A non-uniformly segmented near-field fast beamforming method

Abstract

The application relates to a non-uniform segmented near-field fast beam forming method, belonging to the field of radar signal processing. The method comprises the following steps: determining an array aperture, a distance gain loss tolerance value and a distance range needing segmentation, judging whether the distance range to be segmented meets the segmentation condition; searching distance grid points in a traversal mode, and calculating a matching coefficient of each distance segment; performing matching filtering processing according to the matching coefficient of each distance segment; and splicing the matching filtering results of each distance segment. The application uses the same weight coefficient in each distance segment to replace the accurate three-dimensional near-field weight coefficient, greatly reduces the number of near-field beam forming weight coefficients, and effectively reduces the algorithm complexity on the premise that the distance resolution and ranging accuracy are basically unchanged.

Description

Technical Field

[0001] This invention belongs to the field of radar signal processing and relates to a low-complexity fast search method for targets with non-uniform range segments in a distributed signal-level networked near-field detection mode. Background Technology

[0002] As the aperture of distributed radar continues to increase, the traditional far-field target detection problem is transforming into a near-field detection problem. For far-field target detection, the beamforming weights are only related to the azimuth and elevation angles; however, for near-field target detection, the beamforming weights are not only related to the azimuth and elevation angles but also tightly coupled with the target detection range, constituting a three-dimensional spatial matching process of "azimuth-elevation-range". Therefore, when searching for near-field targets, the number of beamforming weights increases explosively, severely impacting the real-time processing performance of the distributed radar system.

[0003] To improve the real-time performance of the system in near-field search mode, range segmentation is an effective way to reduce computational complexity. This method divides the range into several range intervals, each using the same near-field beamforming weights. While ensuring controllable energy loss, it achieves full coverage of the "azimuth-elevation-range" three-dimensional space with as few range segments as possible, thus meeting real-time processing requirements. The rate of energy attenuation in the range dimension of near-field beamforming mainly depends on the range parameter in the near-field weighting coefficients; that is, the smaller the range parameter in the near-field weighting coefficients, the more severe the energy attenuation in the range dimension. The uniform segmentation method may result in two scenarios: Scenario 1 involves uniformly dividing the entire search range using the segment length obtained from the long distance. This method results in fewer segment lengths, but because the near-field beamforming is less dependent on distance at long distances, the segment lengths are longer, making it impossible to achieve full coverage of the segment using only one set of weighting coefficients at close range. Scenario 2 involves uniformly dividing the entire search range using the segment length obtained from the short distance. While this ensures seamless full coverage of the search range, the relatively short length of each segment means that the number of segments is not significantly reduced, making it difficult to guarantee the system's real-time performance.

[0004] To address the above issues, this study investigates near-field non-uniform range segmentation methods. Considering the range dependence of weighting coefficients, a non-uniform range segmentation approach is adopted. Combining the characteristics of synthetic beam pattern, a fast near-field beamforming method for distributed signal-level networked radar is proposed. This method optimizes the number of non-uniform range segments under controllable energy attenuation, significantly reduces the number of weighting coefficients in near-field beamforming, and improves the real-time processing performance of distributed radar systems. Summary of the Invention

[0005] The technical problem to be solved by this invention is:

[0006] To address the challenge of decreased real-time performance caused by the explosive growth of filter weight coefficients in focused near-field detection mode, this invention proposes a non-uniformly segmented near-field fast beamforming method. This invention employs a non-uniform range segmentation approach. While ensuring controllable gain loss due to weight coefficient mismatch, it reduces the computational complexity of the "azimuth-elevation-range" three-dimensional spatial matching search in near-field detection mode by applying the same weight coefficients to the echo data within each range segment for near-field beamforming processing.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0008] A non-uniformly segmented near-field fast beamforming method, characterized by the following steps:

[0009] Step 1: Determine the array aperture and range gain loss tolerance value, and determine whether the distance from the target to the array center meets the segmentation conditions. If the segmentation conditions are not met, the process ends directly.

[0010] Step 2: Traverse and search for the first optimal distance grid point. The left intersection of its distance dimension response and distance dimension gain tolerance value is the midpoint of the first distance resolution cell, and its right intersection is recorded.

[0011] Step 3: Traverse and search the next segmented grid point, and determine whether the left intersection of the distance dimension response and the distance gain tolerance value of the point is greater than the right intersection recorded in Step 2. If it is less, continue to search the next segmented grid point. If it is greater, record the previous distance point as the best grid point under the condition that it does not exceed the maximum distance value.

[0012] Step 4: Calculate the matching coefficient for each distance segment based on the segmented grid points obtained from the traversal search in Step 3;

[0013] Step 5: Perform matched filtering based on the matching coefficients of each distance segment; concatenate the matched filtering results of each distance segment to obtain an equivalent matched filtering result of the complete distance segment for subsequent processing.

[0014] A further technical solution of the present invention: The segmentation condition in step 1 is specifically: determining whether the distance R from the target to the array center satisfies the near-field condition R > max(R0, R1, L), where the distance dimension response at the focal center distance R0 is... The value at ζ / 2 is equal to ε; where ζ is the range resolution cell, R1 is the maximum value of the radar blind zone set by the user, L is the current aperture of the array, and ε is the range dimension gain loss tolerance value.

[0015] A further technical solution of the present invention: the calculation of the matching coefficient for each distance segment specifically involves:

[0016]

[0017] Where s0(t) is the binary phase code transmission signal, r k For the segmented grid point of the k-th distance segment, and The response in the distance dimension is respectively The left and right intersections with the distance tolerance value Let θ be the latency between the transmitting and receiving nodes, and θ be the elevation angle of the target relative to the array center. The azimuth angle of the target relative to the center of the array.

[0018] A further technical solution of the present invention: the matched filtering output based on the matching coefficient of each distance segment is as follows:

[0019]

[0020] Where χ(t) is the autocorrelation function of the binary code, (·) * This indicates the conjugate operation. and The response in the distance dimension is respectively The left and right intersections with the distance tolerance value.

[0021] A further technical solution of the present invention: the specific method for concatenating the matched filtering results of each distance segment is as follows:

[0022] The matched filter outputs from all K distance segments are concatenated along the t-axis to obtain an approximate output:

[0023]

[0024] Where α is the gain loss coefficient introduced by the segmented processing.

[0025] A computer system is characterized by comprising: one or more processors, and a computer-readable storage medium for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the method described above.

[0026] A computer-readable storage medium is characterized by storing computer-executable instructions, which, when executed, are used to implement the above-described method.

[0027] The beneficial effects of this invention are as follows:

[0028] This invention addresses the challenge of decreased real-time performance caused by the explosive growth of filter weight coefficients in near-field detection mode. By employing non-uniform segmentation of the range dimension, it significantly improves the algorithm's real-time performance. Compared to fine-grained matching in near-field three-dimensional space, this invention uses segmented range processing, replacing precise three-dimensional near-field weight coefficients with the same weight coefficients across all range segments. This drastically reduces the number of near-field beamforming weight coefficients, effectively lowering algorithm complexity while maintaining essentially unchanged range resolution and ranging accuracy. Compared to uniform segmentation, this invention further reduces the number of range segments while ensuring seamless continuity of near-field beam response characteristics across the entire detection range, thus enhancing the algorithm's real-time performance. Attached Figure Description

[0029] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts.

[0030] Figure 1 (a) is a flowchart illustrating the method of the present invention; wherein step 1 determines the array aperture, the range gain loss tolerance value, and the range of distances to be segmented. If the range of distances to be segmented does not meet the segmentation conditions, the process ends directly; step 2 iterates and searches for the first optimal range grid point, and the left intersection of its range response and the range gain tolerance value is the midpoint of the first range resolution cell, and its right intersection is recorded; step 3 iterates and searches for the next segmented grid point, and determines whether the left intersection of the range response and the range gain tolerance value of this point is greater than the right intersection recorded in step 2. If it is less, the search continues for the next segmented grid point; if it is greater, the previous distance point is recorded as the optimal grid point, provided that it does not exceed the maximum distance value; step 4 calculates the matching coefficient of each range segment based on the non-uniform segment center obtained from the iterative search in step 3; step 5 performs matched filtering processing based on the matching coefficient of each range segment. The matched filtering results of each range segment are concatenated to form an equivalent matched filtering result of the complete range segment for subsequent processing.

[0031] Figure 1 (b) is a diagram illustrating the local characteristics of the beam curve. The range response and the range gain loss tolerance have two intersection points, one on the left and one on the right. The distance from the beam center to the left intersection point is less than the distance from the beam center to the right intersection point. Therefore, the search can start from the distance corresponding to the symmetrical point of the left intersection point about the beam center.

[0032] Figure 2 (a) is the structure of the transceiver array in an embodiment of the present invention. The entire receiving array is distributed on the XOY plane and is symmetrical about the x-axis and y-axis, with equal spacing between the array elements on each side; the transmitting array elements are distributed on both sides of the array.

[0033] Figure 2(b) In this embodiment of the invention, three types of targets are set at near, medium and far distances. The grid matching map is output using non-uniform segmentation. The matching area is [2km, 60km] and the number of segments is 9. The target positions are (4km, 45.5°), (20km, 46°) and (45km, 46.5°) respectively.

[0034] Figure 2 (c) is the distance-amplitude map of the grid matching map output by non-uniform segmentation, which sets up three types of targets at near, medium and far distances in an embodiment of the present invention.

[0035] Figure 2 (d) is the angle-amplitude diagram of the grid matching map output by non-uniform segmentation, which sets up three types of targets at near, medium and far distances in the embodiment of the present invention.

[0036] Figure 2 (e) is a grid matching map with three types of targets at near, medium and far distances in this embodiment of the invention, without non-uniform segmentation, and with accurate three-dimensional compensation output. The matching area is [2km, 6km], [18km, 22km] and [43km, 47km], and the matching step size is one distance resolution unit (24.69m). The target positions are (4km, 45.5°), (20km, 46°) and (45km, 46.5°) respectively.

[0037] Figure 2 (f) is the distance-amplitude map of the raster matching map in this embodiment of the invention, which sets up three types of targets at near, medium and far distances and does not perform non-uniform segmented output.

[0038] Figure 2 (g) is the angle-amplitude diagram of the raster matching map in this embodiment of the invention, which sets up three types of targets at near, medium and far distances and does not perform non-uniform segmented output.

[0039] Figure 2 (h) is a comparative diagram showing the distance dimension response at distance center 2120 of the first segment when a gain loss tolerance value of 0.4dB is set in this embodiment of the invention, and the distance dimension response at distance center 2119 when the distance center of the first segment meets the requirements is not.

[0040] Figure 3 This is the result of setting a gain loss tolerance value of 0.4dB and non-uniform segmentation in this embodiment of the invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0042] The basic idea behind this invention is as follows: First, based on the current distributed radar aperture size and the tolerance for gain loss caused by weight coefficient mismatch, determine the maximum and minimum detection range intervals of the radar system that meet the segmentation requirements. Then, using the minimum detection range as the initial range, change the range parameter in the near-field weight coefficients, and combine the beam response and gain loss tolerance to determine the starting point (left intersection point) and ending point (right intersection point) under different range parameters. When the left intersection point coincides with the initial range, the current range segment is updated. Next, use the ending point of the previous range segment as the starting point of the next range segment, and repeat the above operation to finally complete the non-uniform division of the entire range segment.

[0043] To achieve the above objectives, the technical solution of the present invention includes the following steps:

[0044] Step 1: Calculate the current array aperture L based on the beam pointing, determine the range dimension gain loss tolerance value ε, and then determine whether the distance R from the target to the array center satisfies the near-field condition R > max(R0, R1, L), where the range dimension response at the focusing center distance R0 is... The value at ζ / 2 is equal to ε, where ζ is the range resolution unit and R1 is the maximum value of the radar blind zone set by the user.

[0045] Step 2: Divide the region that meets the distance segmentation requirements from Step 1 into segments. First, traverse and search for the first optimal distance segmentation grid point, whose distance dimension response... Left intersection with distance tolerance value ε It equals ζ / 2, and the right intersection point is... The optimal criterion is whether the left intersection point is equal to ζ / 2.

[0046] Step 3: Search for the next optimal distance segmented grid point R y Its distance dimension response The left intersection point with the distance tolerance value ε is The right intersection point is The optimal judgment condition is the left intersection point. The right intersection point smaller than the previous optimal segmented grid point And at the focal center R y+1 Distance dimension response Left intersection with distance tolerance value Greater than We can take advantage of the fact that the distance from the optimal segmented grid point to the left intersection point is less than that to the right intersection point, without needing to start from R. x +1 point-by-point search, can directly from The search begins at this point, greatly reducing computation time.

[0047] Step 4: Update the current optimal distance segment grid point to R y Record its right intersection point Iterative updates are performed until the maximum value of the segmented region is reached. The near-field multidimensional matching function corresponding to the k-th distance segment can be obtained as follows:

[0048]

[0049] Where s0(t) is the binary phase code transmission signal, r k For the segmented grid point of the k-th distance segment, and The response in the distance dimension is respectively The left and right intersections with the distance tolerance value Let θ be the latency between the transmitting and receiving nodes, and θ be the elevation angle of the target relative to the array center. The azimuth angle of the target relative to the center of the array.

[0050] Step 5: From the matching function of the k-th distance segment obtained in Step 4, the matched filter output of the k-th distance segment can be obtained as follows:

[0051]

[0052] Where χ(t) is the autocorrelation function of the binary code, (·) * This indicates the conjugate operation. and The response in the distance dimension is respectively The left and right intersections with the distance tolerance value. Concatenating the matched filter outputs across all K distance segments along the t-axis yields an approximate output:

[0053]

[0054] Where α is the gain loss coefficient introduced by the segmented processing.

[0055] Example:

[0056] Step 1: Calculate the current array aperture as 90m based on the beam pointing, determine the range dimension gain loss tolerance value as 0.4dB, and then determine whether the target distance R from the array center satisfies the near-field condition R > max(R0,R1,L). The range dimension response at the focusing center distance of 148.14m has a value greater than or equal to -0.4dB at half the range cell (37.035m). The range resolution cell is 74.07m. The radar blind zone is artificially set to 2000m, and the range that meets the segmentation requirements is R > 2000m.

[0057] Step 2: Divide the region that meets the distance segmentation requirements from Step 1 into segments. First, traverse and search for the first optimal distance segmentation grid point, whose distance dimension response... The left intersection point with a distance tolerance of 0.4 dB The distance is equal to 2000m, and the right intersection point is 2255m. The optimal judgment condition is that when the distance from the center is 2120m, the distance gain loss at 2000m is 0.3993dB, and when the distance from the center is 2121m, the distance gain loss at 2000m is 0.4057dB.

[0058] Step 3: Search for the next optimal distance segmented grid point R y Its distance dimension response The left intersection point with the distance tolerance value of 0.4dB is 2255m, and the right intersection point is 2584m. The optimal judgment conditions are: when the distance center is 2408m, the distance gain loss at 2255m is 0.3957dB; when the distance center is 2409m, the distance gain loss at 2255m is 0.4006dB.

[0059] Step 4: Update the current optimal distance segment grid point to R y Record its right intersection point Iterate and update until the maximum value of the segmented region is reached. This yields 9 distances from the center and distance segments:

[0060] Table 1. All results of non-uniform segmentation

[0061] Distance segment number Distance from center (m) Distance segment start (m) End of distance segment (m) 1 2120 2000 2255 2 2408 2255 2584 3 2788 2584 3027 4 3310 3027 3652 5 4073 3652 4604 6 5294 4604 6228 7 7563 6228 9626 8 13237 9626 21185 9 53025 21185 60000

[0062] The time delays of the corresponding 9 near-field multidimensional matching functions are [7.06667e-06, 8.02667e-06, 9.29333e-06, 1.10333e-05, 1.35767e-05, 1.76467e-05, 0.00002521, 4.41233e-05, 0.00017675].

[0063] Step 5: Using the matching functions obtained in Step 4 for the 9 distance segments, and concatenating the matched filter outputs for all 9 distance segments along the t-axis, an approximate output can be obtained:

[0064]

[0065] In this example, the gain loss coefficient introduced by segmented processing is approximately 1.

[0066] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the scope of the technology disclosed in the present invention, and such modifications or substitutions should all be covered within the scope of protection of the present invention.

Claims

1. A method for non-uniformly segmented near-field fast beamforming, characterized in that... The steps are as follows: Step 1: Determine the array aperture and range gain loss tolerance values, and check if the distance from the target to the array center meets the segmentation conditions. If the segmentation conditions are not met, the process ends directly. The segmentation conditions in Step 1 are specifically: determining the distance from the target to the array center. Does it meet the near-field condition? Among them, the distance of the focal center Distance dimension response exist The value at that location is equal to ;in, For range-resolved units, The maximum value of the radar blind zone is set by the user. For the current aperture of the array, This represents the tolerance value for gain loss in the distance dimension. Step 2: Traverse and search for the first optimal distance grid point. The left intersection of its distance dimension response and distance dimension gain tolerance value is the midpoint of the first distance resolution cell, and its right intersection is recorded. Step 3: Traverse and search the next segmented grid point, and determine whether the left intersection of the distance dimension response and the distance gain tolerance value of the point is greater than the right intersection recorded in Step 2. If it is less, continue to search the next segmented grid point. If it is greater, record the previous distance point as the best grid point under the condition that it does not exceed the maximum distance value. Step 4: Calculate the matching coefficient for each distance segment based on the segmented grid points obtained from the traversal search in Step 3; Step 5: Perform matched filtering based on the matching coefficients of each distance segment; concatenate the matched filtering results of each distance segment to obtain an equivalent matched filtering result of the complete distance segment for subsequent processing. The specific steps for calculating the matching coefficient for each distance segment are as follows: in, It is a phase code transmission signal. For the first Segmented grid points of a distance segment, and The response in the distance dimension is respectively The left and right intersections with the distance tolerance value For the latency between sending and receiving nodes, The elevation angle of the target relative to the center of the array. The azimuth angle of the target relative to the center of the array; The matched filtering output based on the matching coefficient of each distance segment is as follows: ， in, For the autocorrelation function of the binary code, This indicates the conjugate operation. and The response in the distance dimension is respectively The left and right intersections with the distance tolerance value.

2. The near-field fast beamforming method with non-uniform segmentation according to claim 1, characterized in that: The specific steps for concatenating the matched filtering results for each distance segment are as follows: All The matched filter outputs at each distance segment are By concatenating the elements along the dimensions, we obtain an approximate output: in, Gain loss coefficient introduced for segmented processing.

3. A computer system, characterized in that... include: One or more processors, a computer-readable storage medium for storing one or more programs, wherein, when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the method of claim 1.

4. A computer-readable storage medium, characterized in that... The device stores computer-executable instructions, which, when executed, are used to implement the method of claim 1.

Images