Elevator door abnormal sound positioning method and system based on multiple microphones

Abstract

The invention discloses an elevator door abnormal sound positioning method and system based on multiple microphones. The method comprises the steps that four microphones are installed at the four corners of the top of an elevator car, a reference microphone is selected, a coordinate system is established with the reference microphone as an original point, the position of each microphone in the coordinate system is determined, sound signals generated when the elevator car opens and closes a door are collected, the sound signals are preprocessed, and invalid data are removed. Extracting time domain features, frequency domain features and sound pressure level features of the sound signals, setting a sound pressure threshold value, and marking the sound signals exceeding the threshold value as abnormal sound signals; and according to the reference microphone, calculating time delay information between the other microphones and the reference microphone in combination with the improved cross-correlation function, and determining the orientation of the abnormal sound source according to the time delay information. The elevator door abnormal sound positioning method based on the multiple microphones has the beneficial effects that according to the elevator door abnormal sound positioning method based on the multiple microphones, elevator door abnormal sound positioning is conducted on the microphone installation positions in the complex elevator environment and the narrow space of an elevator car, and the time delay is calculated in combination with the improved cross-correlation function; when abnormal sound appears in the opening and closing process of the elevator car door, the position of the abnormal sound can be accurately judged, and high practicability is achieved.

Description

Technical Field

[0001] The invention relates to elevator technology and acoustic fields, and in particular to a method for locating abnormal noise of an elevator door based on multiple microphones. Background Art

[0002] As an indispensable vertical transportation tool in modern buildings, the safety and reliability of elevators are directly related to the safety and comfort of passengers. The elevator door is an important part of the elevator system and plays a key role in the entry and exit of passengers in the elevator car. In the daily use of elevators, passengers may use their hands or large objects to forcibly block the closing of the elevator door, causing wear between the door and the track; when carrying large or heavy objects, it is easy to collide with the elevator door, resulting in deviations in the alignment of the elevator door, friction or abnormal noise when the door is closed; during daily cleaning, the elevator car environment is humid, which will also affect the components of the elevator door.

[0003] At present, the sound source localization system is to install multiple microphones in different positions according to a certain spatial distribution, process the signals received by each microphone, and finally obtain the position of the sound source. Patent CN107396244A discloses a sound source localization method based on a microphone array. The method uses a linear microphone array for positioning, estimates the delay information through an algorithm, and obtains the target distance information based on the geometric relationship between the array and the sound source; the method has good positioning accuracy in one-dimensional direction, however, the linear array can only be used for positioning in one-dimensional direction. If the sound source is located at different vertical heights, the array cannot obtain relevant information through one-dimensional arrangement because the microphones are not arranged vertically; Patent CN106872944A discloses a sound source localization method and device based on a microphone array. The method obtains the signal delay to calculate the position of the sound source by constructing a rectangular microphone array in the space to be detected in advance; the method has better positioning accuracy than traditional one-dimensional and two-dimensional positioning algorithms in an environment rich in reverberation and noise, however, the method requires at least eight microphones, the cost is high, and the installation range is limited due to the elevator structure, so the rectangular array is not suitable for elevator environments. Summary of the invention

[0004] In order to overcome the deficiencies of the prior art, an object of the present invention is to provide a method for locating abnormal noise of an elevator door based on multiple microphones, so as to improve the accuracy and reliability of locating abnormal noise of an elevator car door.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A method for locating abnormal noise of elevator door based on multiple microphones, characterized in that the method comprises the following steps:

[0007] Step 1: Install four microphones at the four corners of the top of the elevator car, select a reference microphone, establish a coordinate system with the reference microphone as the origin, determine the position of each microphone in the coordinate system, collect the sound signal of the elevator car when the door is opened and closed, pre-process the sound signal, and remove invalid data;

[0008] Step 1.1: Microphone installation and coordinate system establishment

[0009] Four microphones are installed at the four corners of the top of the elevator car door to form a rectangular structure, a reference microphone is selected, a coordinate system is established with the reference microphone as the origin, and the position of each microphone in the coordinate system is determined, including: selecting a microphone at the top corner of the inner side of the elevator car as the reference microphone, and establishing a coordinate system with the reference microphone as the origin, wherein the xoy plane is located at the top of the car, the z-axis is perpendicular to the xoy plane where the top of the car is located, and the positive direction is downward, and the coordinates of the remaining microphones are: (X, 0, 0), (Y, 0, 0), (X, Y, 0), wherein X and Y represent the depth and width of the car, respectively;

[0010] Step 2: extracting the time domain features, frequency domain features and sound pressure level features of the sound signal, setting a sound pressure threshold, and marking the sound signal exceeding the threshold as an abnormal sound signal;

[0011] Step 2.1: Windowing and Fourier transforming the discrete sound source signal collected by the microphone, converting the discrete sound source signal into a frequency domain signal, obtaining the denoised frequency domain signal by spectral subtraction, and finally obtaining the denoised time domain signal by inverse Fourier transform. Windowing and Fourier transform are common signal processing methods, and spectral subtraction is a common denoising method. In an elevator environment, the background noise is stable and the denoising effect is significant. The time domain signal is subjected to median filtering to smooth the signal, highlight the part with higher peak values, and eliminate the influence of burrs, so as to obtain the time domain signal after gain. Median filtering is a common data enhancement method.

[0012] Step 2.2: Set the sound pressure threshold and mark the signal exceeding the threshold as an abnormal sound signal

[0013] Calculate the sound pressure. The sound pressure calculation formula is:

[0014]

[0015] Where p is the current sound pressure, p0 is the reference sound pressure, and for each signal obtained by the windowing operation, the sound pressure level value of each window is recorded to form an array, and the mean (Mean) and standard deviation (StandardDeviation) of all sound pressure level values ​​are calculated, and the threshold is set using the mean and standard deviation:

[0016]

[0017] Threshold=Mean+2×Standard Deviation

[0018] Wherein, N is the total number of windows after the audio signal is segmented; i represents the index of the current sound pressure level data;

[0019] Step 3: According to the reference microphone, calculate the time delay information between the remaining microphones and the reference microphone, and determine the direction of the abnormal sound source according to the time delay information;

[0020] Step 3.1: Calculate the time delay of the signal arrival

[0021] Step 3.1.1: Perform Fourier transform on the processed time domain signal to obtain the frequency domain signal X i (f), calculate the power spectrum X1(f)X2(f) between the microphone signals * ;

[0022] Step 3.1.2: Calculate the signal-to-noise ratio

[0023] Step 3.1.2.1: Calculate the signal power P signal (f), by calculating the power spectrum of the frequency domain signal

[0024] P signal (f) = |X i (f)| 2

[0025] Step 3.1.2.2: Calculate the noise power P using the power obtained by the minimum tracking method noise (f)

[0026]

[0027] Among them, P noise (f, t-1) is the value of the noise power spectrum at time point t-1, and α is the smoothing coefficient.

[0028] Step 3.1.2.3: Calculate the signal-to-noise ratio. The signal-to-noise ratio is calculated by calculating the signal power P signal (f) and the noise power P noise The ratio of (f) is:

[0029]

[0030] Step 3.1.3: Calculate the signal-to-noise ratio weighting function

[0031]

[0032] Among them, γ this the signal-to-noise ratio threshold, and the value is obtained by collecting signal data for a period of time for frequency domain analysis, calculating the signal-to-noise ratio of each frequency, taking 75% of the median of the signal-to-noise ratio distribution as the initial value, and adjusting the threshold through testing.

[0033] Step 3.1.4: Calculate the low frequency weighting function

[0034]

[0035] Among them, f is the frequency of the abnormal sound signal, f c is the low frequency cutoff frequency, f c It can be adjusted according to actual conditions;

[0036] Step 3.1.5: Calculate the time delay:

[0037]

[0038]

[0039] Among them, F is the number of sampling points, X i (f) is the microphone signal x i (t) Frequency domain signal obtained after Fourier transform, X1(f)X2(f) * is the power spectrum between microphone signals (superscript * indicates complex conjugate), W SNR (f) is the signal-to-noise ratio weighting function, W LF (f) is the low frequency weighting function, r 1,2 (τ) is the cross-correlation function between microphone signals, τ 21 To calculate the time delay by finding the peak point of the cross-correlation function between the microphone signals, and then obtain the time delay estimation results of other microphones relative to the reference microphone;

[0040] Step 3.2: Determine the location of the abnormal sound source based on the time delay signal

[0041] Step 3.2.1: Construct equations based on the geometric relationship between the microphone and the location of the noise source

[0042] The geometric positions of microphones M1, M2, M3, M4 and the abnormal sound source satisfy the following formula:

[0043]

[0044] Where (x, y, z) is the coordinate of the abnormal sound source, r1 is the distance between the abnormal sound source and microphone M1, r2 is the distance between the abnormal sound source and microphone M2, r3 is the distance between the abnormal sound source and microphone M3, r4 is the distance between the abnormal sound source and microphone M4, c is the speed of sound, is the azimuth angle, θ is the elevation angle;

[0045] Step 3.2.2: Calculate the position information

[0046] Select M1 as the reference microphone, and r1≈r. Combine the equations in step 3.2.1 to calculate the position information between the abnormal sound source and the reference microphone:

[0047]

[0048] Where c is the speed of sound, τ 21 is the time delay from the abnormal sound source to microphone M1 and from the abnormal sound source to microphone M2, τ 31 is the time delay from the abnormal sound source to microphone M1 and from the abnormal sound source to microphone M3, τ 41 is the time delay from the abnormal sound source to microphone M1 and from the abnormal sound source to microphone M4, r is the distance from the abnormal sound source to the reference microphone, is the azimuth angle, and θ is the elevation angle.

[0049] A system for realizing a method for locating abnormal noise of elevator doors based on multiple microphones, comprising:

[0050] A sound collection module is used to collect sound signal data of the elevator car door during the door opening and closing process, and the sound collection device includes multiple microphones and a pickup;

[0051] The data processing module is used to pre-process the sound signal, remove invalid data, extract the time domain features, frequency domain features, and sound pressure features of the sound data of the elevator car door opening and closing process, and mark the abnormal sound signal, specifically including:

[0052] The discrete sound source signal collected by the microphone is subjected to windowing and Fourier transform processing, the discrete sound source signal is converted into a frequency domain signal, spectral subtraction is used to obtain a denoised frequency domain signal, and finally an inverse Fourier transform is performed to obtain a denoised time domain signal. Windowing and Fourier transform are common signal processing methods, and spectral subtraction is a common denoising method. In an elevator environment, the background noise is stable and the denoising effect is significant;

[0053] Perform median filtering on the time domain signal to smooth the signal, highlight the part with higher peak value, and eliminate the influence of burrs to obtain the time domain signal after gain. Median filtering is a common data enhancement method.

[0054] Setting the sound pressure threshold includes: calculating the sound pressure, setting the sound pressure threshold, and marking the signal exceeding the threshold as an abnormal sound signal;

[0055] The positioning module uses three-dimensional positioning to obtain the position information of the abnormal sound source and the reference microphone according to the time delay signal.

[0056] The beneficial effects of the present invention are as follows: the elevator door abnormal noise positioning method based on multiple microphones provided by the present invention, the proposed microphone installation position can be used to locate the elevator door abnormal noise in the complex environment of the elevator and the narrow space of the elevator car, combined with the improved cross-correlation function to calculate the delay, and when the abnormal noise occurs during the opening and closing process of the elevator car door, the position of the abnormal noise can be accurately determined, which has strong practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 The present invention is a flow chart of a method for locating abnormal noise of elevator door based on multiple microphones.

[0058] Figure 2 Schematic diagram of the placement of the collection device in the elevator car

[0059] Figure 3 The figure is a flow chart of a method for locating abnormal noise of elevator door based on multiple microphones in a specific embodiment of the method.

[0060] Figure 4 Schematic diagram of an elevator door abnormal noise positioning system based on multiple microphones described in this method DETAILED DESCRIPTION

[0061] See attached Figures 1 to 4 , a method for locating abnormal noise of elevator door based on multiple microphones, characterized in that: the method comprises the following steps:

[0062] Figure 1 : is a flow chart of a method for locating abnormal noise of elevator door based on multiple microphones. The specific implementation flow chart of the method for locating abnormal noise of elevator door based on multiple microphones is as follows: Figure 3 , combined with Figure 1 and Figure 3 , a method for locating abnormal noise of elevator door based on multiple microphones comprises the following steps:

[0063] Step 1: Install four microphones at the four corners of the top of the elevator car, select a reference microphone, establish a coordinate system with the reference microphone as the origin, determine the position of each microphone in the coordinate system, collect the sound signal of the elevator car when the door is opened and closed, pre-process the sound signal, and remove invalid data;

[0064] Step 1.1: Microphone installation and coordinate system establishment

[0065] like Figure 2As shown, four microphones are installed at the four corners of the top of the elevator car door to form a rectangular structure, a reference microphone is selected, a coordinate system is established with the reference microphone as the origin, and the position of each microphone in the coordinate system is determined, including: selecting a microphone at the top corner of the inner side of the elevator car as the reference microphone, establishing a coordinate system with the reference microphone as the origin, and labeling each microphone (for example, M1, M2, M3, M4), wherein the xoy plane is located at the top of the car, the z-axis is perpendicular to the xoy plane where the top of the car is located, and the positive direction is downward, and the coordinates of the remaining microphones are: (X, 0, 0), (X, Y, 0), (Y, 0, 0), wherein X and Y represent the depth and width of the car, respectively;

[0066] Step 2: extracting the time domain features, frequency domain features and sound pressure level features of the sound signal, setting a sound pressure threshold, and marking the sound signal exceeding the threshold as an abnormal sound signal;

[0067] Step 2.1: Sample the signal received by the microphone using a certain sampling frequency to obtain a discrete sound signal; common sampling rates include 24kHz, 44.1kHz, 48kHz, etc.

[0068] The discrete signal of the sound source collected by the microphone is subjected to windowing and Fourier transform processing, and the discrete signal of the sound source is converted into a frequency domain signal. The denoised frequency domain signal is obtained by spectral subtraction, and finally the denoised time domain signal is obtained by inverse Fourier transform. Windowing and Fourier transform are common signal processing methods, and spectral subtraction is a common denoising method. In an elevator environment, the background noise is stable and the denoising effect is significant. The time domain signal is subjected to median filtering processing to smooth the signal, highlight the part with higher peak value, and eliminate the influence of burrs at the same time, so as to obtain the time domain signal after gain. Median filtering is a common data enhancement method.

[0069] Step 2.2: Set the sound pressure threshold and mark the signal exceeding the threshold as an abnormal sound signal

[0070] Calculate the sound pressure. The sound pressure calculation formula is:

[0071]

[0072] Among them, p is the current sound pressure, p0 is the reference sound pressure (usually 20μPa), for each signal obtained by the windowing operation, the sound pressure level value of each window is recorded to form an array, the mean (Mean) and standard deviation (Standard Deviation) of all sound pressure level values ​​are calculated, and the threshold is set using the mean and standard deviation:

[0073]

[0074] Threshold=Mean+2×Standard Deviation

[0075] Wherein, N is the total number of windows after the audio signal is segmented; i represents the index of the current sound pressure level data;

[0076] Step 3: According to the reference microphone, the time delay signals between the remaining microphones and the reference microphone are calculated, and the position of the abnormal sound source is determined according to the time delay signal;

[0077] Due to the complex elevator environment, severe noise, reverberation and reflection, the core idea of ​​traditional PHAT weighting is to weight the frequency domain signal to eliminate the influence of amplitude on phase, so as to better estimate the delay of direct sound. However, in the elevator environment, due to strong reverberation and multipath effects, traditional PHAT weighting cannot effectively distinguish between direct sound and reflected sound. Therefore, the weighting function is improved, and the signal-to-noise ratio weighting and low-frequency weighting function are introduced to effectively suppress the influence of reverberation and reflection.

[0078] Step 3.1: Calculate the time delay of the signal arrival

[0079] Step 3.1.1: Perform Fourier transform on the processed time domain signal to obtain the frequency domain signal X i (f), calculate the power spectrum X1(f)X2(f) between the microphone signals * ;

[0080] Step 3.1.2: Calculate the signal-to-noise ratio

[0081] Step 3.1.2.1: Calculate the signal power P signal (f), by calculating the power spectrum of the frequency domain signal

[0082] P signal (f) = |X i (f)| 2

[0083] Step 3.1.2.2: Calculate the noise power P noise (f)

[0084]

[0085] Among them, P noise (f) is the power obtained using the minimum tracking method, that is, tracking the minimum value of the noise power spectrum in multiple time windows, P noise (f, t-1) is the value of the noise power spectrum at time point t-1, and α is the smoothing coefficient.

[0086] Step 3.1.2.3: Calculate the signal-to-noise ratio. The signal-to-noise ratio is calculated by calculating the signal power P signal(f) and the noise power P noise The ratio of (f) is:

[0087]

[0088] Step 3.1.3: Calculate the signal-to-noise ratio weighting function

[0089]

[0090] Among them, γ th is the signal-to-noise ratio threshold, and the value is obtained by collecting signal data for a period of time for frequency domain analysis, calculating the signal-to-noise ratio of each frequency, taking 70% to 80% of the median of the signal-to-noise ratio distribution as the initial value, and adjusting the threshold through testing.

[0091] Step 3.1.4: Calculate the low frequency weighting function

[0092]

[0093] Among them, f is the frequency of the abnormal sound signal, f c is the low frequency cutoff frequency, f c It can be adjusted according to actual conditions;

[0094] Step 3.1.5: Calculate the time delay:

[0095]

[0096] Among them, F is the number of sampling points, X i (f) is the microphone signal x i (t) Frequency domain signal obtained after Fourier transform, X1(f)X2(f) * is the power spectrum between microphone signals (superscript * indicates complex conjugate), W SNR (f) is the signal-to-noise ratio weighting function, W LF (f) is the low frequency weighting function, τ 21 To calculate the time delay by finding the peak point of the cross-correlation function between the microphone signals, and then obtain the time delay estimation results of other microphones relative to the reference microphone;

[0097] Step 3.2: Determine the location of the abnormal sound source based on the time delay signal

[0098] Step 3.2.1: Construct equations based on the geometric relationship between the microphone and the location of the noise source

[0099] like Figure 2As shown, M1 is selected as the reference microphone, and the reference microphone is used as the coordinate origin. The coordinate points of adjacent microphones are: M2 (X, 0, 0), M3 (X, Y, 0), M4 (0, Y, 0), and the Z axis is the vertical direction of the plane where the microphone array is located. Among them, X, Y, and Z represent the depth, width, and height of the car respectively. The values ​​of X, Y, and Z are determined by the structural parameters of the elevator car. The geometric positions of microphones M1, M2, M3, and M4 and the abnormal sound source satisfy the following formula:

[0100]

[0101] Where (x, y, z) is the coordinate of the abnormal sound source, r1 is the distance between the abnormal sound source and microphone M1, r2 is the distance between the abnormal sound source and microphone M2, r3 is the distance between the abnormal sound source and microphone M3, r4 is the distance between the abnormal sound source and microphone M4, c is the speed of sound, is the azimuth angle, θ is the elevation angle;

[0102] Step 3.2.2: Calculate the position information

[0103] Selecting M1 as the reference microphone in step 3.2.1, we can get r1≈r. Combining the equations in step 3.2.1, we can calculate the position information between the abnormal sound source and the reference microphone:

[0104]

[0105] Where c is the speed of sound, τ 21 is the time delay from the abnormal sound source to microphone M1 and from the abnormal sound source to microphone M2, τ 31 is the time delay from the abnormal sound source to microphone M1 and from the abnormal sound source to microphone M3, τ 41 is the time delay from the abnormal sound source to microphone M1 and from the abnormal sound source to microphone M4, r is the distance from the abnormal sound source to the reference microphone, is the azimuth angle, θ is the elevation angle;

[0106] Step 4: Experimental results analysis

[0107] The elevator sound data is collected using four microphones, which are installed at the four corners of the top of the elevator car door to form a rectangular structure. Figure 2As shown, in this case, microphone M1 is used as the reference microphone, and the microphone collects the original sound data of the elevator door opening and closing process. The audio data is denoised and data enhanced, and the sound pressure threshold is set according to the characteristics of the audio data. The formula is explained in step 2.2. If it is detected that the sound pressure in any time window exceeds the threshold, the audio data is marked as having an abnormal sound signal, and subsequent positioning is performed. The process is explained in step 3. In this case, the values ​​of the car parameters X, Y, and Z are: (1.6m, 1.6m, 2.5m).

[0108] In order to verify the performance of the present invention, the proposed method is compared with the traditional generalized cross-correlation (GCC-PAHT) based delay estimation arrival time difference method, beamforming based positioning method, and multiple signal classification (MUSIC) based positioning method. The experimental results in this case are shown in Table 1.

[0109] Table 1 Comparison of results of different positioning methods

[0110]

[0111] It can be seen from the table that the positioning method based on beamforming and the positioning results based on multi-signal classification have larger positioning errors than the positioning results of GCC-PAHT and the present invention. The method proposed in the present invention has the best positioning performance under different signal-to-noise ratio environments.

[0112] The contents described in the embodiments of this specification are merely enumerations of implementation forms of the inventive concept and are for illustrative purposes only. The protection scope of the present invention should not be considered to be limited to the specific forms described in this embodiment, and the protection scope of the present invention also extends to equivalent technical means that can be thought of by ordinary technicians in this field based on the inventive concept.

Claims

1. A method for locating abnormal noise of elevator doors based on multiple microphones, characterized in that: The method comprises the following steps: Step 1: Install four microphones at the four corners of the top of the elevator car, select a reference microphone, establish a coordinate system with the reference microphone as the origin, determine the position of each microphone in the coordinate system, collect the sound signal of the elevator car when the door is opened and closed, pre-process the sound signal, and remove invalid data; Step 2: extracting the time domain features, frequency domain features and sound pressure level features of the sound signal, setting a sound pressure threshold, and marking the sound signal exceeding the threshold as an abnormal sound signal; Step 3: Based on the reference microphone, calculate the time delay information between the remaining microphones and the reference microphone, and determine the direction of the abnormal sound source based on the time delay information.

2. The elevator door abnormal noise positioning method based on multiple microphones according to claim 1, characterized in that: In step 1, four microphones are installed at the top four corners of the elevator car door to form a rectangular structure.

3. The elevator door abnormal noise positioning method based on multiple microphones as claimed in claim 1, characterized in that: In the step 1, a reference microphone is selected, a coordinate system is established with the reference microphone as the origin, and the position of each microphone in the coordinate system is determined, including: selecting a microphone at the inner top corner of the elevator car as the reference microphone, and establishing a coordinate system with the reference microphone as the origin, wherein the xoy plane is located at the top of the car, the z-axis is perpendicular to the xoy plane where the top of the car is located, and the positive direction is downward, and the coordinates of the remaining microphones are: (X, 0, 0), (Y, 0, 0), (X, Y, 0), wherein X and Y represent the depth and width of the car, respectively.

4. The elevator door abnormal noise positioning method based on multiple microphones as claimed in claim 1, characterized in that: In the step 1, the sound signal is preprocessed, including: windowing and Fourier transforming the discrete sound source signal collected by the microphone, converting the discrete sound source signal into a frequency domain signal, obtaining a denoised frequency domain signal by spectral subtraction, and finally obtaining a denoised time domain signal by inverse Fourier transforming, and performing median filtering on the time domain signal to smooth the signal, highlight the part with higher peak value, and eliminate the influence of burrs to obtain a time domain signal after gain.

5. The elevator door abnormal noise positioning method based on multiple microphones as claimed in claim 1, characterized in that: In step 2, the setting of the sound pressure threshold includes: Calculate the sound pressure. The sound pressure calculation formula is: Where p is the current sound pressure, p0 is the reference sound pressure, and for each signal obtained by the windowing operation, the sound pressure level value of each window is recorded to form an array, and the mean (Mean) and standard deviation (StandardDeviation) of all sound pressure level values ​​are calculated, and the threshold is set using the mean and standard deviation: Threshold=Mean+2×Standard Deviation Wherein, N is the total number of windows after the audio signal is segmented; i represents the index of the current sound pressure level data.

6. The elevator door abnormal noise positioning method based on multiple microphones as claimed in claim 1, characterized in that: In step 3, the main expression formula for calculating the time delay is: Among them, F is the number of sampling points, X i (f) is the microphone signal x i (t) Frequency domain signal obtained after Fourier transform, X1(f)X2(f) * is the power spectrum between microphone signals (superscript * indicates complex conjugate), W SNR (f) is the signal-to-noise ratio weighting function, W LF (f) is the low frequency weighting function, τ 21 The time delay is calculated by finding the peak point of the cross-correlation function between the microphone signals, and then the time delay estimation results of other microphones relative to the reference microphone are obtained.

7. The elevator door abnormal noise positioning method based on multiple microphones as claimed in claim 1, characterized in that: In the step 3, determining the position of the abnormal sound source according to the time delay signal specifically includes: Using three-dimensional positioning, the position information of the abnormal sound source and the reference microphone is obtained according to the time delay signal, wherein the reference microphone is marked as M1, and the three-dimensional coordinate system of the three-dimensional positioning takes the reference microphone as the coordinate origin, and the coordinate points of adjacent microphones are: M2 (X, 0, 0), M3 (X, Y, 0), M4 (0, Y, 0), and the Z axis is the vertical direction of the plane where the four microphones are located, wherein X, Y, and Z represent the depth, width, and height of the car, respectively; The position information between the abnormal sound source and the reference microphone is calculated by using a trilateral positioning algorithm.

8. The elevator door abnormal noise positioning method based on multiple microphones as claimed in claim 7, characterized in that: The main formula for calculating the position information between the abnormal sound source and the reference microphone is as follows: Where c is the speed of sound, τ 21 is the time delay from the abnormal sound source to microphone M1 and from the abnormal sound source to microphone M2, τ 31 is the time delay from the abnormal sound source to microphone M1 and from the abnormal sound source to microphone M3, τ 41 is the time delay from the abnormal sound source to microphone M1 and from the abnormal sound source to microphone M4, r is the distance from the abnormal sound source to the reference microphone, is the azimuth angle, and θ is the elevation angle.

9. A system for implementing the method for locating abnormal noise of elevator doors based on multiple microphones as claimed in any one of claims 1 to 8, characterized in that: The system comprises: A sound collection module is used to collect sound signal data of the elevator car door during the door opening and closing process, and the sound collection device includes multiple microphones and a pickup; The data processing module is used to pre-process the sound signal, remove invalid data, extract the time domain features, frequency domain features, and sound pressure features of the sound data of the elevator car door opening and closing process, and mark the abnormal sound signal, specifically including: The discrete sound source signal collected by the microphone is subjected to windowing and Fourier transform processing, the discrete sound source signal is converted into a frequency domain signal, spectral subtraction is used to obtain a denoised frequency domain signal, and finally an inverse Fourier transform is performed to obtain a denoised time domain signal. Windowing and Fourier transform are common signal processing methods, and spectral subtraction is a common denoising method. In an elevator environment, the background noise is stable and the denoising effect is significant; Perform median filtering on the time domain signal to smooth the signal, highlight the part with higher peak value, and eliminate the influence of burrs to obtain the time domain signal after gain. Median filtering is a common data enhancement method. Sound pressure threshold setting, including: calculating sound pressure, setting sound pressure threshold The positioning module uses three-dimensional positioning to obtain the position information of the abnormal sound source and the reference microphone according to the time delay signal.

Images