received by the reference microphone. Analyze the phase difference θ of the two obtained sound signals, and the speed of sound c is...

[0020] The preferred operating frequency f0 is 100-500Hz, the difference between the two positions of the sound source and the distance to the reference microphone is 0.3-1m, and the sound source is located on the z-axis. This ensures good sound signal reception and a suitable phase difference.

[0021] S3) Determine the highest frequency f of the sound signal emitted by the standard sound source S based on the frequency band of interest in the application. max This determines the position coordinates (x, y) of the sound source in the coordinate system. S y S , z S This ensures that the distance between the sound source and the geometric center of the microphone array is less than two wavelengths of the emitted signal. In this case, the sound field can be considered as a near-field model, and the phase difference of the output signals of each microphone is affected by the distance difference.

[0022] For example, if the frequency band of interest is 1kHz to 10kHz, then the maximum frequency emitted by the standard sound source is 10kHz, and the minimum signal wavelength is 0.034m. The sound source's location coordinates can be defined as (0, 0, 100). In this case, the distance between the sound source and the geometric center of the circuit board is approximately three signal wavelengths, which can be considered a near-field model. Based on the formula, the phase difference between the reference microphone and other microphones due to their positions can be derived as follows:

[0023] S4) as Figure 3 As shown, the sound signal emitted by the sound source, after propagation, reaches the reference microphone as x. c (t), based on the far-field assumption and the coordinates of other microphones in the array, estimates the sound signal x reaching the remaining N-1 microphones. i The phase difference between (t) and the reference microphone f is the selected analysis frequency that yields the best results, where the speed of sound c is a stable value when the test environment remains unchanged. Therefore, substituting the formula in step S2, we have... All variables in this formula can be obtained through experimentation or measurement, achieving good accuracy. The ideal phase difference expressed will serve as a reference for subsequent self-calibration.

[0024] like Figure 3 As shown, reference microphone H c The output signal y c (n), corresponding to the other i microphones H to be calibrated i The theoretical output signal y i (n), based on the differences in the subsequent electrical circuits of each microphone and the impact of the test environment on its performance, in addition to the theoretical phase difference Actual microphone to be calibrated H i The output signal y′ i (n) and the signal y from the reference microphone c Compared to (n), there is also an additional phase difference.

[0025] S5) Using a digital adaptive algorithm, the optimized filter for each microphone is obtained iteratively by involving a cost function.

[0026] The digital adaptive algorithm employs methods such as Least Mean Square (LMS), Frequency Domain Least Mean Square (FLMS), and Recursive Least Squares (RMS), with the Frequency Domain Least Mean Square being preferred.

[0027] The cost function is the theoretical output signal y of any microphone. i (n) and the actual output signal y' i The difference between the Fast Fourier Transforms of (n).

[0028] The iteration specifically refers to: setting y i (n) and y' i (n) Perform the difference operation, then perform an FFT on the result to obtain D. i (f)=Y i (f)-Y′ i (f); the cost function of the i-th adaptive filter in the FLMS algorithm Then filter G i (f)=argmin{J i}, its matrix form is G = argmin{J} = argmin{DD} H}, where: G = [G1(f), G2(f), ..., G7(f)] T , D=[D1(f), D2(f),…,D7(f)] TChoose an appropriate step size factor μ for iteration, when the cost function J i When the error between the actual output and the ideal output is minimized, the adaptive filter parameters for each microphone can be obtained.

[0029] Compared with existing technologies, this invention breaks through the limitation of traditional measurement methods that require microphones to be placed close together. By inferring phase difference through geometric position, this method can include phase difference caused by circuit installation and testing environment factors in the calibration range, thus improving the accuracy of microphone arrays in practical applications. Furthermore, based on the technical principle of this method, an adaptive filter suitable for this method is designed, and new provisions are made for its cost function, etc. This adaptive algorithm transforms the previously offline phase difference calibration into an online automated process.

[0030] The above-described specific implementations can be partially adjusted by those skilled in the art in different ways without departing from the principles and purpose of the present invention. The scope of protection of the present invention is defined by the claims and is not limited to the above-described specific implementations. All implementation schemes within the scope of the claims are bound by the present invention.

Claims

1. A method for on-site level adaptive planar microphone array calibration, characterized in that, A coordinate system is established based on a reference microphone. A standard sound source emits the same single-frequency signal at different locations, which is received by the reference microphone, and the phase difference between the signal and other microphones is measured. The frequency band of interest and the sound source location are set, and the sound source is activated. The theoretical values ​​of the output signals of the other microphones are derived from the output signal of the reference microphone. An adaptive algorithm is used to iteratively calculate the optimized filter for each microphone, specifically including: S1) For a planar N-microphone array, select a reference microphone and use its output signal as a reference for the output signals of the other microphones; establish a three-dimensional coordinate system with the acoustic center of the reference microphone as the origin of the planar coordinate system. The coordinates of this microphone are... Further determine the coordinates of the acoustic centers of all microphones in the microphone array. where i = 1, 2, ..., N-1; S2) To determine the speed of sound c in the environment, select two positions in the coordinate system that are at different distances from the acoustic center of the reference microphone, with a distance difference of c. The sound source emits the same operating frequency at two different locations. The signal is received by the reference microphone, and the phase difference between the two received signals is obtained. The speed of sound is calculated. S3) Determine the highest frequency of the sound signal emitted by the standard sound source S based on the frequency band of interest in the application. This determines the position coordinates of the sound source in the coordinate system. To ensure that the distance between the sound source and the geometric center of the microphone array is less than two wavelengths of the emitted signal, the sound field can be regarded as a near-field model, and the phase difference of the output signal of each microphone is affected by the distance difference. S4) The sound signal emitted by the sound source, after propagation, reaches the reference microphone as the sound signal is... Based on the far-field assumption and the coordinates of other microphones in the array, the sound signals reaching the remaining N-1 microphones are estimated. The phase difference between the reference microphone and the sound velocity c is obtained by substituting the result of the calculation in step S2. ,in: The selected analysis frequency yields the best results; S5) Using a digital adaptive algorithm, the optimized filter for each microphone is obtained iteratively by involving a cost function.

2. The field-level adaptive planar microphone array calibration method according to claim 1, characterized in that, The digital adaptive algorithm described above employs the least mean square algorithm, the least mean square algorithm in the frequency domain, or the recursive least squares method.

3. The field-level adaptive planar microphone array calibration method according to claim 1, characterized in that, The cost function is the theoretical output signal of any microphone. and actual output signal The difference between the Fast Fourier Transform and the Fast Fourier Transform.

4. The field-level adaptive planar microphone array calibration method according to claim 3, characterized in that, The iteration specifically refers to: and Do the difference, perform an FFT on the result, and obtain In the FLMS algorithm, the cost function of the i-th optimized filter Then optimize the filter Its matrix form ,in: , Choose an appropriate step size factor Perform iterations, when the cost function When the error between the actual output and the ideal output is minimized, the parameters of the optimized filter for each microphone can be obtained.

Images