Active sonar target positioning method based on ray model multi-path arrival structure matching
Patent Information
- Application Number
- CN202510369574.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-27
- Publication Date
- 2026-09-22
- Estimated Expiration
- 2045-03-27
AI Technical Summary
既可以解决目标深度估计问题，又能够弥补传统主、被动声纳目标定位的局限
[0027]与传统的主、被动声纳定位方法相比，利用主动声纳多途到达结构相关性匹配进行目标定位有以下几个优点：1.距离和深度估计。传统的主动声纳定位方法，往往只能获得目标的距离信息，难以估计目标的深度。而利用主动声纳多途到达结构相关性匹配进行目标定位，可以在主动声纳的距离-深度剖面上获取相关性最大的位置点，从而估计出目标的距离和深度。2.信号控制与增强，提升定位精度。主动声纳能够发射已知信号，并通过分析回波来定位目标，因此可以调整发射信号特性，如频率、脉冲形状和功率等信号参数，充分利用信号特性解算多途到达结构进行相关匹配，更精确地确定目标的位置。3.环境适应性提升。通过预先仿真不同环境条件下的多径到达结构，可以生成一个全面的模板库，适应多种环境变化，能够在不同的季节、温度、盐度等条件下保持较高的定位精度。另外，在建立不同环境下的模板库后，还可以提升实际定位中的信号处理速度。
Smart Images
Figure CN120143114B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the fields of underwater acoustic engineering and signal processing technology, and relates to an active sonar target localization method based on ray model multipath arrival structure matching, which is applicable to the accurate localization of underwater targets by active sonar. Background Technology
[0002] The estimation of underwater target distance and depth based on multipath arrival structure matching using ray model is essentially the detection of multipath arrival structure correlation in active sonar signal transmission channels.
[0003] Traditional active sonar localization methods primarily rely on techniques such as time delay (TOA), time difference of arrival (TDOA), beamforming, and the Doppler effect. These methods have significant limitations in complex environments and are susceptible to multipath effects, low signal-to-noise ratios (SNR), and environmental changes. Multipath effects lead to measurement errors in time delay, phase difference, beamforming, and Doppler shift, affecting localization accuracy. In low SNR environments, echo signal detection becomes difficult, measurement errors in time delay and Doppler shift increase, and the phase and amplitude information of the signal is difficult to extract accurately, making it challenging to accurately determine the target's location. Environmental changes result in varying localization errors, requiring real-time parameter adjustments and hindering accurate real-time target localization. In passive sonar localization, the sound source is not controlled by the sonar system but is naturally generated or autonomously emitted by the underwater target. Its signal strength is weak and easily interfered with by background noise, which reduces the reliability and accuracy of localization to some extent. Summary of the Invention
[0004] To address the shortcomings of traditional active and passive sonar localization methods, this invention proposes an active sonar target localization method based on ray model multipath arrival structure correlation matching. First, using an acoustic toolbox, single-path multipath channels at different distances and depths along the active sonar detection profile are simulated and calculated using a ray model, forming a sample database. Then, deconvolution is performed on the transmitted and received signals of the active sonar to extract the two-path multipath arrival structures. Finally, the calculated two-path multipath arrival structures are time-compressed to transform them into single-path multipath arrival structures. Correlation matching is then performed with the single-path multipath arrival structure samples in the simulation database to find the single-path multipath arrival structure with the highest correlation coefficient, thereby estimating the target's distance and depth. This method solves the target depth estimation problem and overcomes the limitations of traditional active and passive sonar target localization.
[0005] The technical solution of this invention is as follows:
[0006] The active sonar target localization method based on ray model multipath arrival structure matching uses the transmitted and received signals of active sonar to perform deconvolution calculations to obtain the actual multipath arrival structure. The actual multipath arrival structure is then matched with the multipath arrival structure samples obtained by simulation calculation based on the ray model to achieve the localization of underwater targets.
[0007] First, the range-depth detection profile of the active sonar is divided into a grid. Based on the grid division, the marine environment file and the location information of the sound source and receiver are correctly set. Then, based on the ray model, the single-path multipath channel structure signal Arr at the range and depth of each grid is calculated using the acoustic toolbox. origin Sample {i,j} using sampling rate f s The Nyquist theorem must be satisfied, and a "bell-shaped" expansion is performed using a Gaussian function to obtain matching samples Arr{i,j}, where i and j are the indices of grid points at different distances and depths, respectively.
[0008] Secondly, assuming an underwater target is located at a certain distance and depth within the detection range of an active sonar, the active sonar is used to detect and locate it. When the active sonar's transmitted signal reaches and contacts the target, the target reflects, absorbs, and scatters it. A portion of the reflected and scattered waves are received by the active sonar, which constitutes the received signal. Therefore, an active sonar detection system model can be established, using the active sonar's transmitted signal as the system's input signal and the active sonar's received signal as the system's output signal. The process of the transmitted signal reaching the target is the outbound arrival structure, and the process of the target's reflected signal reaching the active sonar transducer is the return arrival structure. The outbound and return arrival structures are considered together as a multipath arrival structure. Furthermore, considering the additive noise affecting signal transmission, the signals satisfy the following relationship:
[0009]
[0010] In equation (1), r(t) is the active sonar received signal (echo signal), s(t) is the active sonar transmitted signal, m(t) is the multipath arrival structure signal, and n(t) is the noise during signal transmission. This is the convolution symbol.
[0011] Performing a Fourier transform on equation (1) yields:
[0012] R(ω)=S(ω)M(ω)+N(ω) (2)
[0013] In equation (2), R(ω), S(ω), M(ω), and N(ω) are the Fourier transforms of r(t), s(t), m(t), and n(t), respectively.
[0014] Therefore, a bandpass filter can be used to remove noise from the echo signal to obtain R. denoise (ω):
[0015] R denoise (ω)=S(ω)M(ω) (3)
[0016] Then, perform an inverse Fourier transform to calculate the multipath arrival structure signal m(t) corresponding to the target:
[0017] m(t) = ifft(R) denoise (ω) / S(ω)) (4)
[0018] The multipath arrival structure signal m(t) calculated in equation (4) is sampled (sampling rate f). s The signal m(k) is obtained by using the same signal Arr{i,j} as the signal. Then, discrete sampling and time compression are performed to obtain the signal arr. Finally, a bell-shaped expansion using a Gaussian function is used to obtain the one-way arrival structure Arr corresponding to the target. target .
[0019] Finally, with Arr target Based on this, sliding correlation matching is performed with the matching samples Arr{i,j} in the simulation database to obtain the normalized correlation coefficient Corcoef:
[0020] Corcoef = xcorr(Arr) target ,Arr{i,j},'coeff') (5)
[0021] In equation (5), the parameter 'coeff' is used to normalize the calculation results, limiting the correlation value range to [-1, 1], which facilitates the comparison of correlation strength between different signals. The xcorr function is the cross-correlation function for calculating two sequences, and can calculate the signal Arr. target The process of sliding relevance matching, which uses the similarity function R[p] of Arr{i,j} at different time offsets, is defined as follows:
[0022]
[0023] In equation (6), p represents the time offset value, q represents the original index of the sequence, and Arr represents the time offset value. target [q] represents the one-way arrival structure signal Arr corresponding to the target. target The q-th value, Arr{i,j}[q+p] represents the (q+p)-th value of the matched sample signal.
[0024] By comparing the correlation coefficient of each matched sample, a correlation coefficient with the signal Arr is found in the matched sample Arr{i,j}. targetThe distance-depth of the single-path multipath arrival structure signal with the highest correlation coefficient is the estimated target location.
[0025] In the above technical solutions, the active sonar multipath arrival structure matching process based on the ray model involves signals s(t), r(t), m(k), arr, and Arr. target A schematic diagram of Arr{i,j} is shown below. Figure 2 As shown.
[0026] The beneficial effects of this invention are:
[0027] Compared with traditional active and passive sonar localization methods, target localization using active sonar multipath arrival structure correlation matching has several advantages: 1. Range and depth estimation. Traditional active sonar localization methods often only obtain the target's range information, making it difficult to estimate the target's depth. However, using active sonar multipath arrival structure correlation matching for target localization can obtain the location points with the highest correlation on the range-depth profile of the active sonar, thereby estimating the target's range and depth. 2. Signal control and enhancement, improving localization accuracy. Active sonar can emit known signals and locate targets by analyzing the echoes. Therefore, the characteristics of the emitted signal, such as frequency, pulse shape, and power, can be adjusted to fully utilize the signal characteristics to calculate multipath arrival structures for correlation matching, more accurately determining the target's position. 3. Improved environmental adaptability. By pre-simulating multipath arrival structures under different environmental conditions, a comprehensive template library can be generated, adapting to various environmental changes and maintaining high localization accuracy under different seasons, temperatures, salinity, and other conditions. In addition, establishing a template library for different environments can also improve the signal processing speed in actual localization. Attached Figure Description
[0028] Figure 1 This is a model of the active sonar detection system used in this invention.
[0029] Figure 2 This is a schematic diagram of the signal matching process for a multipath-to-structure approach.
[0030] Figure 3 This is a flowchart illustrating the implementation of the present invention.
[0031] Figure 4 This is a schematic diagram of a meshed range-depth profile for active sonar.
[0032] Figure 5 This is a sound speed gradient diagram.
[0033] Figures 6(a) and 6(b) show the simulated active sonar transmission and echo signals, respectively.
[0034] Figures 7(a) and 7(b) are comparison diagrams of the multipath arrival structure signal corresponding to the simulated target and the single-path multipath structure with the highest correlation, respectively.
[0035] Figure 8 This is a graph showing the distribution of grid correlation coefficients and the matching results with the target. Detailed Implementation
[0036] The specific embodiments of the present invention are described in detail below with reference to the technical solutions and accompanying drawings.
[0037] This invention provides an active sonar target localization method based on ray model multipath arrival structure matching, the implementation process of which is as follows: Figure 3 As shown, it includes the following steps:
[0038] Step 1: Mesh the range-depth profile of the active sonar, set the marine environment file and the location information of the sound source and receiver, and use the acoustic toolbox to simulate and calculate the one-way multipath arrival structure signal Arr{i,j} at each grid point based on the ray model, as a matching sample. The specific process includes the following steps:
[0039] (1) Based on the experimentally measured sound velocity distribution or historical data from the marine environment database, set the sound velocity gradient and marine environment file. This step uses... Figure 5 The sound velocity gradient of the marine environment shown is used for simulation calculation.
[0040] (2) Based on the distance-depth range of the sound source target, set the distance-depth detection profile of the active sonar. The distance and depth intervals of the grid division are determined according to the positioning accuracy requirements. This step follows... Figure 4 As shown, the active sonar detection profile, with a depth range of 0–100 m and a distance range of 0–10 km, is divided into 21×21 grid points according to a depth search step of 5 m and a distance search step of 0.5 km.
[0041] (3) Based on the ray model, the sound field was calculated using the Bellhop program in the acoustic toolbox. The single-path multipath arrival structure signals Arr located at grid points at different distances and depths were read and saved. origin {i,j}.
[0042] (4) The original single-path multi-path arrival structure consists of two parts of data: amplitude information A{i,j} and time delay information delay{i,j}. Therefore, it needs to be sampled, and the sampling rate is set to f. s =10000Hz, obtaining the sampled signal Arr d {i,j}:
[0043] Arr d {i,j}(round(delay{i,j}*fs ))=abs(A{i,j}) (7)
[0044] In equation (7), round is the floor function and abs is the modulo function.
[0045] (5) Use the Gaussian function to apply Arr d The {i,j} is expanded in a "bell shape" to obtain the matching sample Arr{i,j}.
[0046] Step 2: Simulate the transmission signal s of the active sonar simulate (t) and echo signal r simulate (t), the present invention performs the following simulation calculations:
[0047] (1) Simulate and generate a linear frequency modulated signal s with a pulse duration of 200ms. simulate (t).
[0048] (2) Set the simulated target position to a depth of 40m and a distance of 2km, and simplify its reflection characteristics to a scalar coefficient cons = 0.3.
[0049] (3) Run the Bellhop program in the acoustic toolbox to calculate the sound field, read and save the single-path multipath reach structure arr1 at the grid point where the simulated target is located at the distance and depth. According to the acoustic reciprocity principle, the outbound and return multipath reach structures are both arr1.
[0050] (4) According to Figure 1 The active sonar detection system model in the figure is used to obtain the simulated active sonar echo signal r by convolution operation according to the calculation principle in equation (1). simulate (t):
[0051]
[0052] In equation (8), n(t) represents additive noise during signal transmission.
[0053] The simulated active sonar transmission signal and echo signal are shown in Figure 6(a) and Figure 6(b).
[0054] Step 3: Obtain the simulated multipath arrival structure signal m through deconvolution operation. simulate (t), the principle is shown in equations (2), (3), and (4).
[0055] Step 4: Calculate the simulated multipath arrival structure signal m calculated in Step 3. simulate (t) is sampled, time-compressed, and bell-shaped extended to obtain the target-corresponding single-path multipath arrival structure signal Arr. target The sampling rate is the same as f in step 1. sThe same applies; the purpose is to ensure Arr when calculating the correlation coefficient. target The signal length is the same as that of the matching sample Arr{i,j}.
[0056] Step 5: Calculate the single-path multipath arrival structure signal Arr corresponding to the simulated target obtained in Step 4. targe The correlation coefficient set Corcoef is obtained by performing correlation calculations on the matching samples Arr{i,j} calculated in step 2, and the principle is shown in equations (5) and (6).
[0057] Step 6: By comparing the correlation coefficients at different distances and depths, find the one-way multipath arrival structure signal with the highest correlation (largest correlation coefficient) in the matching samples. The corresponding distance and depth are the estimated values of the target location.
[0058] Figures 7(a) and 7(b) show the Arr signal corresponding to the simulated target's single-path multipath arrival structure. target Comparing the signal with the single-path multipath structure Arr{i,j} that has the highest correlation in the matched samples, it can be seen that the overall structure of the signal is similar, and the time delay and amplitude relationships are basically corresponding.
[0059] Figure 8 A comparison of the target's estimated location obtained by simulating the target location and matching the multipath channel structure clearly shows that the two locations are identical, both at a depth of 40m and a distance of 2.0km. Furthermore, it can be seen that the correlation coefficient of the area surrounding the estimated target location is generally higher than that of other areas, indicating a target distribution hotspot. Therefore, it can be proven that the active sonar target localization method based on ray model multipath arrival structure matching proposed in this invention can indeed achieve accurate target localization.
Claims
1. An active sonar target localization method based on ray model multipath arrival structure matching, characterized in that, The steps are as follows: First, the range-depth detection profile of the active sonar is divided into a grid. Based on the grid division, the marine environmental file and the location information of the sound source and receiver are correctly set. Then, based on the ray model, the single-path multipath channel structure signal at the range and depth of each grid is calculated using the acoustic toolbox. Sample the samples and perform a "bell-shaped" expansion using a Gaussian function to obtain matching samples. ,in and These are the indices corresponding to grid points at different distances and depths; Secondly, assuming an underwater target is located at a certain distance and depth within the detection range of an active sonar, the active sonar is used to detect and locate it. When the active sonar's transmitted signal reaches and contacts the target, the target reflects, absorbs, and scatters the signal. Some of the reflected and scattered waves are received by the active sonar, which is the active sonar's received signal. Therefore, an active sonar detection system model is established, with the active sonar transmitted signal as the system's input signal and the active sonar received signal as the system's output signal. The process of the transmitted signal reaching the target is the outbound arrival structure, and the process of the target reflected signal reaching the active sonar transducer is the return arrival structure. The outbound and return arrival structures are considered together as a multipath arrival structure. In addition, considering the influence of additive noise during signal transmission, the signals satisfy the following relationship: (1) In equation (1), To enable active sonar to receive echo signals, For active sonar transmission signals, For multipath arrival structure signals, Noise during signal transmission, The symbol for convolution; Performing a Fourier transform on equation (1) yields: (2) In equation (2), , , , They are respectively , , , Fourier transform; Therefore, a bandpass filter is first used to remove noise from the echo signal to obtain... : (3) Then, perform an inverse Fourier transform to calculate the multipath arrival structure signal corresponding to the target. : (4) The multipath arrival structure signal calculated in equation (4) The signal is obtained by sampling. Then, discrete sampling and time compression processing are performed to obtain the signal. Then, a "bell-shaped" expansion using a Gaussian function is used to obtain the one-way arrival structure corresponding to the target. ; Finally, with Based on this, matching samples in the simulation database Perform sliding correlation matching to obtain the normalized correlation coefficient. : (5) In equation (5), the parameter Used to normalize the calculation results, so that the relevant values are limited to the range of [-1, 1]; The function is used to calculate the cross-correlation function of two sequences and to calculate the signal. and Similarity functions at different time offsets The process of achieving sliding correlation matching is defined as follows: (6) In equation (6), Represents the time offset value. Represents the original index of the sequence. The structure signal representing the one-way arrival of the target. The One value, The first one representing the matched sample signal One value; By comparing the correlation coefficient of each matched sample, in the matched samples Find the signal The distance-depth of the single-path multipath arrival structure signal with the highest correlation coefficient is the estimated target location.