False alarm point elimination method based on radar array subarray combination

Abstract

The invention relates to a false alarm point elimination method based on radar array subarray combination, which comprises the following steps of: dividing an MIMO (Multiple Input Multiple Output) array of a radar into a plurality of subarrays, and carrying out independent signal processing on each subarray, namely carrying out frequency mixing and filtering on received signals, converting the received signals of each subarray into beat frequency signals, and carrying out signal processing on each subarray; carrying out fast Fourier transform (FFT) calculation, CFAR detection and DOA calculation to generate a potential target point cloud; through the sub-array joint detection step, target point clouds are matched and verified, false alarm points are removed, weighted fusion is carried out in combination with detection results of all sub-arrays, and a final target list is generated; and finally, through post-processing and tracking steps, isolated points are eliminated by using a density clustering algorithm, multi-frame tracking is performed on the target by using a Kalman filter, and possible false alarm points are further eliminated. According to the method, the false alarm rate can be remarkably reduced, especially in the multipath effect and dynamic noise environment, the robustness of target detection is improved at the same time, and it is ensured that the radar works stably and reliably in a complex scene.

Description

Technical Field

[0001] The invention relates to the field of radar technology, and in particular to a false alarm point elimination method based on radar array subarray union. Background Art

[0002] During the use of the radar system, false alarm points will appear due to various influences such as thermal noise (pseudo signals caused by environmental electromagnetic noise), multipath effects (false targets caused by reflection paths), hardware problems (imperfect antenna arrays or signal processing errors), and clutter interference (clutter generated by trees, ground, rain and snow, etc.). However, the real target will be obscured by these false detection results, resulting in missed detection, which will affect the normal operation of the millimeter-wave radar. Therefore, it is very important to effectively eliminate false alarm points. The current common method for eliminating false alarm points is to use a Doppler map or a one-dimensional range image. The Doppler map effect can distinguish between stationary targets and moving targets, which can reduce the false alarm points of stationary targets. In addition, the Doppler map provides the speed information of the target, which helps to identify and track specific speeds. However, in complex environments, the multipath effect makes it difficult to interpret the Doppler map, which results in false alarms. One-dimensional range images can provide structural information of the target, which helps to identify the target type and reduce false alarms. Compared with traditional radar systems, one-dimensional range images can provide higher range resolution, which helps to distinguish close-range targets and have a certain resistance to noise and clutter. However, one-dimensional range images require complex signal processing and large amount of calculation when eliminating false alarm points. The quality is greatly affected by the target posture. Different observation angles may lead to different range images, resulting in false alarms. These two common methods of eliminating false alarm points are easy to eliminate normal targets, resulting in missed detection and affecting the normal operation of millimeter-wave radar. Summary of the invention

[0003] In view of this, the present invention provides a false alarm point elimination method based on radar array sub-array union, which effectively solves the problem of high false alarm rate in multipath effect and dynamic noise environment, and ensures stable and reliable operation in complex scenes.

[0004] To achieve the above object, the present invention provides a false alarm point elimination method based on radar array subarray combination, comprising the following steps: S1, sub-array division and signal processing; The subarray division is to divide the MIMO antenna array of the radar into S subarrays based on the physical position of the array element: array element 1-array element 1+S, array element 2-array element 2+S, ..., array element S-array element S+S, and each subarray processes the signal independently; the S subarrays and a virtual array are subjected to signal processing, and the signal processing includes frequency modulation continuous wave FMCW processing, fast Fourier transform FFT calculation, constant false alarm rate CFAR detection, and radar arrival angle estimation DOA calculation; S2, sub-array joint detection; S201, target matching; The results of the constant false alarm rate CFAR detection of all arrays are spatially compared. If there is a target point cloud If it is detected only in a single array but not in other arrays, the target point cloud is judged is a potential false alarm point; S202, target verification and elimination of false alarm points, verifying the target point cloud according to the sub-array coverage consistency and strength comparison Make judgments; The subarray coverage consistency means that if the target point cloud If it only appears in one array and fails to pass the verification of other arrays, the target point cloud is determined The verification includes distance consistency, speed consistency and space consistency. The distance consistency refers to the target point cloud The distance value The difference between the arrays is less than the preset threshold , the expression is: ; ; in, Indicates The point cloud is in The distance values ​​in the array, Represents different arrays; Satisfying the speed consistency means that the target point cloud The difference in speed values ​​between the arrays is less than the preset threshold , the expression is: ; in, Indicates The point cloud is in The velocity values ​​in an array; Satisfying the spatial consistency means that the target point cloud The difference in the solution angle between the arrays is less than the preset threshold , the expression is: ; Among them, for the same distance, the solution angle value is the angle index, that is, the horizontal coordinate, Indicates The point cloud is in The horizontal coordinate in the array; The intensity comparison verification refers to the target point cloud If a random intensity distribution appears, it is determined to be a false alarm point. The target point cloud The reflection intensity cannot satisfy the following expression: ; in, Indicates The point cloud is in The signal-to-noise ratio in the array; S203, weighted fusion: weighted fusion of the verified target point clouds is performed in combination with the detection strength and confidence of each array to generate a final target point cloud list; S3, post-processing and tracking, use density clustering to cluster the target point cloud to eliminate isolated points, use the Kalman filter to track the target point cloud, and continue to eliminate false alarm points through multi-frame fusion.

[0005] Preferably, signal processing is performed on the divided S sub-arrays and one virtual full array, and the signal processing comprises the following steps: S101, frequency modulated continuous wave (FMCW) processing: using FMCW technology to process the received signal of each subarray, including first mixing the received signal, then filtering the signal, and converting the received signal of each subarray into a beat frequency signal; S102, Fast Fourier Transform (FFT) calculation: the beat frequency signal performs Fast Fourier Transform calculation on the distance and Doppler dimensions respectively to obtain a distance-Doppler two-dimensional graph; S103, constant false alarm rate CFAR detection: using a constant false alarm rate CFAR detection algorithm to detect echo points on the range-Doppler two-dimensional graph to obtain potential targets; S104, radar arrival angle estimation DOA calculation: perform DOA estimation on the potential target, obtain angle information, and generate a point cloud list ,in, Indicates the number of point clouds, Indicates The horizontal coordinate of the point cloud, Indicates The vertical coordinate of the point cloud, Indicates The speed of the point cloud, Indicates The signal-to-noise ratio of a point cloud.

[0006] Preferably, the final target point cloud list generated after weighted fusion is: ; ; in, represents the number of target point clouds that have passed verification, Indicates The horizontal coordinate of the array target point cloud, Indicates The ordinate of the array target point cloud, Indicates The speed of the array target point cloud, Indicates The signal-to-noise ratio of the array target point cloud, Represents the weights of different arrays.

[0007] Compared with the prior art, the present invention has the following beneficial effects: The present invention can effectively eliminate false alarm points caused by thermal noise, multipath effect, hardware problems and clutter interference through the method of subarray division and joint detection, and significantly reduce the false alarm rate under multipath effect and dynamic noise, avoiding the problem of missed detection caused by mistaken deletion of normal targets in the prior art; The present invention can verify the target's distance consistency, speed consistency and spatial consistency in multiple dimensions, and combine weighted fusion to ensure the accuracy and stability of target detection, so that the radar can still maintain high detection performance in complex scenes with a large amount of clutter or dynamic noise; The present invention further eliminates possible false alarm points through multi-frame fusion and Kalman filtering technology, effectively improving the continuity of target tracking, and the present invention fully utilizes the multi-channel information of the radar array to improve the efficiency and accuracy of signal processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] FIG1 is a schematic diagram of the algorithm flow of the present invention; Figure 2 This is a schematic diagram of dividing the 2T3R antenna array according to an embodiment of the present invention; Figure 3 This is an angle diagram of an embodiment of the present invention with poor spatial consistency; Figure 4 This is an angle diagram with good spatial consistency in an embodiment of the present invention. DETAILED DESCRIPTION

[0009] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0010] In order to solve the problem that false alarm point removal is inaccurate and normal targets are easily deleted by mistake in the prior art, the present invention provides the following technical solution: a false alarm point removal method based on radar array subarray combination, comprising the following steps: S1, sub-array division and signal processing; Subarray division: The radar's MIMO antenna array is divided into several subarrays according to the physical position or angular coverage of the array elements, and each subarray processes the signal independently; like Figure 2 As shown, in this embodiment, taking a 2T3R uniform array (2 transmitting antennas, 3 receiving antennas) as an example, the 2T3R radar MIMO antenna array is divided into three sub-arrays based on the physical positions of the array elements: element 1 to element 4, element 2 to element 5, and element 3 to element 6. Together with the virtual full array, there are a total of four arrays. Signal processing is performed on the four arrays, and the signal processing includes the following steps: S101, frequency modulated continuous wave (FMCW) processing: using FMCW technology to process the received signal of each subarray, including first mixing the received signal, then filtering the signal, and converting the received signal of each subarray into a beat frequency signal; S102, Fast Fourier Transform FFT calculation: The beat frequency signal performs Fast Fourier Transform calculation on the distance and Doppler dimensions respectively to obtain a distance-Doppler two-dimensional graph; S103, constant false alarm rate CFAR detection: using a constant false alarm rate CFAR detection algorithm to detect echo points on a range-Doppler two-dimensional graph to obtain potential targets; S104, Radar arrival angle estimation DOA calculation: Perform DOA estimation on potential targets (including fast Fourier transform FFT, digital beamforming DBF, maximum likelihood arrival direction estimation DML method), obtain angle information, and generate a point cloud list ,in, Indicates the number of point clouds, Indicates The horizontal coordinate of the point cloud, Indicates The vertical coordinate of the point cloud, Indicates The speed of the point cloud, Indicates The signal-to-noise ratio of a point cloud.

[0011] S2, sub-array joint detection; S201, target matching, spatially compare the results of the constant false alarm rate CFAR detection of all arrays, if there is a target point cloud If it is detected only in a single array but not in other arrays, the target point cloud is judged is a potential false alarm point; S202, target verification and elimination of false alarm points, verify the target point cloud based on the sub-array coverage consistency and strength comparison Make judgments; Subarray coverage consistency means that if the target point cloud If it only appears in one array and fails to pass the verification of other arrays, the target point cloud is determined is a false alarm point; verification includes distance consistency, speed consistency and space consistency. Meeting distance consistency means that the target point cloud The distance value The difference between the arrays is less than the preset threshold , the expression is: ; ; in, Indicates The point cloud is in The distance values ​​in the array, Represents different arrays; Satisfying speed consistency means that the target point cloud The difference in speed values ​​between the arrays is less than the preset threshold , the expression is: ; in, Indicates The point cloud is in The velocity values ​​in an array; According to the field of view of the antenna and the collective relationship between the subarrays, check the spatial consistency of the target point cloud. False alarm points usually appear at certain angles or are randomly distributed and cannot be verified by all arrays. Meeting spatial consistency means that the target point cloud The difference in the solution angle between the arrays is less than the preset threshold , the expression is: ; in, Indicates The point cloud is in The horizontal coordinate in the array; The spatial consistency is reflected in the horizontal coordinate of the point cloud, that is, the solution angle value, such as Figure 2 The uniform radar array and sub-array division of 2T3R are shown in Figure 1. The angle spectrum and angle resolution values ​​are checked respectively to observe the spatial consistency. The results are as follows: Figure 3 and Figure 4 As shown, the horizontal axis represents the angle index (taking the angle dimension FFT point number as 128 as an example), and the vertical axis represents the energy value of the angle spectrum. Figure 3 The angle diagram with poor spatial consistency is shown, that is, the false alarm point, with a distance of 4.68 and an angle index of 54. At this time, the angle value of the entire array is -0.73125, the angle value of sub-array 1 (element 1-element 4) is 2.34, the angle value of sub-array 2 (element 2-element 5) is 0.36562, and the angle value of sub-array 3 (element 3-element 6) is -4.0219. It can be seen that the angle index of each array is quite different (for the same distance, the angle index represents the angle value, that is, the horizontal axis); Figure 4The angle diagram with good spatial consistency is shown, that is, the target point, the distance is 1.89, the angle index is 75, at this time the full array solution angle value is 0.32484, the sub-array 1 (array element 1-array element 4) solution angle value is 0.32484, the sub-array 2 (array element 2-array element 5) solution angle value is 0.29531, and the sub-array 3 (array element 3-array element 6) solution angle value is 0.26578. The angle index of each array is slightly different. Intensity contrast verification refers to the target point cloud If there is a random intensity distribution, it is judged as a false alarm point. This is because in a multi-array, the reflection intensity RCS of the real target should have a small fluctuation, and the false alarm point usually has a random intensity distribution. If the reflection intensity (expressed as the signal-to-noise ratio (SNR) of the point cloud) fails to satisfy the following expression, it is determined to be a false alarm point. The expression is: ; in, Indicates The point cloud is in The signal-to-noise ratio in the array; That is, only when the point cloud satisfies: ; If the point cloud is a real target point cloud, then the point cloud is considered to be a false alarm point; otherwise, the point cloud is considered to be a false alarm point. S203, weighted fusion: weighted fusion of the verified target point clouds is performed in combination with the detection strength and confidence of each array to generate a final target point cloud list; A high confidence level indicates that the detection strength of the array is better, that is, the signal-to-noise ratio (SNR) of the point cloud is better. Then the final point cloud assigns a higher weight to the array. After weighted fusion, the final target point cloud list is generated as follows: ; in, represents the number of target point clouds that have passed verification, Indicates The horizontal coordinate of the array target point cloud, Indicates The ordinate of the array target point cloud, Indicates The speed of the array target point cloud, Indicates The signal-to-noise ratio of the array target point cloud, represents the weights of different arrays and satisfy: ; S3, post-processing and tracking, use density clustering to cluster the target point cloud to eliminate isolated points, use the Kalman filter to track the target point cloud, and continue to eliminate false alarm points through multi-frame fusion.

[0012] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A false alarm point elimination method based on radar array sub-array combination, characterized in that: The following steps are involved: S1, sub-array division and signal processing; The sub-array division is to divide the MIMO antenna array of the radar into S sub-arrays based on the physical position of the array element: array element 1-array element 1+S, array element 2-array element 2+S, ..., array element S-array element S+S, and each sub-array processes the signal independently; Performing signal processing on the S sub-arrays and one virtual array, wherein the signal processing includes frequency modulation continuous wave FMCW processing, fast Fourier transform FFT calculation, constant false alarm rate CFAR detection, and radar angle of arrival estimation DOA calculation; S2, sub-array joint detection; S201, target matching; The results of the constant false alarm rate CFAR detection of all arrays are spatially compared. If there is a target point cloud If it is detected only in a single array but not in other arrays, the target point cloud is judged is a potential false alarm point; S202, target verification and elimination of false alarm points, verifying the target point cloud according to the sub-array coverage consistency and strength comparison Make judgments; The subarray coverage consistency means that if the target point cloud If it only appears in one array and fails to pass the verification of other arrays, the target point cloud is determined The verification includes distance consistency, speed consistency and space consistency. The distance consistency refers to the target point cloud The distance value The difference between the arrays is less than the preset threshold , the expression is: ； ； in, Indicates The point cloud is in The distance values ​​in the array, Represents different arrays; Satisfying the speed consistency means that the target point cloud The difference in speed values ​​between the arrays is less than the preset threshold , the expression is: ； in, Indicates The point cloud is in The velocity values ​​in an array; Satisfying the spatial consistency means that the target point cloud The difference in the solution angle between the arrays is less than the preset threshold , the expression is: ； Among them, for the same distance, the solution angle value is the angle index, that is, the horizontal coordinate, Indicates The point cloud is in The horizontal coordinate in the array; The intensity comparison verification refers to the target point cloud If a random intensity distribution appears, it is determined to be a false alarm point. The target point cloud The reflection intensity cannot satisfy the following expression: ； in, Indicates The point cloud is in The signal-to-noise ratio in the array; S203, weighted fusion: weighted fusion of the verified target point clouds is performed in combination with the detection strength and confidence of each array to generate a final target point cloud list; S3, post-processing and tracking, use density clustering to cluster the target point cloud to eliminate isolated points, use the Kalman filter to track the target point cloud, and continue to eliminate false alarm points through multi-frame fusion.

2. The method for eliminating false alarm points based on radar array subarray combination according to claim 1, characterized in that: Signal processing is performed on the divided S sub-arrays and one virtual full array, and the signal processing includes the following steps: S101, frequency modulated continuous wave (FMCW) processing: using FMCW technology to process the received signal of each subarray, including first mixing the received signal, then filtering the signal, and converting the received signal of each subarray into a beat frequency signal; S102, Fast Fourier Transform (FFT) calculation: the beat frequency signal performs Fast Fourier Transform calculation on the distance and Doppler dimensions respectively to obtain a distance-Doppler two-dimensional graph; S103, constant false alarm rate CFAR detection: using a constant false alarm rate CFAR detection algorithm to detect echo points on the range-Doppler two-dimensional graph to obtain potential targets; S104, radar arrival angle estimation DOA calculation: perform DOA estimation on the potential target, obtain angle information, and generate a point cloud list ,in, Indicates the number of point clouds, Indicates The horizontal coordinate of the point cloud, Indicates The vertical coordinate of the point cloud, Indicates The speed of the point cloud, Indicates The signal-to-noise ratio of a point cloud.

3. The method for eliminating false alarm points based on radar array subarray combination according to claim 1, characterized in that: The final target point cloud list generated after weighted fusion is: ； ； in, represents the number of target point clouds that have passed verification, Indicates The horizontal coordinate of the array target point cloud, Indicates The ordinate of the array target point cloud, Indicates The speed of the array target point cloud, Indicates The signal-to-noise ratio of the array target point cloud, Represents the weights of different arrays.

Images