Gunshot direction finding method based on rising edge arrival time estimation in multipath propagation environment

Abstract

The invention discloses a gunshot direction finding method based on rising edge arrival time estimation in a multipath propagation environment, and the method comprises the steps: carrying out the filtering of microphone array data received in the multipath environment, and carrying out the direction finding of a gunshot trajectory shock wave signal and a muzzle wave signal in a high-frequency band and a low-frequency band respectively; performing pulse waveform rising edge detection on the gunshot signal pulse to obtain arrival time estimation of the initial rising edge of the direct wave of the gunshot signal of each array element; calculating the difference value of the arrival time of the direct wave of the gunshot signal of each array element relative to the reference array element to obtain a microphone array arrival time difference vector; and according to the time difference of arrival vector of the microphone array and the formation coefficient matrix, a direction vector of the direct wave of the gunshot signal is solved by adopting a least square method, and then angle estimation of the direction of the direct wave of the gunshot signal is obtained. According to the method, the multipath interference resistance of the time difference of arrival estimation method in gunshot direction finding calculation is greatly improved, and the gunshot source positioning accuracy in a multipath environment is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of gunshot signal direction finding and acoustic signal processing, and particularly relates to a gunshot direction finding method based on rising edge time of arrival estimation in a multipath propagation environment. Background Art

[0002] Among various detection means, the gunshot positioning technology based on the acoustic principle has become an important means for detecting and positioning the enemy due to its low power consumption, simple operation, all-airspace detection ability, strong anti-interference performance, and excellent performance under non-line-of-sight conditions. With the evolution of modern military strategies and tactics, the battlefield environment is undergoing profound changes. In addition to traditional plain warfare, urban street fighting and mountain warfare are also gradually becoming important forms of modern warfare. In these complex combat environments, the high-rise urban buildings, intricate alleys, and rugged and variable mountain terrains seriously affect the propagation direction, energy distribution, and spectral characteristics of acoustic signals. In these multipath environments, the reflected wave signals will reach the microphone array along the propagation path, and the finally received signal is no longer a simple transformation of the direct signal after propagation attenuation, but a complex waveform formed by the superposition of the direct wave signal and the reflected wave signal. The time-domain waveform of the overlapping gunshot signal is severely distorted, and this distortion will change the time-domain waveform characteristics and time-frequency spectrum characteristics of the original gunshot signal, thereby affecting the positioning accuracy of the gunshot source and increasing the difficulty of accurately perceiving the enemy situation. Against this background, acoustic signal processing in complex combat environments faces severe challenges, and the accurate analysis and positioning of enemy targets, especially weapons of mass destruction, have become one of the key technologies determining the outcome of modern battles.

[0003] Among the current numerous technical means, estimating the angle between the target to be measured and the reference direction of the receiving array by the Time Difference of Arrival (TDOA) of the signal arriving at the microphone array is one of the commonly used methods. The microphone array, with its characteristics of low cost and no large amount of electromagnetic radiation, provides a passive detection means and performs excellently in some environments with high electromagnetic compatibility requirements. The TDOA technology uses multiple microphones to measure the time difference of arrival of the same signal, and then calculates the direction or position of the target signal source. Currently, most gunshot signal direction-finding methods based on TDOA estimation have high positioning accuracy in an ideal propagation environment. However, in a multipath environment, due to the aliasing of the reflected wave signal, the time-domain waveform and spectral energy distribution of the direct wave signal change greatly, and the effects of conventional TDOA-based direction-finding methods all decrease significantly, making it difficult to provide accurate sound source positioning. Early rising edge detection usually adopted the zero-crossing detection method, estimating the TOA of the signal by recording the time of the signal zero-crossing sampling points. This method performs relatively stably in a low signal-to-noise ratio environment, but the sound source positioning error in a multipath reverberation environment is still large (Zeng Xiangyang, Cai Huaizhen. Indoor dual-microphone sound source localization based on rising zero-crossing detection [J]. Audio Engineering, 2014(2): 26-30.). To solve this problem and enable the TDOA-based positioning method to be effectively applied in complex multipath environments such as cities and mountains. Summary of the Invention

[0004] The purpose of the present invention is to provide, in view of the problems existing in the above-mentioned prior art, a gunshot direction-finding method based on the rising edge arrival time estimation of the muzzle wave and ballistic shock wave signals of gunshots in a multipath propagation environment for a gunshot detection system facing a small number of array elements with an indefinite quantity. A gunshot detection system facing a small number of array elements with an indefinite quantity is proposed, and a method for gun sound source localization based on double-threshold rising edge detection. This method combines a noise threshold and a peak threshold to more accurately estimate the signal TDOA, thereby significantly enhancing the robustness to multipath effects and effectively improving the accuracy of sound source direction-finding in a complex multipath environment.

[0005] The technical solution for achieving the purpose of the present invention is as follows: On the one hand, a gunshot direction-finding method based on the rising edge arrival time estimation in a multipath propagation environment is provided, and the method includes the following steps:

[0006] Step 1, in a multipath propagation environment, use a microphone array to collect gunshot signals, and for each microphone array element channel, extract the muzzle wave signal pulse of the gunshot and the ballistic shock wave signal pulse of the gunshot;

[0007] Step 2, for each signal pulse extracted in Step 1, respectively perform rising edge detection on the pulse waveform to obtain the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels;

[0008] Step 3: Using the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels obtained in Step 2, calculate the time difference of arrival (TDOA) of the direct wave of the gunshot signal of each microphone array element relative to the reference array element respectively.

[0009] Step 4: Based on the TDOA of the direct wave of the gunshot signal of each microphone array element relative to the reference array element obtained in Step 3, use the least squares method to solve the direction vector s of the direct wave of the gunshot signal.

[0010] Step 5: Calculate the angle estimation of the direction of the direct wave of the gunshot signal according to the direction vector s.

[0011] On the other hand, a gunshot direction finding system based on rising edge time of arrival estimation in a multipath propagation environment is provided. The system includes:

[0012] The first module is used to collect gunshot signals by using a microphone array, and for each microphone array element channel, extract the muzzle wave signal pulse and the ballistic shock wave signal pulse of the gunshot.

[0013] The second module is used to, for each signal pulse extracted by the first module, respectively detect the rising edge of the pulse waveform to obtain the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels.

[0014] The third module is used to use the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels obtained by the second module to calculate the TDOA of the direct wave of the gunshot signal of each microphone array element relative to the reference array element respectively.

[0015] The fourth module is used to, based on the TDOA of the direct wave of the gunshot signal of each microphone array element relative to the reference array element obtained by the third module, use the least squares method to solve the direction vector s of the direct wave of the gunshot signal.

[0016] The fifth module is used to calculate the angle estimation of the direction of the direct wave of the gunshot signal according to the direction vector s.

[0017] On the other hand, a computer device is provided, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the gunshot direction finding method based on rising edge time of arrival estimation in a multipath propagation environment is implemented.

[0018] On the other hand, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the gunshot direction finding method based on rising edge time of arrival estimation in a multipath propagation environment is implemented.

[0019] Compared with the prior art, the remarkable advantages of the present invention are as follows:

[0020] 1) A new and effective gunshot signal direction finding strategy in a multipath propagation environment is proposed, namely a gunshot signal direction finding method based on double-threshold rising-edge TDOA estimation, which can more accurately estimate the signal TDOA, thus significantly enhancing the robustness to multipath effects and effectively improving the accuracy of sound source direction finding in a complex multipath environment.

[0021] 2) By optimizing the signal processing and multipath interference identification methods, the robustness of the system in a dynamic environment is enhanced. Even in a complex environment with strong reflection or scattering, it can still stably provide high-precision positioning results.

[0022] 3) The structure is simple. While improving the multipath resistance ability, it also ensures the real-time performance and computational efficiency of the algorithm, providing a reliable and robust solution for the efficient analysis and processing of large-scale field-measured gunshot data.

[0023] The present invention will be further described in detail below with reference to the accompanying drawings. Description of the Drawings

[0024] Figure 1 It is a flowchart of the method of the present invention.

[0025] Figure 2 It is a schematic diagram of sound source direction finding of a three-dimensional array of a small gunshot detection system in an embodiment and a qualitative analysis diagram of the direction vector s in each quadrant. Figure 2 In (a) is a schematic diagram of sound source direction finding of a three-dimensional array of a small gunshot detection system. Figure 2 In (b) is a qualitative analysis diagram of the direction vector s in each quadrant.

[0026] Figure 3 It is a time-domain waveform diagram of a gunshot signal in an ideal environment and a time-domain waveform diagram of a gunshot signal with multipath propagation in an embodiment. Figure 3 In (a) and (b) are respectively the time-domain waveform diagram and the time-frequency spectrum diagram of the measured muzzle wave signal of a gunshot with multipath propagation.

[0027] Figure 4 It is a schematic diagram of rising-edge detection of a gunshot signal in an embodiment. The TDOA of the gunshot signal is obtained according to the noise threshold and the peak threshold. Figure 4 In (a) and (b) are respectively the schematic diagrams of rising-edge detection of a gunshot signal of a reference microphone element and other microphone elements.

[0028] Figure 5 It shows the error situation between the incident angle of the gunshot signal estimated by the rising-edge detection-based method and the actual angle when the multipath signal passes through from different incident angles in an environment with a signal-to-noise ratio of 10 dB. The calculated average angle error is less than 5°. Detailed Implementation Modes

[0029] In view of the acoustic positioning requirements, the present invention proposes an effective gunshot signal direction finding method in a multipath propagation environment, that is, a gunshot direction finding method based on time-of-arrival (TOA) estimation of the rising edge in the time domain. The method proposed by the present invention has high positioning accuracy for gunshot signals in a multipath environment. Therefore, the present invention enhances the robustness of the system in a dynamic environment. Even in a complex environment with strong reflections or scattering, it can still stably provide high-precision positioning results, providing a new idea with high speed and efficiency and a reliable and robust new tool for the application of sound source positioning in a complex multipath environment.

[0030] In one embodiment, in combination with Figure 1 , a gunshot direction finding method based on rising edge time-of-arrival estimation in a multipath propagation environment is provided. The method includes the following steps:

[0031] Step 1, in a multipath propagation environment, use a microphone array to collect gunshot signals. For each microphone element channel, extract the muzzle wave signal pulse and the ballistic shock wave signal pulse of the gunshot.

[0032] Step 2, for each signal pulse extracted in Step 1, respectively perform pulse waveform rising edge detection to obtain the TOA of the direct wave rising edge of the gunshot signal in all microphone element channels.

[0033] Step 3, use the TOA of the direct wave rising edge of the gunshot signal in all microphone element channels obtained in Step 2 to calculate the time difference of arrival (TDOA) of the direct wave of the gunshot signal of each microphone element relative to the reference element.

[0034] Step 4, based on the TDOA of the direct wave of the gunshot signal of each microphone element relative to the reference element obtained in Step 3, use the least squares method to solve the direction vector s of the direct wave of the gunshot signal.

[0035] Step 5, calculate the angle estimation of the direction of the direct wave of the gunshot signal according to the direction vector s.

[0036] Further, in one embodiment, the microphone array in Step 1 is a planar array composed of no less than 3 microphones or a three-dimensional array composed of no less than 4 microphones, and all microphone element channels synchronously collect gunshot signals and convert them into digital array signals.

[0037] Further, in one embodiment, the extraction of the muzzle wave signal pulse and the ballistic shock wave signal pulse of the gunshot for each microphone element channel in Step 1 specifically includes:

[0038] For each microphone element channel, through a cut-off frequency of f lThe low-pass filter obtains the muzzle wave signal processing branch, and then, through the muzzle wave recognition method based on the multi-scale sub-band energy set characteristics, the muzzle wave signal pulse of the gunshot is intercepted;

[0039] For each microphone array element channel, a high-pass filter with a cut-off frequency of f h is used to obtain the ballistic shock wave signal processing branch, and then, through the ballistic shock wave recognition method based on singular value decomposition, the ballistic shock wave signal pulse of the gunshot is intercepted.

[0040] Here, in some embodiments, the muzzle wave recognition method based on the multi-scale sub-band energy set characteristics specifically includes:

[0041] (1) Segment the continuous audio data based on the Gaussian mixture model, and intercept the full-band audible segments;

[0042] (2) Filter the full-band audible segments obtained in (1) through a multi-scale sub-band filter bank to obtain sub-band audible segments;

[0043] (3) Align the starting points of the full-band audible segments intercepted in (1) and the filtered sub-band audible segments in (2); specifically including:

[0044] (3.1) Equally spaced frame the sub-band audible segments filtered by the filter with a center frequency of 300 Hz obtained in (2);

[0045] (3.2) Calculate the energy of each segment according to the formula to find the maximum energy max(En) and the minimum energy min(En), where En is the energy of the nth sub-band audible segment, N is the length of each sub-band audible segment, and xi is the value of the i-th point in each sub-band audible segment;

[0046] (3.3) Set an adaptive threshold according to the maximum energy max(En) and the minimum energy min(En) obtained in (3.2). When the energy of three consecutive segments of the signal exceeds the threshold Threshold, mark the first segment of these three segments as the starting segment. Among them, the threshold calculation formula is as follows:

[0047] Threshold = min(En) + τ × (max(En) - min(En))

[0048] In the formula, τ is a constant;

[0049] (3.4) Search for the amplitude spectrum peak by moving the full-band audible segments obtained in (1) and the 4 filtered sub-band audible segments in (2) forward by three segment lengths from the starting segment position, find the first peak position, and record this peak position as the starting point;

[0050] (3.5) According to the group delay characteristics of the 4 sub-band filters at their respective center frequencies, delay the offset amounts respectively to align the starting time of the voiced segments;

[0051] (4) Perform variable-scale framing and feature extraction on the voiced segments after starting point alignment in (3); specifically including:

[0052] (4.1) Perform variable-scale framing on the voiced segments after alignment in (3), and extract 5 frames before the aligned starting time;

[0053] (4.2) Extract the short-time energy as features for each of the above 5 frames. Denote the short-time energy of the nth frame of the mth sub-band segment as:

[0054]

[0055] where m = 1, 2, 3, 4 is the serial number of the sub-band segment; n = 1, 2, 3, 4, 5 is the serial number of each corresponding frame; N is the frame length, is the amplitude of a single sample point in the frame. Then the short-time energy feature vectors obtained from the full-band voiced segments and the sub-band voiced segments are:

[0056]

[0057] (4.3) Extract the short-time energy ratio feature. The short-time energy ratio is specifically the ratio of the short-time energy of the sub-band voiced segments output by the sub-band filter bank to the short-time energy of the full-band voiced segments:

[0058]

[0059] where m' = 1, 2, 3, 4 is the serial number of the sub-band segment; n = 1, 2, 3, 4, 5 is the serial number of each corresponding frame. Then the short-time energy ratio feature vectors of the 4 groups of sub-band voiced segments and the full-band voiced segments are:

[0060]

[0061] (4.4) Re-piece together the obtained short-time energy features and short-time energy ratio features to obtain the energy set features. The multi-scale sub-band energy set feature MS of each voiced segment is expressed as:

[0062]

[0063] (5) Perform muzzle wave classification according to the multi-scale sub-band energy set features extracted in (4) to obtain the best recognition result.

[0064] Preferably, in some embodiments, the ballistic shock wave recognition method based on singular value decomposition specifically includes:

[0065] 1. Signal preprocessing

[0066] Band - pass filtering: First, perform band - pass filtering on the collected gunshot signal. The pass - band frequency of the filter is 100 Hz to 600 Hz, aiming to suppress high - frequency noise and the reflected wave of the shock wave, and retain the low - frequency signal of the muzzle shock wave. Signal framing: Frame the filtered signal. The length of each frame is 0.5 ms, and the overlap between frames is 50%.

[0067] 2. Feature extraction

[0068] Time - domain feature extraction: Extract feature parameters from the time - domain waveform of the signal, mainly including:

[0069] The positive - peak rise duration T of the shock wave sr and the positive - pressure duration T s1 and the negative - pressure duration T s2 and the duration T from the positive - pressure peak to the negative - pressure peak s3 ;

[0070] The maximum value P of the positive - peak of the shock wave max and the minimum value P of the negative - peak min and the ratio of the absolute values of the peaks

[0071] The positive - pressure rise rate R of the shock wave up and the positive - pressure decay rate R down .

[0072] The duration T of the muzzle shock wave m and the positive - peak rise duration T mr and the positive - peak duration T m1 and the negative - peak duration T m2 and the duration T from the positive - pressure peak to the negative - pressure peak m3 .

[0073] The maximum pressure value P of the muzzle shock wave max and the minimum value P min and the power P m and the positive - peak power P m1 and the negative - peak power P m2 and the power P from the positive - pressure peak to the negative - pressure peak m3 and the positive - pressure rise rate R up .

[0074] 3. Feature standardization

[0075] Standardization processing: Perform standardization processing on the extracted original features to eliminate the influence of different feature dimensions and value ranges. The standardization formula is:

[0076]

[0077] where, μi is the mean value of the i-th feature, and σ i is the standard deviation of the i-th feature.

[0078] 4. Singular Value Decomposition (SVD)

[0079] Calculation of the correlation coefficient matrix: Calculate the correlation coefficient matrix of the standardized feature matrix.

[0080] Singular value decomposition: Perform singular value decomposition on the correlation coefficient matrix to obtain the singular values and the corresponding left singular vectors. The singular value decomposition formula is:

[0081]

[0082] where U is the left singular vector matrix, Σ is the singular value matrix, and V is the right singular vector matrix.

[0083] 5. Weight Vector Fitting

[0084] Weight vector selection: According to the contribution rate and cumulative contribution rate of the singular values, select the left singular vectors corresponding to the main singular values as the weight vectors. Usually, select the left singular vectors corresponding to the first few singular values with larger contribution rates.

[0085] Feature space mapping: Map the standardized feature matrix into the feature space and calculate the weighted value:

[0086] d i = ||G T F i ′||2

[0087] where G is the weight vector and Fi′ is the standardized feature vector.

[0088] 6. Recognition Threshold Training

[0089] Threshold setting: According to the weighted values of the training samples, set the recognition threshold. Usually, decrease the minimum value of the weighted values by a certain proportion as the threshold:

[0090] Thr = α × min{d1, d2, K, d n}

[0091] where α is a constant, usually taken as 0.8.

[0092] 7. Recognition Classification

[0093] Feature weighting: Perform weighting processing on the feature vectors of the test samples and calculate the weighted value:

[0094] Flag = ||G T F′||2

[0095] Threshold comparison: Compare the weighted value with the recognition threshold. If the weighted value exceeds the threshold, the signal is determined to be a target signal (shock wave or muzzle blast); otherwise, it is determined to be a non-target signal.

[0096] 8. Recognition result

[0097] Result output: According to the recognition result, output the category of the signal (shock wave or muzzle blast), and record the arrival time of the signal (TOA).

[0098] Furthermore, in one of the embodiments, for each signal pulse obtained in step 1 in step 2, by detecting the rising edge of the pulse waveform respectively, the TOA of the direct wave rising edge of the gunshot signal in all microphone array element channels is obtained, specifically including:

[0099] Step 2-1, for the gunshot signal pulse extracted in step 1, define the number of microphone array elements as I, and the signal of the i-th microphone array element channel as x i (n), and the number of sample points of the protection distance is n0; for each microphone array element channel, select L sample point data before the sample point sequence number n of the signal to be measured to obtain the background noise level estimate N i :

[0100]

[0101] Set the first rising edge detection threshold of the direct wave pulse of the gunshot signal:

[0102]

[0103] where a < b, that is, the detection threshold value of the ballistic shock wave signal pulse is greater than that of the muzzle wave signal pulse. Take the sample point sequence number n w,i when the signal amplitude first exceeds the detection threshold T 1,i as the first rising edge time estimate of the direct wave pulse of the gunshot signal, that is, TOA w,i = n 1,i , i = 1, 2,..., I;

[0104] Step 2-2, define T c as the number of sample points slightly larger than the positive half-cycle width of the gunshot signal pulse. For each microphone array element channel, with the first rising edge time estimate TOA w,i obtained in step 2-1 as the starting point, and TOA w,i + T c as the end point, search for the first peak P p,i of the gunshot signal pulse, and set the second rising edge detection threshold of the direct wave pulse of the gunshot signal:

[0105] T p,i = cP p,i, 0 < c < 1, i = 1, 2, …, I

[0106] At the pulse peak P p,i Take the sample point number n when the signal amplitude first exceeds the detection threshold T p,i for the first time 2,i as the estimation of the second rising edge time of the direct wave pulse of the gunshot signal, i.e., TOA p,i = n 2,i , i = 1, 2,...I;

[0107] Step 2 - 3: For each microphone array element channel, based on the two estimations of the rising edge time of the direct wave pulse of the gunshot signal obtained in Step 2 - 1 and Step 2 - 2, take the average value to obtain the TOA estimation of the direct wave of the gunshot signal:

[0108]

[0109] where TOA i represents the TOA of the rising edge of the direct wave of the gunshot signal corresponding to the i-th microphone array element channel.

[0110] Furthermore, in one embodiment, Step 3 specifically includes:

[0111] For the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels obtained in Step 2, taking microphone array element 1 as the reference array element, calculate the TDOA of the direct wave of the gunshot signal of each microphone array element relative to the reference array element respectively, to obtain the TDOA vector b = [Δτ 21 , Δτ 31 ,..., Δτ I1 T , where:

[0112] Δτ i1 = TOA i - TOA1, i = 2, 3,..., I

[0113] where TOA i represents the TOA of the rising edge of the direct wave of the gunshot signal corresponding to the i-th microphone array element channel, TOA1 represents the TOA of the rising edge of the direct wave of the gunshot signal corresponding to the reference array element, and I is the number of microphone array elements.

[0114] Furthermore, in one embodiment, Step 4 for solving the direction vector s of the direct wave of the gunshot signal by using the least squares method based on the TDOA of the direct wave of the gunshot signal of each microphone array element relative to the reference array element obtained in Step 3 specifically includes:

[0115] Step 4 - 1: According to the coordinate vector m of each microphone array element i = [x​i , y i , z i T , i = 1, 2, ... I. Taking microphone 1 as the reference element, calculate the connection vectors from the reference element to other elements respectively, and obtain the formation coefficient matrix A of the microphone array:

[0116] A = [m2 - m1, m3 - m1, ... m I - m1] T

[0117] where x i , y i , z i are the i-th microphone elements respectively, and I is the number of microphone elements.

[0118] Step 4 - 2: Using the formation coefficient matrix A of the microphone array obtained in Step 4 - 1 and the gunshot signal direct wave TDOA vector b obtained in Step 3, solve the direction vector s of the gunshot signal direct wave by the least squares method:

[0119]

[0120] In the formula, is the atmospheric sound speed, s x , s y , s z are the components of the direction vector s in the x, y, and z directions respectively.

[0121] Furthermore, in one embodiment, the angle estimation of the gunshot signal direct wave direction in Step 5 specifically includes:

[0122] Calculate the azimuth angle θ of the gunshot signal direct wave according to the direction vector s as:

[0123]

[0124] The elevation angle is:

[0125]

[0126] Then determine the quadrant where the azimuth angle θ is located according to the sign values of s x and s y .

[0127] In one embodiment, a gunshot direction finding system based on rising edge time of arrival estimation in a multipath propagation environment is provided. The system includes:

[0128] ​The first module is used to collect gunfire signals using a microphone array. For each microphone element channel, the muzzle wave signal pulse and the ballistic shock wave signal pulse of the gunfire are extracted;

[0129] The second module is used to, for each signal pulse extracted by the first module, detect the rising edge of the pulse waveform respectively, and obtain the TOA of the rising edge of the direct wave of the gunfire signal in all microphone element channels;

[0130] The third module is used to calculate the time difference of arrival TDOA of the direct wave of the gunfire signal of each microphone element relative to the reference element respectively by using the TOA of the rising edge of the direct wave of the gunfire signal in all microphone element channels obtained by the second module;

[0131] The fourth module is used to solve the direction vector s of the direct wave of the gunfire signal by using the least squares method based on the time difference of arrival TDOA of the direct wave of the gunfire signal of each microphone element relative to the reference element obtained by the third module;

[0132] The fifth module is used to calculate the angle estimation of the direction of the direct wave of the gunfire signal according to the direction vector s.

[0133] For the specific limitations of the gunfire direction finding system based on the time of arrival estimation of the rising edge in a multipath propagation environment, reference can be made to the limitations of the gunfire direction finding method based on the time of arrival estimation of the rising edge in a multipath propagation environment in the above text, which will not be elaborated here. Each module in the above gunfire direction finding system based on the time of arrival estimation of the rising edge in a multipath propagation environment can be implemented in whole or in part by software, hardware and their combination. The above modules can be embedded in the processor in the computer device in the form of hardware or be independent of it, or can be stored in the memory in the computer device in the form of software, so as to facilitate the processor to call and execute the operations corresponding to the above modules.

[0134] In one embodiment, a computer device is provided, including a memory, a processor and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following is implemented:

[0135] Step 1, in a multipath propagation environment, use a microphone array to collect gunfire signals. For each microphone element channel, extract the muzzle wave signal pulse and the ballistic shock wave signal pulse of the gunfire;

[0136] Step 2, for each signal pulse extracted in Step 1, detect the rising edge of the pulse waveform respectively, and obtain the TOA of the rising edge of the direct wave of the gunfire signal in all microphone element channels;

[0137] Step 3: Using the TOA of the rising edge of the direct wave of the gunshot signal in all microphone element channels obtained in Step 2, calculate the time difference of arrival (TDOA) of the direct wave of the gunshot signal for each microphone element relative to the reference element.

[0138] Step 4: Based on the TDOA of the direct wave of the gunshot signal for each microphone element relative to the reference element obtained in Step 3, use the least squares method to solve for the direction vector s of the direct wave of the gunshot signal.

[0139] Step 5: Calculate the angle estimate of the direction of the direct wave of the gunshot signal according to the direction vector s.

[0140] For the specific limitations of each step, reference can be made to the limitations of the gunshot direction finding method based on the rising edge time of arrival estimation in a multipath propagation environment described above, which will not be elaborated here.

[0141] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it implements:

[0142] Step 1: In a multipath propagation environment, use a microphone array to collect gunshot signals. For each microphone element channel, extract the muzzle wave signal pulse and the ballistic shock wave signal pulse of the gunshot.

[0143] Step 2: For each signal pulse extracted in Step 1, respectively perform pulse waveform rising edge detection to obtain the TOA of the rising edge of the direct wave of the gunshot signal in all microphone element channels.

[0144] Step 3: Using the TOA of the rising edge of the direct wave of the gunshot signal in all microphone element channels obtained in Step 2, calculate the time difference of arrival (TDOA) of the direct wave of the gunshot signal for each microphone element relative to the reference element.

[0145] Step 4: Based on the TDOA of the direct wave of the gunshot signal for each microphone element relative to the reference element obtained in Step 3, use the least squares method to solve for the direction vector s of the direct wave of the gunshot signal.

[0146] Step 5: Calculate the angle estimate of the direction of the direct wave of the gunshot signal according to the direction vector s.

[0147] For the specific limitations of each step, reference can be made to the limitations of the gunshot direction finding method based on the rising edge time of arrival estimation in a multipath propagation environment described above, which will not be elaborated here.

[0148] As a specific example, in one of the embodiments, the present invention is further verified and described in detail.

[0149] Since the direction of the muzzle wave signal of gunshots is the shooting direction of the firearm, the following takes the muzzle wave signal of gunshots as the research object, and illustrates the present invention through specific examples of the muzzle wave signal of gunshots received by a four-element planar array. The analysis of the ballistic shock wave signal of gunshots can be obtained in the same way. The specific implementation steps are as follows:

[0150] Step 1, for the multipath propagation environment, use a microphone array to collect gunshot signals; for each microphone element channel, through low-pass filtering and muzzle wave identification processing, intercept the muzzle wave signal pulse of the gunshot:

[0151] Step 1-1, for the multipath propagation environment, use no less than 3 microphones to form a planar array or use no less than 4 microphones to form a three-dimensional array. All microphone element channels synchronously collect gunshot signals and convert them into digital array signals. In a specific example of the present invention, for a four-element planar microphone array, assume that the position vector of the i-th microphone m i =[x i ,y i ,z i T , i = 1, 2, 3, 4. Since it is a planar array, then the microphone z i =0, i = 1, 2, 3, 4.

[0152] Step 1-2, for each microphone element channel, use a low-pass filter with a cut-off frequency of f l to obtain the muzzle wave signal processing branch, and through the muzzle wave identification method based on the multi-scale sub-band energy set feature, intercept the muzzle wave signal pulse of the gunshot. In a specific example of the present invention, the cut-off frequency of the low-pass filter is 800 Hz.

[0153] Step 2, for the gunshot signal pulse intercepted in Step 1, through the detection of the rising edge of the pulse waveform, obtain the TOA of the rising edge of the direct wave of the gunshot signal in all microphone element channels:

[0154] Step 2-1, for the gunshot signal pulse intercepted in Step 1, define the number of microphone elements as I, and the signal of the i-th microphone element channel as x i (n), and the number of samples of the protection distance is n0; for each microphone element channel, select L sample data before the sample number n of the signal to be measured to obtain the background noise level estimate N i . In this example, the number of microphone elements is 4, the number of selected sample points L is 50, and the number of samples of the protection distance n0 is 2000:

[0155]

[0156] ​Since the detection threshold value of the ballistic shock wave signal pulse is greater than that of the muzzle wave signal pulse, in this example, the noise threshold for the ballistic shock wave signal is generally set to 10 times the background noise base; the noise threshold for the muzzle wave signal is set to 3 times the background noise base, that is:

[0157]

[0158] Take the sample point number n w,i when the signal amplitude first exceeds the detection threshold T 1,i as the estimation of the first rising edge moment of the direct wave pulse of the gunshot signal, that is, TOA w,i = n 1,i , i = 1, 2, 3, 4.

[0159] Step 2-2, define T c as the number of sample points slightly larger than the positive half-cycle width of the gunshot signal pulse. In a specific example of the present invention, for the muzzle wave signal of the gunshot, the value of T c is selected as 0.00013 s, and for the ballistic shock wave signal of the gunshot, the value of T c is selected as 0.00006 s. For each microphone array element channel, starting from the first rising edge moment estimation TOA w,i obtained in Step 2-1, and taking TOA w,i + T c as the end point, search for the first peak P p,i of the gunshot signal pulse, and set the detection threshold for the second rising edge of the direct wave pulse of the gunshot signal:

[0160]

[0161] Before the pulse peak P p,i , take the sample point number n p,i when the signal amplitude first exceeds the detection threshold T 2,i as the estimation of the second rising edge moment of the direct wave pulse of the gunshot signal, that is, TOA p,i = n 2,i , i = 1, 2, 3, 4.

[0162] Step 2-3, for each microphone array element channel, according to the two estimations of the rising edge moments of the direct wave pulses of the gunshot signal obtained in Step 2-1 and Step 2-2, take the average value to obtain the TOA estimation of the direct wave of the gunshot signal:

[0163]

[0164] Step 3: Using the TOA of the rising edge of the direct wave of the muzzle blast signal in all microphone array element channels obtained in Step 2, with microphone element 1 as the reference element, calculate the TDOA of the direct wave of the gunshot signal of each element relative to the reference element, and obtain the TDOA vector b of the direct wave of the gunshot signal of the microphone array: b = [Δτ 21 , Δτ 31 , Δτ 41 T , where:

[0165] Δτ i1 = TOA i - TOA1, i = 2, 3, 4

[0166] Step 4: Using the TDOA of the direct wave of the gunshot signal of each element relative to the reference element obtained in Step 3, solve the direction vector s of the direct wave of the gunshot signal by the LS method, and then calculate the angle estimate of the direction of the direct wave of the gunshot signal according to the direction vector s:

[0167] Step 4 - 1: According to the coordinate vector m i = [x i , y i , z i T , i = 1, 2, 3, 4, with microphone 1 as the reference element, calculate the connection vectors from the reference element to other elements respectively, and obtain the array form coefficient matrix A of the microphone array:

[0168] A = [m2 - m1, m3 - m1, m4 - m1] T

[0169] Step 4 - 2: Using the array form coefficient matrix A of the microphone array obtained in Step 4 - 1 and the TDOA vector b of the direct wave of the gunshot signal obtained in Step 3, solve the direction vector of the direct wave of the gunshot signal by the LS method:

[0170]

[0171] In the formula, is the atmospheric sound speed.

[0172] Step 4 - 3: Using the direction vector s of the direct wave of the gunshot signal obtained in Step 4 - 2, calculate the azimuth angle of the direct wave of the gunshot signal respectively:

[0173]

[0174] Elevation angle:

[0175]

[0176] Then according to s x and s y ​​The symbol value determines the quadrant in which the azimuth angle θ lies.

[0177] Figure 2 In (a) is a schematic diagram of the sound source direction finding of a three-dimensional array of a small gunshot detection system, where the elevation angle of the sound source signal is and the azimuth angle is θ. Figure 2 In (b) is a qualitative analysis diagram of the direction vector s in each quadrant.

[0178] Figure 3 In (a) and (b) are respectively the time-domain waveform diagram and the time-frequency spectrum diagram of the measured muzzle wave signal of gunshots in a multipath propagation environment. It can be seen that multipath propagation causes fluctuations in the muzzle wave signal of gunshots. Correspondingly, the frequency spectrum diagram no longer presents a single pulse signal, and the time-domain and frequency-domain characteristics of the signal are greatly affected.

[0179] Figure 4 is a schematic diagram of the rising edge detection of the muzzle wave signal of gunshots in a multipath environment. The threshold 1 marked in the figure is the noise detection threshold set based on the background noise base, and the threshold 2 is the peak detection threshold set according to the first peak of the rising edge of the signal. Through the joint detection of the two thresholds, two TOA estimation points of the signal can be determined respectively. Finally, by calculating the mean value of these two TOA estimation points, the final estimated value of the TOA of the signal by the rising edge detection is obtained.

[0180] Figure 5 Shows the direction finding error of the muzzle wave signal of gunshots estimated based on the rising edge detection under different incident angles. In a multipath propagation environment with a signal-to-noise ratio of 10 dB, the incident angle ranges from 0° to 360°, with a step of 15°. For each incident angle, 50 experiments are repeated by the angle measurement method estimated based on the rising edge detection, and the average value of the error between the estimated angle and the actual angle is calculated. The statistical results show that the average error of the gunshot direction finding angle estimated based on the rising edge time of arrival is less than 5°, showing good direction finding accuracy.

[0181] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for gunfire direction finding based on rising edge arrival time estimation in a multipath propagation environment, characterized in that: The method comprises the following steps: Step 1: In a multipath propagation environment, a microphone array is used to collect gunshot signals, and for each microphone array element channel, a gunshot muzzle wave signal pulse and a gunshot ballistic shock wave signal pulse are extracted; Step 2: For each signal pulse extracted in step 1, the rising edge of the pulse waveform is detected to obtain the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels; Step 3, using the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels obtained in step 2, respectively calculate the time difference of arrival TDOA of the direct wave of the gunshot signal of each microphone array element relative to the reference array element; Step 4, based on the time difference of arrival TDOA of the direct wave of the gunshot signal of each microphone array element relative to the reference array element obtained in step 3, the direction vector s of the direct wave of the gunshot signal is solved by using the least square method; Step 5, calculating the angle estimation of the direct wave direction of the gunshot signal according to the direction vector s.

2. The method for gunfire direction finding based on rising edge arrival time estimation in a multipath propagation environment according to claim 1, characterized in that: The microphone array described in step 1 is a planar array consisting of no less than 3 microphones or a stereo array consisting of no less than 4 microphones, and all microphone array element channels synchronously collect gunshot signals and convert them into digital array signals.

3. The method for gunfire direction finding based on rising edge arrival time estimation in a multipath propagation environment according to claim 1, characterized in that: The extraction of gunshot muzzle wave signal pulses and gunshot ballistic shock wave signal pulses for each microphone array element channel in step 1 specifically includes: For each microphone array channel, the cutoff frequency is f l The muzzle wave signal processing branch is obtained by a low-pass filter, and then the muzzle wave signal pulse of the gunshot is intercepted by a muzzle wave recognition method based on the multi-scale sub-band energy set characteristics; For each microphone array channel, a cutoff frequency of f is used. h The ballistic shock wave signal processing branch is obtained by a high-pass filter, and then the ballistic shock wave signal pulse of the gunshot is intercepted through a ballistic shock wave identification method based on singular value decomposition.

4. The method for gunfire direction finding based on rising edge arrival time estimation in a multipath propagation environment according to claim 1, characterized in that: Step 2 is to detect the rising edge of the pulse waveform for each signal pulse obtained in step 1, and obtain the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels, which specifically includes: Step 2-1: for the gunshot signal pulse extracted in step 1, define the number of microphone array elements as I, and the channel signal of the i-th microphone array element as x i (n), the number of protection distance samples is n0; for each microphone array channel, select L sample data before the sample number n of the signal to be measured to obtain the background noise level estimate N i : Set the first rising edge detection threshold of the direct wave pulse of the gunshot signal: Where a < b, that is, the detection threshold value of the ballistic shock wave signal pulse is greater than the muzzle wave signal pulse, and the signal amplitude first exceeds the detection threshold T w,i The sample point number n 1,i As the first rising edge time estimation of the direct wave pulse of the gunshot signal, that is, TOA w,i =n 1,i ,i=1,2,...,I; Step 2-2, define T c The number of samples slightly larger than the width of the positive half-cycle of the gunshot signal pulse is used to estimate the TOA of each microphone array channel using the first rising edge time obtained in step 2-1. w,i As a starting point, TOA w,i +T c As the end point, search for the first gunshot signal pulse peak value P p,i , and set the second rising edge detection threshold of the direct wave pulse of the gunshot signal: T p,i cP p,i ,0<c<1,i=1,2,…,I At the pulse peak P p,i The signal amplitude exceeds the detection threshold T for the first time. p,i The sample point number n 2,i As the second rising edge time estimation of the direct wave pulse of the gunshot signal, that is, TOA p,i =n 2,i ,i=1,2,...I; Step 2-3, for each microphone array element channel, the TOA estimate of the direct wave of the gunshot signal is obtained by taking the average of the two gunshot signal direct wave pulse rising edge times obtained in step 2-1 and step 2-2: In the formula, TOA i It represents the TOA of the rising edge of the direct wave of the gunshot signal corresponding to the i-th microphone array element channel.

5. The method for gunfire direction finding based on rising edge arrival time estimation in a multipath propagation environment according to claim 1, characterized in that: Step 3 specifically includes: For the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels obtained in step 2, take microphone array element 1 as the reference array element, calculate the direct wave TDOA of the gunshot signal of each microphone array element relative to the reference array element, and obtain the direct wave TDOA vector b of the gunshot signal of the microphone array = [Δτ 21 ,Δτ 31 ,…,Δτ I1 ] T ,in: Δτ i1 =OLD i -TOA1,i=2,3,...,I In the formula, TOA i represents the TOA of the rising edge of the direct wave of the gunshot signal corresponding to the i-th microphone array element channel, TOA1 represents the TOA of the rising edge of the direct wave of the gunshot signal corresponding to the reference array element, and I is the number of microphone array elements.

6. The method for gunfire direction finding based on rising edge arrival time estimation in a multipath propagation environment according to claim 5, characterized in that: Step 4, based on the time difference of arrival TDOA of the direct wave of the gunshot signal of each microphone array element relative to the reference array element obtained in step 3, uses the least square method to solve the direction vector s of the direct wave of the gunshot signal, specifically including: Step 4-1, based on the coordinate vector m of each microphone array element i =[x i ,y i ,z i ] T , i=1,2,...I, take microphone 1 as the reference array element, calculate the connection vectors from the reference array element to other array elements, and obtain the array coefficient matrix A of the microphone array: A=[m2-m1,m3-m1,...m I -m1] T Among them, x i ,y i ,z i are the i-th microphone array element, I is the number of microphone array elements, Step 4-2, using the array coefficient matrix A of the microphone array obtained in step 4-1 and the TDOA vector b of the direct wave of the gunshot signal obtained in step 3, use the least squares method to solve the direction vector s of the direct wave of the gunshot signal: In the formula, is the atmospheric sound speed, s x ,s y ,s z are the components of the direction vector s in the x, y, and z directions respectively.

7. The method for gunfire direction finding based on rising edge arrival time estimation in a multipath propagation environment according to claim 6, characterized in that: The step 5 of calculating the angle estimation of the direction of the direct wave of the gunshot signal according to the direction vector s specifically includes: The azimuth angle θ of the direct wave of the gunshot signal is calculated according to the direction vector s: Pitch Angle for: Then according to s x and y The sign value of determines the quadrant in which the azimuth angle θ lies.

8. A gunshot direction finding system based on rising edge arrival time estimation in a multipath propagation environment based on the method according to any one of claims 1 to 7, characterized in that: The system comprises: The first module is used to collect gunshot signals using a microphone array, and extract gunshot muzzle wave signal pulses and gunshot ballistic shock wave signal pulses for each microphone array element channel; The second module is used to detect the rising edge of the pulse waveform for each signal pulse extracted by the first module, and obtain the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels; The third module is used to calculate the time difference of arrival TDOA of the direct wave of the gunshot signal of each microphone array element relative to the reference array element by using the TOA of the rising edge of the direct wave of the gunshot signal in all microphone array element channels obtained in the second module; The fourth module is used to calculate the direction vector s of the direct wave of the gunshot signal using the least square method based on the time difference of arrival TDOA of the direct wave of the gunshot signal of each microphone array element relative to the reference array element obtained in the third module; The fifth module is used to calculate the angle estimation of the direct wave direction of the gunshot signal according to the direction vector s.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images