Coherent accumulation method, target detection method, device, equipment and radar system

Abstract

The embodiment of the invention provides a coherent accumulation method, a target detection method, devices, equipment and a radar system, and the method comprises the steps: obtaining a point corresponding to a to-be-confirmed point in a next frame for the to-be-confirmed point of a current frame; and based on the angle of the to-be-confirmed point in the current frame, performing coherent accumulation on the data of the plurality of antenna channels of the to-be-confirmed point in the point corresponding to the next frame to obtain a coherent accumulation value of the to-be-confirmed point in the point corresponding to the next frame. According to the invention, coherent accumulation is carried out on the point corresponding to the to-be-confirmed point in the next frame by using the angle of the to-be-confirmed point in the current frame, the scanning angle during coherent accumulation can be directly determined, variable search or traversal does not need to be carried out, the calculation amount of coherent accumulation is reduced, and the efficiency of coherent accumulation is improved.

Description

Technical Field

[0001] This application relates to the field of radar technology, and in particular to a coherent accumulation method, target detection method, apparatus, equipment and radar system. Background Technology

[0002] When performing target detection based on radar signals, the signal-to-noise ratio of the echo signal can be improved through coherent accumulation, thereby enhancing the radar's detection capability, especially the detection capability of weak targets.

[0003] Coherent accumulation requires that the signals from multiple channels be essentially in phase. When a target moves within the radar beam, its echo signal may exhibit a phenomenon spanning range cells, thus affecting the radar's detection performance. Since the target's direction of motion is unknown, related technologies often set variables such as scanning angle, range, and velocity, and perform target detection by iterating through the results of coherent accumulation under different variables. This method requires multiple coherent accumulations, resulting in a large computational load and poor real-time performance. Summary of the Invention

[0004] This application provides a coherent accumulation method, target detection method, apparatus, device, and radar system. It realizes the determination of the angle of the point to be confirmed in the current frame by using the information of the point to be confirmed in the previous frame, including angle, distance index, and Doppler index, and performs coherent accumulation on the point corresponding to the point to be confirmed in the next frame. It does not require variable scanning such as angle scanning, has a small amount of computation, and improves the efficiency of coherent accumulation.

[0005] The embodiments of this application provide the following technical solutions:

[0006] A first aspect of this application provides a coherent accumulation method, comprising: for a point to be confirmed in the current frame, obtaining the point corresponding to the point to be confirmed in the next frame; and coherently accumulating the data of multiple antenna channels of the point corresponding to the point to be confirmed in the next frame based on the angle of the point to be confirmed in the current frame, to obtain the coherent accumulation value of the point corresponding to the point to be confirmed in the next frame.

[0007] In one possible implementation, obtaining the point corresponding to the point to be confirmed in the next frame includes: determining the Doppler index of the point to be confirmed in the next frame; using the Doppler index of the point to be confirmed in the next frame as the Doppler index and the distance index of the point to be confirmed as the distance index to obtain the point corresponding to the point to be confirmed in the next frame.

[0008] In one possible implementation, determining the Doppler index of the point to be confirmed in the next frame includes: determining the Doppler index of the point to be confirmed in the next frame based on the velocity obtained after deblurring the point in the current frame.

[0009] In one possible implementation, the Doppler index of the point to be confirmed in the next frame is determined based on the velocity obtained after deblurring the point in the current frame, including: determining the blur velocity of the point to be confirmed in the next frame based on the velocity obtained after deblurring the point in the current frame and the detection velocity limit of the next frame; and determining the Doppler index of the point to be confirmed in the next frame based on the blur velocity of the point to be confirmed in the next frame.

[0010] In one possible implementation, the blur velocity of the point to be confirmed in the next frame is determined based on the velocity obtained after deblurring the point in the current frame and the detection velocity limit of the next frame. This includes: taking the integer part and the fractional part of the quotient of the velocity obtained after deblurring the point in the current frame divided by the detection upper limit velocity of the next frame; if the integer part is odd, then the blur velocity of the point to be confirmed in the next frame is calculated based on the fractional part and the first relation.

[0011] In one possible implementation, the method further includes: if the integer part is even, calculating the blur rate of the point to be confirmed in the next frame based on the fractional part and the second relation.

[0012] In one possible implementation, determining the Doppler index of the point to be confirmed in the next frame based on the blur velocity of the point to be confirmed in the next frame includes: determining the target Doppler index based on the blur velocity of the point to be confirmed in the next frame; obtaining a reference point by using the target Doppler index as the Doppler index and the distance index of the point to be confirmed as the distance index; and searching for the Doppler index of the point to be confirmed in the next frame within a preset range of the reference point in the distance-Doppler map of the next frame.

[0013] In one possible implementation, the Doppler index of the point to be confirmed in the next frame is obtained from a preset range of reference points in the distance-Doppler image of the next frame. This includes: determining the Doppler index corresponding to the maximum value within the preset range of reference points in the distance-Doppler image of the next frame, which is the Doppler index of the point to be confirmed in the next frame.

[0014] In one possible implementation, the target Doppler index is determined based on the blur velocity of the point to be confirmed in the next frame, including: rounding the quotient of the blur velocity of the point to be confirmed in the next frame divided by the velocity resolution of the next frame to obtain the initial Doppler index; if the blur velocity of the point to be confirmed in the next frame is less than 0, the target Doppler index is determined to be the sum of the maximum values ​​of the initial Doppler index and the Doppler index of the next frame.

[0015] In one possible implementation, the method further includes: if the blur velocity of the point to be confirmed in the next frame is greater than or equal to 0, then the target Doppler index is determined to be the initial Doppler index plus 1.

[0016] In one possible implementation, based on the angle of the point to be confirmed in the current frame, coherent accumulation is performed on the data of multiple antenna channels of the point corresponding to the point in the next frame to obtain the coherent accumulation value of the point corresponding to the point in the next frame. This includes: performing phase compensation on the data of multiple antenna channels of the point corresponding to the point in the next frame to obtain phase-compensated data; calculating the weighting coefficients of multiple antenna channels based on the angle of the point to be confirmed in the current frame; and performing coherent accumulation on the phase-compensated data based on the weighting coefficients of multiple antenna channels to obtain the coherent accumulation value of the point corresponding to the point in the next frame.

[0017] In one possible implementation, the point to be confirmed is a point in a long-distance interval whose value exceeds a pre-detection threshold; the distance index of the long-distance interval is greater than a preset distance index.

[0018] In one possible implementation, the method further includes: determining the pre-detection threshold of each distance cell in the long-distance interval of the current frame based on the noise floor estimate of each distance cell and the detection threshold factor of the current frame; and determining points in the long-distance interval of the current frame that exceed the pre-detection threshold of their respective distance cells as points to be confirmed.

[0019] In one possible implementation, the method further includes: performing noise floor estimation on the Doppler data of each distance cell in the long-range interval of the current frame to obtain the noise floor estimate value of each distance cell in the long-range interval of the current frame.

[0020] In one possible implementation, the method further includes: performing noncoherent accumulation on the range-Doppler matrix of the current frame to obtain a noncoherent accumulation matrix of the current frame; and estimating the noise floor of the Doppler data of each distance cell in the long-range interval of the noncoherent accumulation matrix of the current frame to obtain the noise floor estimate of each distance cell in the long-range interval of the current frame.

[0021] In one possible implementation, noise floor estimation is performed on the Doppler data of each distance cell in the long-range interval of the noncoherent accumulation matrix of the current frame to obtain the noise floor estimate of each distance cell in the long-range interval of the current frame. This includes: calculating the mean of the Doppler data of each distance cell in the long-range interval of the noncoherent accumulation matrix of the current frame; repeatedly replacing the maximum value in the Doppler data of the distance cell with the mean until the difference between two adjacent mean values ​​is less than a preset value to obtain the Doppler data of the distance cell after adjustment; and determining the noise floor estimate of the distance cell based on the variance of the Doppler data of the distance cell after adjustment.

[0022] In one possible implementation, the method further includes: determining the detection threshold factor of the current frame based on the signal-to-noise ratio of the first type of points and the second type of points in the previous frame to be confirmed.

[0023] In one possible implementation, the detection threshold factor of the current frame is determined based on the signal-to-noise ratio (SNR) of the first type of points and the second type of points in the previous frame to be confirmed. This includes: calculating the minimum SNR of the first type of points in the previous frame to obtain a first SNR; calculating the maximum SNR of the second type of points in the previous frame to obtain a second SNR; and determining the detection threshold factor of the current frame based on the first SNR and the second SNR.

[0024] In one possible implementation, determining the detection threshold factor of the current frame based on a first signal-to-noise ratio and a second signal-to-noise ratio includes: determining the smaller value of the first signal-to-noise ratio and the second signal-to-noise ratio as the detection threshold factor of the current frame.

[0025] In one possible implementation, the method further includes: classifying the points to be confirmed in the previous frame based on the coherent accumulation gain of the points to be confirmed in the previous frame to obtain the first type of points and the second type of points in the previous frame; the coherent accumulation gain of the points to be confirmed in the previous frame is determined based on the coherent accumulation value of the points corresponding to the points to be confirmed in the current frame.

[0026] In one possible implementation, the points to be confirmed in the previous frame are classified based on the coherent accumulation gain of the points to be confirmed in the previous frame, resulting in first-class points and second-class points in the previous frame. This includes: determining the points to be confirmed whose sum of signal-to-noise ratio and coherent accumulation gain is greater than the constant false alarm rate threshold as first-class points, and determining the points to be confirmed whose sum of signal-to-noise ratio and coherent accumulation gain is less than or equal to the constant false alarm rate threshold as second-class points; the constant false alarm rate threshold is the detection threshold used in constant false alarm rate detection.

[0027] In one possible implementation, the method further includes: performing constant false alarm rate (CFAR) detection on the coherently accumulated data of the previous frame to obtain the CFAR threshold for each point to be confirmed in the previous frame; wherein, the data of the point to be confirmed in the coherently accumulated data of the previous frame is the coherently accumulated value of the point to be confirmed in the previous frame.

[0028] In one possible implementation, the method further includes: obtaining an initial detection threshold factor, and using the initial detection threshold factor as the detection threshold factor for the first frame.

[0029] In one possible implementation, the method further includes determining an initial detection threshold factor based on radar parameters.

[0030] In one possible implementation, the initial detection threshold factor is determined based on the radar parameters, including: determining the minimum resolvable signal-to-noise ratio of the radar based on the radar parameters and the radar equations; and determining the initial detection threshold factor based on the minimum resolvable signal-to-noise ratio.

[0031] In one possible implementation, the method further includes: if the number of points to be confirmed obtained based on the pre-detection threshold of the first frame is less than a preset number, then the initial detection threshold factor is reduced, so as to redetermine the pre-detection threshold of the first frame based on the reduced initial detection threshold factor.

[0032] A second aspect of this application provides a target detection method, comprising: performing target detection based on data coherently accumulated from each frame; wherein, the data of the point corresponding to the point to be confirmed in the previous frame in the data coherently accumulated from each frame is the coherent accumulation value of that point; the coherent accumulation value is obtained based on the method provided in the first aspect of this application.

[0033] In one possible implementation, target detection is performed based on the data accumulated by coherence across frames, including: target detection based on constant false alarm rate (CFAR) detection, which involves performing target detection on the data accumulated by coherence across frames.

[0034] A third aspect of this application provides a coherent accumulation device, comprising: a corresponding point acquisition module, configured to acquire, for a point to be confirmed in the current frame, a point corresponding to the point to be confirmed in the next frame; and a coherent accumulation module, configured to coherently accumulate data of multiple antenna channels of the point to be confirmed in the current frame based on the angle of the point to be confirmed in the current frame, and to obtain the coherent accumulation value of the point corresponding to the point to be confirmed in the next frame.

[0035] A fourth aspect of this application provides a target detection apparatus for: performing target detection based on data coherently accumulated from each frame; wherein, the data of the point corresponding to the point to be confirmed in the previous frame in the data coherently accumulated from each frame is the coherent accumulation value of that point; the coherent accumulation value is obtained based on the method provided in the first aspect of this application.

[0036] A fifth aspect of this application provides an electronic device, including a memory and a processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, causing the processor to perform the method provided in the first or second aspect of this application.

[0037] A sixth aspect of this application provides a radar system including a control unit; the control unit is used to execute the method provided in the first or second aspect of this application.

[0038] A seventh aspect of this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method provided in the first or second aspect of this application.

[0039] An eighth aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the method provided in the first or second aspect of this application.

[0040] The coherent accumulation method, target detection method, apparatus, device, and radar system provided in this application, when performing coherent accumulation on radar data of each frame, in order to determine the angle of the reflected signal point including the target point, utilizes the mapping relationship of the same point to be confirmed in adjacent frames. Based on the angle of the point to be confirmed in the current frame, coherent accumulation is performed on the data of multiple antenna channels of the point corresponding to the point to be confirmed in the next frame. Taking advantage of the characteristic that the angle of the same target point in adjacent frames remains basically unchanged, the angle of the point to be confirmed in the previous frame is used to perform coherent accumulation on the corresponding point in the current frame, thereby improving the signal-to-noise ratio and the accuracy of target detection. Since it is not necessary to search for multiple possible angles, but directly uses the known angle of the previous frame, the computational load of coherent accumulation is reduced, and the efficiency of target detection is improved.

[0041] In addition to the technical problems solved by the embodiments of this application, the technical features constituting the technical solutions, and the beneficial effects brought about by the technical features of these technical solutions described above, other technical problems that can be solved by the memory allocation method, data processing method, device, and accelerator chip provided by the embodiments of this application, other technical features included in the technical solutions, and the beneficial effects brought about by these technical features will be further described in detail in the specific implementation. Attached Figure Description

[0042] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0043] Figure 1 This is a schematic diagram of the coherent accumulation process in existing technologies;

[0044] Figure 2 A flowchart illustrating the coherent accumulation method provided in the embodiments of this application. Figure 1 ;

[0045] Figure 3 A schematic diagram illustrating the Doppler index search process of the point to be confirmed in the next frame, as provided in an embodiment of this application;

[0046] Figure 4 A flowchart illustrating the coherent accumulation method provided in the embodiments of this application. Figure 2 ;

[0047] Figure 5 A schematic diagram of the layout of a radar antenna array provided in an embodiment of this application;

[0048] Figure 6 A flowchart illustrating a pre-detection threshold determination method provided in this application embodiment. Figure 1 ;

[0049] Figure 7 A flowchart illustrating the pre-detection threshold determination method provided in this application embodiment. Figure 2 ;

[0050] Figure 8 A flowchart illustrating the pre-detection stage provided in an embodiment of this application;

[0051] Figure 9 A schematic diagram of the RDM and pre-detection threshold provided in the embodiments of this application;

[0052] Figure 10 For this application Figure 7 A flowchart illustrating step S706 in the illustrated embodiment;

[0053] Figure 11 A flowchart illustrating the pre-detection threshold determination method provided in this application embodiment. Figure 3 ;

[0054] Figure 12 A schematic diagram of the long-distance interval pre-detection results in an exemplary embodiment provided in this application;

[0055] Figure 13a This is a schematic diagram of the CFAR detection results before coherent accumulation, as exemplified in this application.

[0056] Figure 13b This is a schematic diagram of the CFAR detection results after coherent accumulation, which is an exemplary embodiment of this application.

[0057] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0058] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0059] Radar is widely used in intelligent driving and driver assistance technologies. To ensure the angular resolution and positioning accuracy of radar, a multiple-input multiple-output (MIMO) antenna array, containing multiple transmitting antennas and multiple receiving antennas, is generally used. During radar signal processing, the echo signal returned by each antenna channel is typically processed into a range-Doppler matrix (RDM), and the CFAR detection method is usually used to detect and identify targets within the RDM.

[0060] To improve the signal-to-noise ratio (SNR), coherent accumulation of the RDM of multiple antenna channels is usually required before CFAR detection. This enhances the energy at the target point, while the noise energy is partially weakened after accumulation due to its random phase, thus achieving the goal of improving the SNR.

[0061] Coherent accumulation requires determining the weighting coefficients of the antenna channel. The angle of the incident signal affects these coefficients; the angle of the incident signal is the azimuth or elevation angle of the signal source relative to the receiving device. In radar systems, coherent accumulation precedes angle estimation; therefore, the angle of the signal source cannot be determined during coherent accumulation of the current frame.

[0062] Figure 1 A schematic diagram of the coherent accumulation process in existing technologies, such as... Figure 1 As shown, after obtaining the echo data output from each antenna channel of the radar, a Fast Fourier Transform (FFT) is used to generate the signal for each antenna channel. Figure 1 The RDM of antenna channels 1 to N is described. During coherent accumulation, each angle is scanned within a certain angle range. Coherent accumulation of the RDM of each antenna channel is performed using the weighted vector at each angle to obtain the coherently accumulated data, which is then sent to the CFAR engine for detection. If a target is detected, the angle scanning stops and coherent accumulation is completed; if no target is detected, the angle scanning continues to perform another coherent accumulation using the weighted vector at the new angle.

[0063] The aforementioned angle scanning method often requires multiple coherent accumulations, and it also requires setting a reasonable angle range and scanning step size, resulting in a large amount of computational load for coherent accumulation, making it unsuitable for batch scenarios.

[0064] To address the aforementioned issues, this application provides a coherent accumulation method that determines the points mapped between adjacent frames, uses the angles of a set of mapped points in the previous frame to coherently accumulate the data from multiple antenna channels of the mapped points in the subsequent frame, and uses the angle of the mapped point in the previous frame as the angle of that point, eliminating the need for angle scanning, reducing the computational load of coherent accumulation, and improving the efficiency of target detection.

[0065] To make the above-mentioned objectives, features, and advantages of the embodiments of this application more apparent and understandable, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0066] Figure 2 A flowchart illustrating the coherent accumulation method provided in the embodiments of this application. Figure 1 This method can be executed by any module with corresponding data processing capabilities, such as a device specifically designed for coherent accumulation, a target detection device, or a processor in a radar system. Figure 2 As shown, the coherent accumulation method includes the following steps:

[0067] Step S201: For the point to be confirmed in the current frame, obtain the point corresponding to the point to be confirmed in the next frame.

[0068] The distance index and Doppler index of the point to be confirmed in the next frame are the distance index and Doppler index of the determined point to be confirmed in the next frame. The distance index of the point to be confirmed in the next frame can be the distance index of the point to be confirmed.

[0069] The distance index is used to represent the position of the echo signal of the corresponding point in the distance dimension of the RDM, while the Doppler index represents the position of the echo signal of the corresponding point in the Doppler dimension.

[0070] Optionally, obtaining the point corresponding to the point to be confirmed in the next frame includes: determining the Doppler index of the point to be confirmed in the next frame; using the Doppler index of the point to be confirmed in the next frame as the Doppler index and the distance index of the point to be confirmed as the distance index to obtain the point corresponding to the point to be confirmed in the next frame.

[0071] For distance index, the distance index of the point to be confirmed in the current frame can be directly determined as the distance index of the point to be confirmed in the next frame, that is, the distance index of the point to be confirmed in the next frame; for Doppler index, the Doppler index of the point to be confirmed in the current frame can be determined based on the velocity of the point to be confirmed in the previous frame.

[0072] The velocity of the point to be confirmed in each frame can be provided by the radar's signal processing module.

[0073] The point to be confirmed in the next frame is usually also the point to be confirmed in the next frame. This step realizes the mapping of points to be confirmed in different frames.

[0074] Because the radar operates in TDMA (Time Division Multiple Address) mode, velocity ambiguity occurs when the Doppler frequency of the echo is greater than half the repetition frequency of the radar's transmitted pulse. Generally, CRT (Chinese Remainder Theorem) can be used for deambiguation to obtain the true velocity. However, this method causes the Doppler index of the same point to change in adjacent RDM frames, making it impossible to directly use the Doppler index of the current frame as the Doppler index of the next frame; a search within a certain range is required. Since CRT-based velocity deambiguation requires at least two frames of data to obtain the true velocity, this application starts coherent accumulation from the third frame as the first frame.

[0075] Optionally, determining the Doppler index of the point to be confirmed in the next frame includes: determining the Doppler index of the point to be confirmed in the next frame based on the velocity obtained after deblurring the point in the current frame.

[0076] The speed obtained after deblurring can be called the true speed or the unblurred speed. Any algorithm can be used to deblurr and obtain the true speed corresponding to the blurred speed, such as a CRT-based speed deblurring algorithm.

[0077] After deblurring the blurred velocity of the point to be confirmed in the current frame to obtain the true velocity, the true velocity of the point to be confirmed in the current frame is mapped to the space where the Doppler dimension of the next frame is located to obtain the Doppler index of the point to be confirmed in the next frame.

[0078] By resolving velocity ambiguity, a mapping relationship for the same point is established in adjacent frames, providing a reliable foundation for coherent accumulation based on information from the same point in the previous frame.

[0079] Optionally, based on the velocity obtained after deblurring the point to be confirmed in the current frame, the Doppler index of the point to be confirmed in the next frame is determined, including: determining the blur velocity of the point to be confirmed in the next frame based on the velocity obtained after deblurring the point to be confirmed in the current frame and the detection velocity limit of the next frame; and determining the Doppler index of the point to be confirmed in the next frame based on the blur velocity of the point to be confirmed in the next frame.

[0080] The detection rate limit for a given frame is the product of the maximum value of the Doppler index in the RDM, such as 64, and the Doppler resolution of that frame.

[0081] The detection speed limit includes an upper limit and a lower limit, with the lower limit having the opposite speed direction to the upper limit. When the fuzzy speed is greater than or equal to 0, the detection speed limit is the upper limit; when the fuzzy speed is less than 0, the detection speed limit is the lower limit, to maintain consistency with the direction of the fuzzy speed.

[0082] Specifically, the blurred velocity of the point to be confirmed in the next frame can be determined by dividing the true velocity of the point to be confirmed in the current frame (i.e., the velocity obtained after deblurring) by the detection velocity limit of the next frame. The calculated quotient can be substituted into a calculation formula, and the blurred velocity of the point to be confirmed in the next frame can be determined through this formula.

[0083] After obtaining the blurred velocity of the point to be confirmed in the next frame, the blurred velocity is mapped to the space where the RDM Doppler dimension of the next frame is located, and the Doppler index of the point to be confirmed in the next frame can be obtained.

[0084] To improve the accuracy of Doppler index determination, a search can be performed within a smaller range to obtain the Doppler index of the point to be confirmed in the next frame.

[0085] Optionally, based on the blur velocity of the point to be confirmed in the next frame, the Doppler index of the point to be confirmed in the next frame is determined, including: determining the target Doppler index based on the blur velocity of the point to be confirmed in the next frame; using the target Doppler index as the Doppler index and the distance index of the point to be confirmed as the distance index to obtain a reference point; and searching for the Doppler index of the point to be confirmed in the next frame from a preset range of the reference point in the distance-Doppler image of the next frame.

[0086] The preset range is used to characterize the region near the reference point, such as the region located in the same Doppler cell as the reference point. The range-Doppler map is used to show the distribution of echo intensity or relative echo intensity in two dimensions: range and Doppler.

[0087] For example, the preset range can be a square or circular area centered on a reference point.

[0088] After obtaining the blurred velocity of the point to be confirmed in the next frame, the blurred velocity is mapped to the space where the RDM Doppler dimension of the next frame is located, and the Doppler index corresponding to the reference point, i.e. the target Doppler index, can be obtained; based on the echo intensity of each point within the preset range of the reference point in the distance-Doppler image of the next frame, the Doppler index of the point to be confirmed in the next frame is obtained.

[0089] For example, the Doppler index of the point with the strongest echo intensity within the preset range of the reference point can be directly determined as the Doppler index of the point to be confirmed in the next frame; or the Doppler index of the point with echo intensity exceeding the echo intensity of the reference point by at least the preset intensity within the preset range can be determined as the Doppler index of the point to be confirmed in the next frame.

[0090] Optionally, the Doppler index of the point to be confirmed in the next frame can be obtained from a preset range of reference points in the distance-Doppler image of the next frame. This includes: determining the Doppler index corresponding to the maximum value within the preset range of reference points in the distance-Doppler image of the next frame, which is the Doppler index of the point to be confirmed in the next frame.

[0091] The Doppler index corresponding to the maximum value within the preset range of the reference point is taken from the Doppler index of the point with the highest echo intensity in the distance-Doppler map, which is located within the preset range of the reference point.

[0092] Figure 3 This is a schematic diagram illustrating the Doppler index search process of the point to be confirmed in the next frame, as provided in the embodiments of this application. Figure 3 For a point to be confirmed in the Nth frame, the blurred velocity of the point to be confirmed in the N+1th frame is determined by the true velocity of the point to be confirmed. Based on the blurred velocity and the distance index of the point to be confirmed in the Nth frame, a reference point is mapped in the distance-Doppler map of the N+1th frame. With the reference point as the center, the point corresponding to the point to be confirmed in the N+1th frame is searched from the preset range of the reference point.

[0093] Using the Doppler index of the point with the highest echo intensity as the Doppler index of the point to be confirmed in the next frame results in high search efficiency and low logical complexity. Taking echo intensity into account during the search improves the accuracy of mapping the point to be confirmed across different frames.

[0094] Step S202: Based on the angle of the point to be confirmed in the current frame, perform coherent accumulation of the data of multiple antenna channels of the point corresponding to the point to be confirmed in the next frame to obtain the coherent accumulation value of the point corresponding to the point to be confirmed in the next frame.

[0095] After determining the point to be confirmed in the next frame, during coherent accumulation, the weighting coefficients of each channel in multiple antenna channels can be determined using the angle of the point to be confirmed in the current frame (including at least one of azimuth and elevation angles). Based on the determined weighting coefficients, the data of multiple antenna channels of the point to be confirmed in the next frame are superimposed to obtain the coherent accumulation value of the corresponding point. The coherent accumulation value is then used to replace the original value of the corresponding point to obtain the coherently accumulated data of the next frame.

[0096] Since the weighting coefficients are also affected by the array manifold of the antenna, weighting formulas for different array manifolds can be established in advance. Based on the array manifold of the radar antenna, the target weighting formula can be determined from multiple weighting formulas. The angle of the point to be confirmed is substituted into the target weighting formula to obtain the weighting coefficients of multiple antenna channels, thereby realizing the coherent accumulation of data from multiple antenna channels of the point to be confirmed in the next frame.

[0097] After coherently accumulating the data from multiple antenna channels at a point, such as RDM or RMDnc, the coherent accumulation value of that point can be obtained. Here, RMDnc is the matrix obtained after non-coherent accumulation of RDM data from multiple antenna channels in the same frame.

[0098] After obtaining the coherently accumulated data for each frame, target points can be detected from the points to be confirmed in each frame using CFAR detection, thus achieving target detection.

[0099] The coherent accumulation method provided in this embodiment, when performing coherent accumulation on radar data of each frame, in order to determine the angle of the reflected signal point including the target point, implements the mapping relationship of the same point to be confirmed in adjacent frames. Based on the angle of the point to be confirmed in the current frame, it performs coherent accumulation on the data of multiple antenna channels of the point corresponding to the point to be confirmed in the next frame. Taking advantage of the characteristic that the angle of the same target point in adjacent frames is basically unchanged, it uses the angle of the point to be confirmed in the previous frame to perform coherent accumulation on the corresponding point in the current frame, thereby improving the signal-to-noise ratio and the accuracy of target detection. Since it does not need to search for multiple possible angles, but directly uses the known angle of the previous frame, it reduces the computational load of coherent accumulation and improves the efficiency of target detection.

[0100] Figure 4 A flowchart illustrating the coherent accumulation method provided in the embodiments of this application. Figure 2 This embodiment is in Figure 2 Based on the illustrated embodiment, the steps for determining the Doppler index and the steps for coherent accumulation are further refined. For example... Figure 4 As shown, the coherent accumulation method provided in this embodiment may specifically include the following steps:

[0101] Step S401: For the point to be confirmed in the current frame, take the integer part and the fractional part of the quotient of the speed obtained after deblurring the point in the current frame and dividing it by the detection upper limit speed of the next frame.

[0102] Let V be the velocity of the point to be confirmed after deblurring in the current frame. N The detection speed limit for the next frame is V. N+1 max Then the integer part in can be expressed as: in = |floor(V N / V N+1 max )|, where floor() is the floor function, used to return the largest integer not greater than a real number; the decimal part de can be expressed as: de = |V N / V N+1 max |-in.

[0103] Step S402: If the integer part is odd, then calculate the blur speed of the point to be confirmed in the next frame based on the fractional part and the first relation.

[0104] Step S403: If the integer part is even, calculate the blur rate of the point to be confirmed in the next frame based on the fractional part and the second relation.

[0105] After obtaining the integer part in and the fractional part de, determine whether the integer part in is odd; if so, the blur velocity Vˊ of the point to be confirmed in the next frame is calculated based on the first relation; if not, Vˊ is calculated based on the second relation.

[0106] The first relation is a piecewise relation: V N If the value is greater than or equal to 0, then Vˊ=(de-1)×V N+1 max If V N If less than 0, then Vˊ=-(de-1)×V N+1 max The second relation is a piecewise relation: if V N If V' is greater than or equal to 0, then V' = de × V N+1 max If V N If less than 0, then Vˊ=-de×V N+1 max .

[0107] Step S404: The quotient of the blur velocity of the point to be confirmed in the next frame divided by the velocity resolution of the next frame is rounded to obtain the initial Doppler index.

[0108] After obtaining the blurred velocity Vˊ of the point to be confirmed in the next frame, the initial Doppler index is determined as the rounded result of the quotient of Vˊ and the velocity resolution of the next frame.

[0109] For example, the initial Doppler index can be represented as: round(Vˊ / V) N+1 res ), where V N+1 res The velocity resolution is set to the next frame, and round() is the rounding function.

[0110] Step S405: If the blur velocity of the point to be confirmed in the next frame is less than 0, then the target Doppler index is determined to be the sum of the maximum values ​​of the initial Doppler index and the Doppler index of the next frame.

[0111] Step S406: If the blur velocity of the point to be confirmed in the next frame is greater than or equal to 0, then the target Doppler index is determined to be the initial Doppler index plus 1.

[0112] When Vˊ is less than 0, the target Doppler index is: round(Vˊ / V) N+1 res When Vˊ is greater than or equal to 0, the target Doppler index is: round(Vˊ / V) + dBins; N+1 res )+1, where dBins is the maximum value of the Doppler index of the next frame.

[0113] The reference point is obtained by using the target Doppler index as the Doppler index and the distance index of the point to be confirmed in the current frame as the distance index.

[0114] Step S407: Determine the Doppler index corresponding to the maximum value within the preset range of the reference point in the distance-Doppler image of the next frame, which is the Doppler index of the point to be confirmed in the next frame.

[0115] Step S408: Using the Doppler index of the point to be confirmed in the next frame as the Doppler index and the distance index of the point to be confirmed as the Doppler index, the point corresponding to the point to be confirmed in the next frame is obtained.

[0116] Step S409: Perform phase compensation on the data of multiple antenna channels of the point to be confirmed in the next frame to obtain the phase-compensated data.

[0117] Since the actual velocity of the point to be confirmed in the current frame is known, phase compensation can be performed on the data of multiple antenna channels of the point corresponding to the point in the next frame based on the actual velocity of the point to be confirmed in the current frame, so as to obtain the phase-compensated data.

[0118] For example, Figure 5 This is a schematic diagram of the layout of a radar antenna array provided in an embodiment of this application, as shown below. Figure 5 As shown, the antenna array includes 3 transmitting antennas and 4 receiving antennas. The spacing between the transmitting antennas is 2λ, and the spacing between the receiving antennas is 0.5λ, where λ is the wavelength of the electromagnetic wave. Based on this, 12 virtual antennas can be simulated using signal processing technology.

[0119] for Figure 5 The antenna array shown can have the following coefficients for phase compensation:

[0120] c = [1, 1, 1, 1, e -jΔΦ ,e -jΔΦ ,e -jΔΦ ,e -jΔΦ ,…,e -j(nT-1)ΔΦ ]

[0121]

[0122] Among them, V NnT represents the actual velocity of the point to be confirmed in the Nth frame; nT represents the number of transmitting antennas; T represents the actual velocity of the point to be confirmed in the Nth frame. c This is the transmission period of the linear frequency modulated pulse Chirp. Figure 5 Taking nT as an example of 3.

[0123] Step S410: Calculate the weighting coefficients of multiple antenna channels based on the angle of the point to be confirmed in the current frame.

[0124] For example, Figure 5 The weighted neighboring function w, composed of the weighting coefficients of each antenna channel in the antenna array shown, is:

[0125]

[0126] Where, θ N Let nT be the angle of the point to be confirmed in the Nth frame, nT be the number of transmitting antennas, and nR be the number of receiving antennas.

[0127] Step S411: Based on the weighting coefficients of multiple antenna channels, perform coherent accumulation on the phase-compensated data to obtain the coherent accumulation value of the point to be confirmed in the next frame.

[0128] By using the aforementioned method of mapping adjacent frame points and coherent accumulation, combined with the nature of velocity fuzziness, accurate mapping of adjacent frame points is achieved, improving the accuracy of determining the corresponding point of the point to be confirmed in the next frame; by using the angle of the point to be confirmed in the previous frame to perform coherent accumulation on the corresponding point in the current frame, it is not necessary to traverse multiple angles, thus reducing the computational load of coherent accumulation.

[0129] Optionally, the point to be confirmed is a point in the long-distance interval whose value exceeds the pre-detection threshold. The long-distance interval is an interval whose distance index is greater than a preset distance index.

[0130] The pre-detection threshold is a threshold used in the pre-detection stage, which precedes the target detection stage. The target detection stage is used to detect target points from the unconfirmed points detected in the pre-detection stage, and CFAR detection is typically used.

[0131] By employing the coherent accumulation provided in the aforementioned embodiments for long-range intervals, the detection accuracy of weak targets at long distances is significantly improved, while the range of coherent accumulation is reduced, further reducing the computational load and improving detection efficiency.

[0132] To determine the points to be confirmed in each frame, it is also necessary to first determine the pre-detection threshold for each frame, so that points whose values ​​exceed the pre-detection threshold are identified as points to be confirmed. Figure 6 A flowchart illustrating a pre-detection threshold determination method provided in this application embodiment. Figure 1 ,like Figure 6 As shown, the pre-detection threshold determination method includes the following steps:

[0133] Step S601: Based on the signal-to-noise ratio of the first type of points and the second type of points in the current frame to be confirmed, determine the detection threshold factor for the next frame; wherein, the points to be confirmed in the current frame are obtained by filtering the pre-detection threshold of the current frame.

[0134] The detection threshold factor is used to determine the pre-detection threshold. Points to be confirmed in the current frame are those whose values ​​in the radar data of the current frame exceed the pre-detection threshold of the current frame. The first type of points are those predicted as target points, while the second type are those predicted as noise.

[0135] The radar data in the current frame can be echo data obtained after the echo signal is processed by the antenna channel, or it can be data obtained after processing the echo data, such as RDM, or data obtained after RDM is noncoherently accumulated, etc.

[0136] Step S602: Determine the pre-detection threshold for the next frame based on the detection threshold factor of the next frame.

[0137] A frame is the basic unit of radar data transmission and reception. A radar frame is a complete set of data acquired by the radar during continuous scanning, from start to finish. This set of data can include information such as the distance, velocity, and angle of a point.

[0138] For each frame of radar data collected, the pre-detection threshold of the frame needs to be determined based on the detection threshold factor of that frame; points with values ​​higher than the pre-detection threshold are identified as points to be confirmed; then the points to be confirmed in the frame are divided into two categories, namely the first category and the second category. The first category of points is predicted as target points during pre-screening, while the second category of points is predicted as noise.

[0139] After determining the points to be confirmed in each frame, the signal-to-noise ratio (SNR) of each point is calculated. The SNR is the ratio of the amplitude or power of the radar data for the point to be confirmed to the amplitude or power of the background noise.

[0140] Specifically, in the RDM or the RDM obtained after non-coherent accumulation of the current frame, the noise floor is estimated in units of range bins to obtain the noise floor estimate; the ratio of the value of each point to be confirmed in the RDM or the RDM obtained after non-coherent accumulation to the noise floor estimate of the range bin is calculated to obtain the signal-to-noise ratio of each point to be confirmed in the current frame.

[0141] Optionally, the method further includes: performing noise floor estimation on the Doppler data of each distance cell in the distance-Doppler matrix of the current frame to obtain the noise floor estimate value of each distance cell in the current frame.

[0142] The range-Doppler matrix is ​​a two-dimensional matrix used to represent the distance and relative velocity between a point within the radar's detection range and the radar. Rows represent different range cells, and columns represent different velocity (Doppler) cells. The value of each matrix element reflects the echo intensity of the point under that combination of range and velocity.

[0143] The range-Doppler matrix (RDM) for each frame can be generated based on the Fast Fourier Transform (FFT) of each frame of echo data.

[0144] For each range cell in each frame of RDM, noise floor estimation is performed based on the Doppler data of that range cell to obtain the noise floor estimate value of that range cell.

[0145] For example, multiple sampling points obtained within a distance cell can be determined, and the average power of all sampling points within the distance cell can be determined as the noise floor estimate of the distance cell.

[0146] In some embodiments, the noise floor estimate of the distance cell can be updated by continuously replacing the maximum value with the average power of the sampling points until the difference between two adjacent noise floor estimates is small. The last noise floor estimate is then used as the final output noise floor estimate of the distance cell to calculate the signal-to-noise ratio of the point to be confirmed within the distance cell. In other words, after obtaining the noise floor estimate, the signal-to-noise ratio of the point to be confirmed can be calculated based on the noise floor estimate.

[0147] Since noise may vary with distance, noise floor estimation is performed using distance cells in RDM, which improves the matching degree between the noise floor estimation and the actual noise level. At the same time, it simplifies the complex noise estimation problem into processing each distance cell separately, which simplifies the signal processing flow and reduces computational complexity.

[0148] After obtaining the signal-to-noise ratio (SNR) of each point to be confirmed in the current frame, the detection threshold factor for the next frame can be determined based on the statistical values ​​of the SNR of the first type of points and the statistical values ​​of the SNR of the second type of points. These statistical values ​​can be the minimum, maximum, average, median, etc.

[0149] Specifically, the detection threshold factor for the next frame can be determined based on the minimum signal-to-noise ratio of the first type of points and the maximum signal-to-noise ratio of the second type of points.

[0150] After obtaining the detection threshold factor for the next frame, the product of the detection threshold factor and the preset noise amplitude or power is calculated to obtain the pre-detection threshold for the next frame. Thus, the pre-detection threshold of the next frame is used to filter out the points to be confirmed in the next frame during the pre-detection stage.

[0151] After obtaining the detection threshold factor for the next frame, the detection threshold factor can be multiplied by the noise floor estimate of each distance cell in the next frame to obtain the pre-detection threshold of each distance cell in the next frame. Thus, the pre-detection threshold of each distance cell in the next frame can be used to filter out the points to be confirmed in each distance cell in the next frame during the pre-detection stage.

[0152] In some embodiments, the determination of the point to be confirmed can be performed only for the long-distance interval. That is, in each distance cell within the long-distance interval of each frame, the point to be confirmed for that distance cell is determined based on the pre-detection threshold of the distance cell.

[0153] Optionally, based on the detection threshold factor of the next frame, a pre-detection threshold for the next frame is determined, including: based on the detection threshold factor of the next frame and the noise floor estimate of each distance cell in the next frame, a pre-detection threshold for each distance cell in the next frame is determined, so as to identify points in each distance cell of the next frame that exceed the pre-detection threshold of the distance cell as points to be confirmed.

[0154] Let the noise floor estimate of a certain distance cell in the next frame be σ. 2 If the detection threshold factor for the next frame is α, then the expression for the pre-detection threshold T of this distance cell is: T = α × σ 2 Therefore, in the pre-detection stage, points whose values ​​in the RDM of the next frame exceed the pre-detection threshold of the distance cell are identified as points to be confirmed, and the signal-to-noise ratio of the points to be confirmed is calculated, and so on.

[0155] The pre-detection threshold is dynamically adjusted based on the noise level of the current frame and the previous frame, which improves the accuracy of the pre-detection threshold determination, thereby improving the accuracy of target detection and reducing the occurrence of false alarms.

[0156] The pre-detection threshold determination method provided in this application, specifically for CFAR detection, aims to improve the accuracy of the detection threshold used in pre-detection before CFAR detection. It offers a scheme for dynamically determining the pre-detection threshold, utilizing the signal-to-noise ratio (SNR) distribution of first-class and second-class points in the current frame to be confirmed. Based on this SNR factor, a dynamic pre-detection threshold for detecting points to be confirmed in the next frame is calculated. By dynamically determining the pre-detection threshold using the noise and target SNR levels of the current frame, the accuracy of both pre-detection threshold determination and point confirmation is improved, thereby reducing the computational load and false detection rate of CFAR detection.

[0157] When the point to be confirmed is limited to a point within a long-distance interval, the pre-detection threshold of each distance unit in the long-distance interval can be obtained by adaptively adjusting the pre-detection threshold determination method provided in the embodiments of this application.

[0158] Figure 7A flowchart illustrating the pre-detection threshold determination method provided in this application embodiment. Figure 2 This embodiment is in Figure 6 Based on the illustrated embodiment, the steps for determining the detection threshold factor are further refined, and steps related to noise floor estimation and classification of points to be confirmed are added. For example... Figure 7 As shown, the pre-detection threshold determination method provided in this embodiment may specifically include the following steps:

[0159] Step S701: Obtain the distance-Doppler matrix of multiple antenna channels in the current frame.

[0160] The radar frames can be traversed in chronological order, with the currently traversed frame as the current frame. Based on the echo data from multiple antenna channels in the current frame, the range-Doppler matrices of multiple antenna channels in the current frame can be obtained. Each antenna channel corresponds to a range-Doppler matrix.

[0161] Step S702: Perform noncoherent accumulation of the range-Doppler matrices of multiple antenna channels in the current frame to obtain the noncoherent accumulation matrix of the current frame.

[0162] Noncoherent accumulation is a signal processing technique that directly superimposes the amplitude or power of a signal or data. The value of a point in the noncoherent accumulation matrix is ​​the sum of the values ​​of that point in the distance-Doppler matrix of multiple antenna channels.

[0163] noncoherent accumulation matrix RDM nc It can be represented as:

[0164]

[0165] Among them, RDM i Let be the distance-Doppler matrix of antenna channel i, and M be the total number of antenna channels.

[0166] Noncoherent accumulation can improve the signal-to-noise ratio, thereby improving the accuracy of target detection, especially weak target detection.

[0167] Step S703: Perform noise floor estimation on the Doppler data of each distance cell in the noncoherent accumulation matrix of the current frame to obtain the noise floor estimate value of each distance cell in the current frame.

[0168] Optionally, noise floor estimation is performed on the Doppler data of each distance cell in the noncoherent accumulation matrix of the current frame to obtain the noise floor estimate of each distance cell in the current frame. This includes: calculating the mean of the Doppler data of each distance cell in the noncoherent accumulation matrix of the current frame; repeatedly replacing the maximum value in the Doppler data of the distance cell with the mean until the difference between two adjacent mean values ​​is less than a preset value to obtain the adjusted Doppler data of the distance cell; and determining the noise floor estimate of the distance cell based on the variance of the adjusted Doppler data of the distance cell.

[0169] Each distance cell corresponds to a specific distance range, and the Doppler data of the distance cell is the value of the RDM that falls within that distance cell.

[0170] In RDM, the unit of data can be decibels (dB) to describe relative echo intensity. The default value is a configurable parameter, such as 1.

[0171] To improve the accuracy of noise floor estimation and avoid introducing target data during noise floor estimation, a method was implemented to continuously replace the maximum value with the mean, thereby obtaining a stable noise floor estimate.

[0172] The mean value u of the Doppler data of the distance cell is obtained in the j-th calculation. j Then, the mean u j Replace the maximum value in the Doppler data of the distance cell with the mean u of the replaced Doppler data. j+1 Determine σ j -σ j+1 Is the condition <1 (preset value) satisfied, where σ j For u j The value obtained by converting to decibels, if u j If the unit is decibel, then u j =σ j If not, then continue with u j+1 Replace the maximum value in the Doppler data of that distance cell, and so on; if so, then based on satisfying σ j -σ j+1 σ <1 j+1 Determine the noise floor estimate of the distance cell, for example, determine the noise floor estimate of the distance cell as σ. j+1 The square of.

[0173] Step S704: Based on the detection threshold factor of the current frame and the noise floor estimate of each distance cell in the current frame, determine the pre-detection threshold of each distance cell in the current frame.

[0174] Specifically, for each distance cell in the current frame, the pre-detection threshold of that distance cell is the product of the detection threshold factor of the current frame and the noise floor estimate of that distance cell.

[0175] Step S705: Points in each distance cell of the current frame that exceed the pre-detection threshold of that distance cell are identified as points to be confirmed, and the signal-to-noise ratio of the points to be confirmed is calculated based on the noise floor estimate of the distance cell where the points to be confirmed are located.

[0176] After obtaining the pre-detection threshold of each distance cell in each frame, points to be confirmed are selected from each distance cell based on the pre-detection threshold of each distance cell. That is, points whose values ​​in the RDM exceed the pre-detection threshold of their respective distance cells are determined as points to be confirmed, thus completing the pre-detection stage.

[0177] After predictive detection, we can proceed to the CFAR detection stage, also known as the target detection stage. CFAR detection is used to determine the target point based on the noncoherent accumulation matrix RDM. nc The value in is used to detect the target point from the points to be confirmed.

[0178] Figure 8 This is a flowchart illustrating the pre-detection stage provided in an embodiment of this application, as shown below. Figure 8 As shown, the pre-detection stage mainly includes the following steps:

[0179] Step S801: For each distance interval of the current frame, obtain the Doppler data for that distance interval;

[0180] Step S802: Calculate the mean u of the Doppler data for this distance interval. j This is then converted to a decibel value, denoted as σ. j And determine the position k where the Doppler data in this distance interval reaches its maximum value;

[0181] Step S803, use the mean u j Replace the value at position k of the Doppler data within this distance interval, calculate the mean of the new Doppler data, and convert it to a decibel value, denoted as σ. j+1 ;

[0182] Step S804, determine σ j -σ j+1 Is it less than 1? If not, return to step S802 to replace the current maximum value with the newly obtained mean; if yes, proceed to step S805.

[0183] Step S805, determine σ j+1 The square of is the noise floor estimate for that distance interval.

[0184] Step S806: Based on the detection threshold factor of each distance interval in the current frame and the noise floor estimate of each distance interval in the current frame, determine the pre-detection threshold of each distance interval in the current frame.

[0185] Step S807: Filter points in each distance interval by using a pre-detection threshold, identify points that exceed the pre-detection threshold as points to be confirmed, and save the distance index and Doppler index of the points to be confirmed.

[0186] For the first frame, the detection threshold factor can be a value set based on experience.

[0187] Optionally, the pre-detection threshold determination method further includes: obtaining an initial detection threshold factor, and using the initial detection threshold factor as the detection threshold factor for the first frame.

[0188] The initial detection threshold factor can be a value set based on experience, which is less than the minimum threshold factor value for CFAR detection in the same scene, such as 5dB.

[0189] In the first frame, since there is no previous frame, the detection threshold factor uses the set initial detection threshold factor. Because the detection threshold factor will be dynamically adjusted in subsequent frames, the initial detection threshold factor can be set to a lower value to avoid missed detections.

[0190] Initial detection threshold factors can be pre-assigned for various detection scenarios, resulting in a preset correspondence. Different detection scenarios may involve different radar parameters, or variations in the detected targets, their motion attributes, or the distance between the target and the radar. Based on the current detection scenario, the corresponding initial detection threshold factor can be read from the preset correspondence and used as the detection threshold factor for the first frame of that scenario. This is then combined with radar data from the first frame, such as RDM, RDM... nc The noise floor estimate is used to obtain the pre-detection threshold for the first frame.

[0191] Optionally, the pre-detection threshold determination method also includes: determining an initial detection threshold factor based on radar parameters.

[0192] Specifically, the initial detection threshold can be determined based on the radar's transmit power, antenna gain, signal bandwidth, and the signal processing algorithm used.

[0193] For example, an initial detection threshold can be determined based on the radar's minimum resolvable signal-to-noise ratio. The initial detection threshold can be a value slightly smaller than the minimum resolvable signal-to-noise ratio.

[0194] The minimum resolvable signal-to-noise ratio of the radar can be obtained through prior simulation tests.

[0195] By utilizing the radar's own parameters, the initial detection threshold is determined, which improves the accuracy of the initial detection threshold determination and avoids the slow convergence speed and large amount of computation caused by setting the initial detection threshold too low.

[0196] Optionally, an initial detection threshold factor is determined based on radar parameters, including: determining the minimum resolvable signal-to-noise ratio of the radar based on radar parameters and radar equations; and determining the initial detection threshold factor based on the minimum resolvable signal-to-noise ratio.

[0197] The radar equations are fundamental formulas describing radar performance, establishing the relationship between radar received power and transmitted power, antenna gain, target cross-section, range, and other parameters. Radar performance, in turn, affects the radar's minimum resolvable signal-to-noise ratio (SNR). Higher transmitted power, higher antenna gain, and higher received power all contribute to improving the SNR and reducing the radar's minimum resolvable SNR.

[0198] The minimum resolvable signal-to-noise ratio (SNR) of the radar is calculated by using the radar's own parameters and radar equations, thus achieving automatic determination of the minimum resolvable SNR and improving the automation level of the algorithm.

[0199] Optionally, the pre-detection threshold determination method further includes: if the number of points to be confirmed obtained based on the pre-detection threshold of the first frame is less than a preset number, then the initial detection threshold factor is reduced, so as to redetermine the pre-detection threshold of the first frame based on the reduced initial detection threshold factor.

[0200] The preset quantity is a configurable parameter, which can be a fixed value or determined based on the attributes of the target being detected in the detection scene, the radar's range resolution, and velocity resolution.

[0201] Since the detection threshold factor is positively correlated with the determined pre-detection threshold, after determining the pre-detection threshold for the first frame based on the initial detection threshold factor, if the number of points to be confirmed obtained through the pre-detection stage is too small (less than the preset number), it indicates that the pre-detection threshold is too high, i.e., the initial detection threshold factor is too high. In this case, the initial detection threshold factor is adjusted to a smaller value, and the pre-detection threshold for the first frame is re-determined. The initial detection threshold factor can be adjusted in fixed steps, or the adjustment value can be determined based on the difference between the number of points to be confirmed and the preset number.

[0202] The initial detection threshold factor can be adjusted multiple times until the number of points to be confirmed in the first frame is greater than or equal to the preset number.

[0203] Through the aforementioned steps, the initial detection threshold factor is adjusted based on the feedback mechanism, thus avoiding the problem of missed detections due to an excessively high initial detection threshold factor.

[0204] pass Figure 8 The pre-detection process shown determines the points to be confirmed in each frame. These points are represented using corresponding distance and Doppler indices.

[0205] For example, Figure 9 This is a schematic diagram of the RDM and pre-detection threshold provided in the embodiments of this application, as shown below. Figure 9 As shown, the graph corresponding to RDM includes three coordinate axes: the distance axis (unit: m), the Doppler axis (unit: m / s), and the relative intensity axis (unit: dB). Figure 9 Each square in the diagram corresponds to a distance cell or Doppler cell. Figure 9 The RDM shown can be the RDM obtained after noncoherent accumulation. nc Pre-detection threshold, such as Figure 9 As shown in the darker areas, RDM nc Points with relative intensity exceeding that of the dark area are marked as points to be confirmed.

[0206] After determining the points to be confirmed in each frame, coherent accumulation can be performed on the points to be confirmed to enhance the signal-to-noise ratio of the target points in the points to be confirmed.

[0207] After obtaining the points to be confirmed in each frame through pre-detection, the signal-to-noise ratio of the points to be confirmed is calculated based on the value of the points to be confirmed in the RDM and the distance cell where the points to be confirmed are located, so as to determine the detection threshold factor for the next frame.

[0208] Step S706: Based on the coherent accumulation gain of the points to be confirmed in the current frame, classify the points to be confirmed in the current frame to obtain the first type of points and the second type of points in the current frame.

[0209] The coherent accumulation gain of the point to be confirmed is used to describe the radar data of the point to be confirmed, such as in RDM or RDM. nc The value in the equation represents the degree to which the gain increases after coherent accumulation.

[0210] For example, the coherent accumulation gain g c It can be represented as: g c =10lg(xˊ / x), where x is the value of the point to be confirmed without coherent accumulation, and xˊ is the value of the point to be confirmed after coherent accumulation.

[0211] The value of the uncoordinated accumulation of the point to be confirmed can be the value of the point to be confirmed in the RDM or in the RDM. nc The value in the RDM of the point to be confirmed; the value after coherent accumulation can be the value obtained by coherently accumulating the values ​​of the point to be confirmed in the RDM of multiple antenna channels, or the value of the point to be confirmed in the current frame RDM. nc The value in the RDM corresponds to the point to be confirmed in the next frame. nc The value obtained by accumulating the values ​​in the reference.

[0212] For example, points to be confirmed where the coherent accumulation gain is greater than a preset gain can be identified as first-type points, and the remaining points to be confirmed can be identified as second-type points. The preset gain can be a fixed value or it can be determined based on the signal-to-noise ratio of the points to be confirmed, such as the difference between the constant false alarm rate threshold at the point to be confirmed and the signal-to-noise ratio of the point to be confirmed.

[0213] The coherent accumulation gain of the point to be confirmed in the current frame can also be represented by the coherent accumulation gain of the point corresponding to the point to be confirmed in the next frame.

[0214] Optionally, based on the coherent accumulation gain of the points to be confirmed in the current frame, the points to be confirmed in the current frame are classified to obtain the first type of points and the second type of points in the current frame. This includes: determining the points to be confirmed whose sum of signal-to-noise ratio and coherent accumulation gain is greater than the constant false alarm rate threshold as the first type of points, and determining the points to be confirmed whose sum of signal-to-noise ratio and coherent accumulation gain is less than or equal to the constant false alarm rate threshold as the second type of points; the constant false alarm rate threshold is the detection threshold used in constant false alarm rate detection.

[0215] Optionally, the method further includes: for the point to be confirmed in the current frame, determining the distance index and Doppler index of the point to be confirmed in the next frame, and obtaining the point corresponding to the point to be confirmed in the next frame; based on the angle of the point to be confirmed in the current frame, performing coherent accumulation on the data of multiple antenna channels of the point corresponding to the point to be confirmed in the next frame, and obtaining the coherent accumulation value of the point corresponding to the point to be confirmed in the next frame; and based on the coherent accumulation value of the point corresponding to the point to be confirmed in the next frame, determining the coherent accumulation gain of the point to be confirmed.

[0216] Optionally, the method further includes: performing constant false alarm rate (CFAR) detection on the coherently accumulated data of the current frame to obtain the CFAR threshold for each point to be confirmed in the current frame; wherein, the data of the point to be confirmed in the coherently accumulated data of the current frame is the coherently accumulated value of the point to be confirmed in the current frame.

[0217] Step S707: Calculate the minimum SNR of the first type of points among the points to be confirmed in the current frame to obtain the first SNR, and calculate the maximum SNR of the second type of points among the points to be confirmed in the current frame to obtain the second SNR.

[0218] Step S708: Based on the first signal-to-noise ratio and the second signal-to-noise ratio, determine the detection threshold factor for the next frame. The next frame is then taken as the current frame, and step S701 is returned to determine the detection threshold factor for subsequent frames. This process continues until no further frames exist.

[0219] For example, the detection threshold factor for the next frame can be determined as the average of the first signal-to-noise ratio and the second signal-to-noise ratio.

[0220] Optionally, based on the first signal-to-noise ratio and the second signal-to-noise ratio, the detection threshold factor for the next frame is determined, including: determining the smaller value of the first signal-to-noise ratio and the second signal-to-noise ratio as the detection threshold factor for the next frame.

[0221] The first signal-to-noise ratio (SNR) represents the minimum SNR of a set of points that are likely to be targets, and can be used to characterize the minimum SNR of distinguishable targets; the second SNR represents the maximum SNR of a set of points that are likely to be noise, and can be used to characterize the highest level of noise.

[0222] Choosing the smaller of the first and second signal-to-noise ratios as the detection threshold factor for the next frame avoids the problem of missed detections caused by setting the detection threshold factor too high. At the same time, by comprehensively considering the smallest distinguishable target and the highest noise level in the current frame, the detection threshold factor is determined, which improves the accuracy of the detection threshold factor determination.

[0223] In this embodiment, the signal-to-noise ratio is improved by noncoherent accumulation of the multi-antenna channel RDM, thereby improving the accuracy of target detection; the noncoherent accumulation matrix RDM... nc Noise floor estimation at the distance unit granularity improves the closeness of the noise floor estimate to the real noise, thereby improving the accuracy of signal-to-noise ratio (SNR) calculation and pre-detection threshold calculation. Since the SNR is significantly improved after coherent accumulation, after the points to be confirmed in the corresponding frame are selected by the pre-detection threshold of the corresponding frame during pre-detection, target and noise classification is performed on the points to be confirmed based on the coherent accumulation gain, which improves the accuracy of classification. The detection threshold factor of the next frame is determined by the minimum SNR of the points to be confirmed belonging to the target class and the maximum SNR of the points belonging to the noise class. Combined with the noise estimate, the pre-detection threshold of the next frame is calculated, so that the pre-detection threshold can better adapt to changes in signal and noise, improving the accuracy and robustness of target detection.

[0224] Optionally, the method further includes: determining a pre-detection threshold for the current frame based on the noise floor estimate of each distance cell in the far-distance interval of the current frame and the detection threshold factor of the current frame; and detecting points to be confirmed from each distance cell in the far-distance interval of the current frame based on the pre-detection threshold of the current frame.

[0225] Optionally, the method further includes: performing noise floor estimation on the Doppler data of each distance cell in the long-range interval of the current frame to obtain the noise floor estimate value of each distance cell in the long-range interval of the current frame.

[0226] Optionally, the method further includes: performing noncoherent accumulation on the distance-Doppler matrix of the current frame to obtain the noncoherent accumulation matrix of the current frame; and estimating the noise floor of the Doppler data of each distance cell in the far-distance interval of the noncoherent accumulation matrix of the current frame to obtain the noise floor estimate of each distance cell in the far-distance interval of the current frame.

[0227] Optionally, noise floor estimation is performed on the Doppler data of each distance cell in the long-range interval of the noncoherent accumulation matrix of the current frame to obtain the noise floor estimate of each distance cell in the long-range interval of the current frame. This includes: calculating the mean of the Doppler data of each distance cell in the long-range interval of the noncoherent accumulation matrix of the current frame; repeatedly replacing the maximum value in the Doppler data of the distance cell with the mean until the difference between two adjacent mean values ​​is less than a preset value to obtain the Doppler data of the distance cell after adjustment; and determining the noise floor estimate of the distance cell based on the variance of the Doppler data of the distance cell after adjustment.

[0228] Figure 10 For this application Figure 7 The flowchart of step S706 in the illustrated embodiment is as follows: Figure 10 As shown, step S706 may specifically include the following steps:

[0229] Step S1001: Obtain the data accumulated from the previous frame's coherence.

[0230] The data after coherent accumulation in each frame is the data obtained after coherent accumulation based on the coherent accumulation algorithm provided in the embodiments of this application, and the points for coherent accumulation are points to be confirmed.

[0231] Based on the information of the point to be confirmed in the previous frame, including angle, velocity, etc., the data of multiple antenna channels of the point corresponding to the point to be confirmed in the current frame, such as RDM or RDM nc The data in the data are coherently accumulated to obtain the coherent accumulated value xˊ of the point in the current frame; the coherent accumulated value xˊ is used to replace the distance-Doppler matrix RDM or the non-coherent accumulated matrix RDM. nc By determining the original value x of each point, we can obtain the data accumulated by the coherent parameters of the current frame.

[0232] Before coherent accumulation, phase compensation is required to compensate for the phase changes caused by the target's motion relative to the radar.

[0233] Step S1002: Perform constant false alarm rate (CFAR) detection on the coherently accumulated data of the previous frame to obtain the CFAR threshold for each point to be confirmed in the previous frame.

[0234] After obtaining the coherently accumulated data for each frame, CFAR detection is performed to determine target points from the points to be confirmed in each frame. During CFAR detection, background clutter power is estimated based on data from points within a certain range of the points to be confirmed. Based on the estimated background clutter power and a preset false alarm probability, a constant false alarm rate (CFAR) threshold is determined. Target detection is then performed based on this CFAR threshold, and points to be confirmed whose data exceeds the CFAR threshold are identified as target points. If no points to be confirmed whose data exceeds the CFAR threshold exist, then no target is identified.

[0235] Step S1003: Determine the points to be confirmed where the sum of the signal-to-noise ratio and the coherent accumulation gain is greater than the constant false alarm rate threshold as the first type of points, and determine the points to be confirmed where the sum of the signal-to-noise ratio and the coherent accumulation gain is less than or equal to the constant false alarm rate threshold as the second type of points.

[0236] For each point to be confirmed in each frame, after obtaining the constant false alarm rate threshold of the point to be confirmed, the sum of the signal-to-noise ratio of the point to be confirmed and the coherent accumulation gain of the point to be confirmed is calculated, and it is determined whether the sum is greater than the constant false alarm rate threshold of the point to be confirmed; if so, the point to be confirmed is determined to be a first-class point, otherwise, the point to be confirmed is determined to be a second-class point.

[0237] The coherent accumulation gain of a point to be confirmed in a certain frame is determined based on the coherent accumulation value of the point corresponding to that point in the next frame, specifically based on the ratio of the coherent accumulation value of that point to the original value.

[0238] Because the target point itself has a high signal-to-noise ratio, its gain is high after coherent accumulation. On the other hand, the noise itself has a low signal-to-noise ratio, and its gain is low after coherent accumulation due to phase incoherence. Therefore, the target point and noise in the points to be confirmed can be distinguished by the signal-to-noise ratio and the coherent accumulation gain, thus improving the accuracy of the classification of the points to be confirmed. Since the threshold of constant false alarm rate detection will be adaptively adjusted to adapt to complex environments, the classification of the points to be confirmed can be achieved by comparing with the constant false alarm rate threshold, further improving the accuracy of the classification.

[0239] Figure 11 A flowchart illustrating the pre-detection threshold determination method provided in this application embodiment. Figure 3 ,like Figure 11 As shown, the pre-detection threshold determination method provided in this embodiment specifically includes the following steps:

[0240] Step S1101, set the detection threshold factor for the first frame: specifically, determine the initial detection threshold factor as the detection threshold factor for the first frame;

[0241] Step S1102: Perform noncoherent accumulation of the RDM of multiple antenna channels in the current frame to obtain the noncoherent accumulation matrix RDM. nc ;

[0242] Step S1103, using RDM nc The noise floor is estimated using Doppler data in each distance cell. The noise floor estimate is obtained. The pre-detection threshold of the current frame is determined as the product of the detection threshold factor of the current frame and the noise floor estimate. Points exceeding the pre-detection threshold are identified as points to be confirmed. The signal-to-noise ratio of the points to be confirmed is calculated.

[0243] Step S1104: Perform coherent accumulation on the point corresponding to the point to be confirmed in the next frame;

[0244] Step S1105: Replace RDM with coherent accumulated value. nc The values ​​at the corresponding positions are used to obtain the coherent accumulation matrix, and CFAR detection is performed based on the coherent accumulation matrix;

[0245] Step S1106: Determine the coherent accumulation gain of the point to be confirmed based on the ratio of the coherent accumulation value of the point corresponding to the point to be confirmed in the next frame to the value before the coherent accumulation of that point.

[0246] Step S1107: Determine whether the sum of the signal-to-noise ratio and coherent accumulation gain of the point to be confirmed exceeds the CFAR detection threshold, i.e., the constant false alarm rate threshold.

[0247] Step S1108: If the CFAR detection threshold is exceeded, the point to be confirmed is determined to belong to the first type of point; if the CFAR detection threshold is not exceeded, the point to be confirmed is determined to belong to the second type of point; calculate the minimum signal-to-noise ratio (SNR) of the first type of point. 1,min And the maximum signal-to-noise ratio (SNR) of the second type of points. 2,max ;

[0248] Step S1109, obtain SNR 1,min and SNR 2,max The smaller value in is used as the detection threshold factor for the next frame; return to step S1102, take the next frame as the current frame, perform CFAR detection for the next frame, and dynamically determine the detection threshold factor for subsequent frames.

[0249] Exemplary embodiments

[0250] Taking millimeter-wave radar as an example, assuming the millimeter-wave radar uses TDMA transmission mode when acquiring data, with 512 sampling points, 256 identical frequency-modulated continuous waves transmitted per frame, a range resolution of 0.61m, and a maximum detectable range of approximately 156m, and setting target 1 and target 2 to distances of 30m and 120m respectively, a velocity of 30m / s, and angles of 5.7° and 1.4° respectively,... Figure 5 The antenna array shown is the antenna array of this embodiment.

[0251] The antenna array components to which the technical solutions provided in this application are applicable are not limited to... Figure 5 The array shown can also be other arrays. Figure 5 The array shown is for illustrative purposes only.

[0252] First, perform FFT transformations on the data from multiple antenna channels acquired in the Nth frame in both the range dimension and Doppler dimension to obtain the RDM of multiple antenna channels; then, perform noncoherent accumulation on the RDM of multiple antenna channels in the same frame to obtain the noncoherent accumulation matrix RDM. nc .

[0253] Using intervals exceeding 100m as long-distance intervals, RDM is utilized in long-distance intervals. nc For each range cell in the RDM, noise floor estimation is performed to obtain the noise floor estimate. Combined with the detection threshold factor of the current frame, a pre-detection threshold for each range cell in the long-range interval of the current frame is obtained. The pre-detection threshold can be the product of the detection threshold factor and the noise floor estimate. Based on the pre-detection threshold for each range cell in the long-range interval of each frame, the RDM of each frame is... nc Pre-detection is performed to obtain multiple points to be confirmed in each frame.

[0254] For example, Figure 12 A schematic diagram of the long-distance interval pre-detection results in an exemplary embodiment provided in this application, as shown below. Figure 12 As shown, the points to be confirmed that exceed the pre-detection threshold in the long-distance range include target 2 in the aforementioned embodiment, whose coordinates are [27, 198, 104.651]. In the distance dimension, 1 represents a distance of approximately 0.6m (156m / 256), so the distance of target 2 is approximately 120m.

[0255] For each frame's pre-detection result, i.e., the point to be confirmed in the long-distance interval of each frame, save the distance index, Doppler index, distance, velocity, and angle of the point to be confirmed.

[0256] Based on the distance index and Doppler index of the point to be confirmed in the Nth frame, and the true velocity obtained by deblurring, determine the blurred velocity Vˊ of the point to be confirmed in the N+1th frame.

[0257] When Vˊ is less than 0, the target Doppler index is: round(Vˊ / V) N+ 1res When Vˊ is greater than or equal to 0, the target Doppler index is: round(Vˊ / V) + dBins; N+ 1res +1, dBins is the maximum value of the Doppler index for the next frame.

[0258] The reference point for frame N+1 is obtained by using the distance index and the target Doppler index of frame N. The maximum value is searched in the vicinity of the reference point, i.e., within the preset range. The Doppler index of the maximum value is recorded as the Doppler index of the point to be confirmed in frame N+1. The distance index remains unchanged, and the point to be confirmed in frame N+1 is obtained.

[0259] Using the distance index and Doppler index of the point to be confirmed in the Nth frame (as described above), data from multiple antenna channels of that point are extracted. Phase compensation is then performed using the angle and velocity of the point to be confirmed in the Nth frame, where the Doppler phase is: for Figure 5The antenna array layout shown has the following phase compensation coefficients for the 12 virtual antennas: c = [1, 1, 1, 1, e] -jΔΦ ,e -jΔΦ ,e -jΔΦ ,e -jΔΦ ,e -j2ΔΦ ,e -j2ΔΦ ,e -j2ΔΦ ,e -j2ΔΦ ]. Among them, V N T represents the actual velocity of the point to be confirmed in frame N; C The transmission period of the linear frequency modulated pulse Chirp is given. Coherent accumulation is performed on the phase-compensated data. Figure 12 The weighting coefficients for the 12 virtual antennas in the antenna array shown are:

[0260] Based on the magnitude of the gain before and after coherent accumulation, the coherent accumulation gain of the point to be confirmed in the Nth frame corresponding to that point is calculated. Based on the coherent accumulation gain, the detection threshold factor of the N+1th frame is determined.

[0261] Specifically, by determining whether the sum of coherent accumulation gain and signal-to-noise ratio exceeds the CFAR detection threshold, the points to be confirmed in the Nth frame are classified into Class I points and Class II points. The smaller of the minimum signal-to-noise ratio of Class I points and the maximum signal-to-noise ratio of Class II points is determined as the detection threshold factor for the N+1th frame.

[0262] Figure 13a This is a schematic diagram of the CFAR detection results before coherent accumulation, as an exemplary embodiment of this application. Figure 13b This is a schematic diagram of the CFAR detection results after coherent accumulation, as an exemplary embodiment of this application, as shown below. Figure 13a and Figure 13b As shown, the horizontal axis X represents distance in meters (m), and the vertical axis Y represents the relative intensity of the echo in dB. Figure 13a The distance dimension signal in the equation is the original signal of each point without coherent accumulation. Figure 13b The distance dimension signal in the equation represents the signal at each point after coherent accumulation, i.e., the coherent accumulation value. The CFAR detection threshold is the aforementioned constant false alarm rate threshold. Before coherent accumulation, target 2 with coordinates (120.156, 100.81) in the long-range interval could not be detected. After coherent accumulation, due to the enhanced signal-to-noise ratio of the target, target 2 with coordinates (123.206, 114.66) near 120m exceeds the CFAR detection threshold and can be detected.

[0263] After coherent accumulation, based on the coherent accumulation gain of the points to be confirmed, the points to be confirmed in the first frame are classified, and the minimum SNR of the first type of points (the type of points belonging to the target) and the maximum SNR of the second type of points (the type of points belonging to the noise) are calculated. The smaller value is taken as the detection threshold factor for the second frame to determine the pre-detection threshold of the second frame, thereby realizing the pre-detection and target detection of the second frame, and so on.

[0264] Corresponding to the coherent accumulation method provided in the foregoing embodiments, this application also provides a corresponding apparatus. This application provides a coherent accumulation apparatus, comprising: a corresponding point acquisition module, used to acquire, for a point to be confirmed in the current frame, the point corresponding to the point to be confirmed in the next frame; and a coherent accumulation module, used to coherently accumulate data from multiple antenna channels of the point to be confirmed in the current frame based on the angle of the point to be confirmed in the current frame, obtaining the coherent accumulation value of the point corresponding to the point to be confirmed in the next frame.

[0265] In one possible implementation, the corresponding point acquisition module includes: a Doppler index determination unit, used to determine the Doppler index of the point to be confirmed in the next frame; and a corresponding point determination unit, used to obtain the point corresponding to the point to be confirmed in the next frame by using the Doppler index of the point to be confirmed in the next frame as the Doppler index and the distance index of the point to be confirmed as the distance index.

[0266] In one possible implementation, the Doppler index determination unit is specifically used to: determine the Doppler index of the point to be confirmed in the next frame based on the velocity obtained after the point to be confirmed is deblurred in the current frame.

[0267] In one possible implementation, the Doppler index determination unit includes: a blur velocity determination subunit, used to determine the blur velocity of the point to be confirmed in the next frame based on the velocity obtained after deblurring the point in the current frame and the detection velocity limit of the next frame; and a Doppler index determination subunit, used to determine the Doppler index of the point to be confirmed in the next frame based on the blur velocity of the point to be confirmed in the next frame.

[0268] In one possible implementation, the fuzzy velocity determination subunit is specifically used to: take the integer part and the fractional part of the quotient of the velocity obtained by deblurring the point to be confirmed in the current frame and dividing it by the detection upper limit velocity of the next frame; if the integer part is odd, then calculate the fuzzy velocity of the point to be confirmed in the next frame based on the fractional part and the first relation; if the integer part is even, then calculate the fuzzy velocity of the point to be confirmed in the next frame based on the fractional part and the second relation.

[0269] In one possible implementation, the Doppler index determination subunit includes: a target index determination service for determining a target Doppler index based on the blur velocity of the point to be confirmed in the next frame; a reference point determination service for obtaining a reference point using the target Doppler index as the Doppler index and the distance index of the point to be confirmed as the distance index; and a Doppler index search service for searching for the Doppler index of the point to be confirmed in the next frame from a preset range of reference points in the distance-Doppler map of the next frame.

[0270] In one possible implementation, the Doppler index search service is specifically used to: determine the Doppler index corresponding to the maximum value within a preset range of distance-Doppler map reference points in the next frame, which is the Doppler index of the point to be confirmed in the next frame.

[0271] In one possible implementation, the target index determination service is specifically used to: round down the quotient of the blurred velocity of the point to be confirmed in the next frame divided by the velocity resolution of the next frame to obtain the initial Doppler index; if the blurred velocity of the point to be confirmed in the next frame is less than 0, then the target Doppler index is determined to be the sum of the maximum value of the initial Doppler index and the Doppler index of the next frame; if the blurred velocity of the point to be confirmed in the next frame is greater than or equal to 0, then the target Doppler index is determined to be the initial Doppler index plus 1.

[0272] In one possible implementation, the coherent accumulation module is specifically used for: performing phase compensation on the data of multiple antenna channels of the point to be confirmed in the next frame to obtain phase-compensated data; calculating the weighting coefficients of multiple antenna channels based on the angle of the point to be confirmed in the current frame; and performing coherent accumulation on the phase-compensated data based on the weighting coefficients of multiple antenna channels to obtain the coherent accumulation value of the point to be confirmed in the next frame.

[0273] In one possible implementation, the point to be confirmed is a point in a long-distance interval whose value exceeds a pre-detection threshold; the distance index of the long-distance interval is greater than a preset distance index.

[0274] In one possible implementation, the coherent accumulation device further includes: a pre-detection threshold determination module, used to determine the pre-detection threshold of each distance cell in the long-distance interval of the current frame based on the noise floor estimate of each distance cell in the long-distance interval of the current frame and the detection threshold factor of the current frame; and a pre-detection module, used to determine points in the long-distance interval of the current frame that exceed the pre-detection threshold of their respective distance cells as points to be confirmed.

[0275] In one possible implementation, the coherent accumulation device further includes a first noise floor estimation module, used to: perform noise floor estimation on the Doppler data of each distance cell in the long-range interval of the current frame, and obtain the noise floor estimation value of each distance cell in the long-range interval of the current frame.

[0276] In one possible implementation, the coherent accumulation device further includes: a noncoherent accumulation module for performing noncoherent accumulation of the range-Doppler matrix of the current frame to obtain the noncoherent accumulation matrix of the current frame; and a second noise floor estimation module for performing noise floor estimation on the Doppler data of each distance cell in the long-range interval of the noncoherent accumulation matrix of the current frame to obtain the noise floor estimate value of each distance cell in the long-range interval of the current frame.

[0277] In one possible implementation, the second noise floor estimation module is specifically used to: calculate the mean of the Doppler data of each distance cell in the long-distance interval of the non-coherent accumulation matrix of the current frame; repeatedly replace the maximum value in the Doppler data of the distance cell with the mean until the difference between two adjacent mean values ​​is less than a preset value, thereby obtaining the adjusted Doppler data of the distance cell; and determine the noise floor estimate of the distance cell based on the variance of the adjusted Doppler data of the distance cell.

[0278] In one possible implementation, the coherent accumulation device further includes a threshold factor determination module, used to: determine the detection threshold factor of the current frame based on the signal-to-noise ratio of the first type of points and the second type of points in the previous frame to be confirmed.

[0279] In one possible implementation, the threshold factor determination module includes: a signal-to-noise ratio (SNR) statistics unit, used to calculate the minimum SNR of the first type of points in the previous frame to obtain a first SNR, and to calculate the maximum SNR of the second type of points in the previous frame to obtain a second SNR; and a threshold factor determination unit, used to determine the detection threshold factor of the current frame based on the first SNR and the second SNR.

[0280] In one possible implementation, the threshold factor determination unit is specifically used to: determine the smaller of the first signal-to-noise ratio and the second signal-to-noise ratio as the detection threshold factor for the current frame.

[0281] In one possible implementation, the coherent accumulation device further includes a classification module, used to: classify the points to be confirmed in the previous frame based on the coherent accumulation gain of the points to be confirmed in the previous frame, to obtain the first type of points and the second type of points in the previous frame; the coherent accumulation gain of the points to be confirmed in the previous frame is determined based on the coherent accumulation value of the points to be confirmed in the current frame.

[0282] In one possible implementation, the classification module is specifically used to: determine the points to be confirmed where the sum of the signal-to-noise ratio and the coherent cumulative gain is greater than the constant false alarm rate threshold as first-class points, and to determine the points to be confirmed where the sum of the signal-to-noise ratio and the coherent cumulative gain is less than or equal to the constant false alarm rate threshold as second-class points; the constant false alarm rate threshold is the detection threshold used in constant false alarm rate detection.

[0283] In one possible implementation, the device further includes a constant false alarm rate (CFAR) detection module, used to: perform CFAR detection on the coherently accumulated data of the previous frame to obtain the CFAR threshold for each point to be confirmed in the previous frame; wherein, the data of the point to be confirmed in the coherently accumulated data of the previous frame is the coherently accumulated value of the point to be confirmed in the previous frame.

[0284] In one possible implementation, the coherent accumulation device further includes an initial factor acquisition module, used to: acquire an initial detection threshold factor, and use the initial detection threshold factor as the detection threshold factor of the first frame.

[0285] In one possible implementation, the coherent accumulation device further includes an initial factor determination module for determining an initial detection threshold factor based on radar parameters.

[0286] In one possible implementation, the initial factor determination module is specifically used to: determine the minimum resolvable signal-to-noise ratio of the radar based on the radar parameters and radar equations; and determine the initial detection threshold factor based on the minimum resolvable signal-to-noise ratio.

[0287] In one possible implementation, the coherent accumulation device further includes an initial factor adjustment module, used to: reduce the initial detection threshold factor if the number of points to be confirmed obtained based on the pre-detection threshold screening of the first frame is less than a preset number, so as to redetermine the pre-detection threshold of the first frame based on the reduced initial detection threshold factor.

[0288] The coherent accumulation apparatus provided in this application can be used to execute the technical solution of the coherent accumulation method provided in any of the above embodiments of this application. Its implementation principle and technical effect are similar, and will not be described again here.

[0289] Corresponding to the target detection method provided in the foregoing embodiments, this application also provides a corresponding apparatus. This application provides a target detection apparatus for: performing target detection based on data coherently accumulated from each frame; wherein, the data of the point corresponding to the point to be confirmed in the previous frame in the data coherently accumulated from each frame is the coherent accumulation value of that point; the coherent accumulation value is obtained based on the coherent accumulation method provided in this application.

[0290] In one possible implementation, target detection is performed based on the data accumulated by coherence across frames, including: target detection based on constant false alarm rate (CFAR) detection, which involves performing target detection on the data accumulated by coherence across frames.

[0291] The target detection device provided in this application embodiment can be used to execute the target detection method provided in any of the above embodiments of this application. Its implementation principle and technical effect are similar, and will not be repeated here.

[0292] This application also provides an electronic device, including a memory and a processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, causing the processor to execute the technical solutions of the pre-detection threshold determination method, threshold factor determination method, or target detection method provided in any of the above embodiments of this application.

[0293] It should be noted that the beneficial effects of the electronic device provided in this application embodiment are the same as those of the method embodiment described above, and will not be elaborated further in this embodiment.

[0294] This application also provides a radar system, including a control unit, for executing the technical solutions of the pre-detection threshold determination method, threshold factor determination method, or target detection method provided in any of the above embodiments of this application.

[0295] Furthermore, the radar system may also include a transmitter, an antenna, a receiver, and a signal processing unit. The antenna is connected to both the transmitter and the receiver. The transmitter is responsible for transmitting radar signals, and the antenna is responsible for radiating the radar signals generated by the transmitter into space and receiving the echo signals reflected back from the target, transmitting the echo signals to the receiver for processing. The receiver mainly amplifies and filters the echo signals to improve the signal-to-noise ratio of the echo signals.

[0296] The signal processing unit is connected to both the receiver and the control unit. It is responsible for further processing and analyzing the signal output from the receiver, such as distance measurement, velocity measurement, and direction measurement. Through a series of processing steps, useful information for target detection can be extracted, and radar data or images can be generated. The control unit, based on the radar data output from the signal processing unit, further processes the radar data by executing the pre-detection threshold determination method, threshold factor determination method, or target detection method provided in any of the above embodiments.

[0297] For example, the radar system is a millimeter-wave radar system, a high-frequency ground-wave radar system, etc.

[0298] This application also provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, are used to implement the methods of any of the foregoing embodiments.

[0299] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the method of any of the foregoing embodiments.

[0300] The various embodiments or implementation methods described in this specification are presented in a progressive manner. Each embodiment focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referred to each other.

[0301] It should be noted that the embodiments referred to in the specification, such as "some embodiments," "exemplary embodiments," "one embodiment," or simply "embodiment," may include specific features, structures, or characteristics, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, such phrases do not necessarily refer to the same embodiment. Moreover, when a specific feature, structure, or characteristic is described in connection with an embodiment, implementing such a feature, structure, or characteristic in conjunction with other embodiments, whether explicitly described or not, is within the knowledge scope of those skilled in the art.

[0302] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method of coherent accumulation, characterized by, The method comprises the following steps: For a to-be-confirmed point of a current frame, a corresponding point of the to-be-confirmed point in a next frame is obtained; Based on an angle of the to-be-confirmed point in the current frame, data of multiple antenna channels of the corresponding point of the to-be-confirmed point in the next frame is co-phased accumulated to obtain a co-phased accumulation value of the corresponding point of the to-be-confirmed point in the next frame.

2. The method of claim 1, wherein, The step of obtaining the corresponding point of the to-be-confirmed point in the next frame comprises the following steps: A Doppler index of the to-be-confirmed point in the next frame is determined; The Doppler index of the to-be-confirmed point in the next frame is taken as a Doppler index, and a distance index of the to-be-confirmed point is taken as a distance index to obtain the corresponding point of the to-be-confirmed point in the next frame.

3. The method of claim 2, wherein, The step of determining the Doppler index of the to-be-confirmed point in the next frame comprises the following steps: Based on a speed of the to-be-confirmed point obtained after deblurring of the current frame, the Doppler index of the to-be-confirmed point in the next frame is determined.

4. The method of claim 3, wherein, The step of determining the Doppler index of the to-be-confirmed point in the next frame based on the speed of the to-be-confirmed point obtained after deblurring of the current frame comprises the following steps: Based on the speed of the to-be-confirmed point obtained after deblurring of the current frame and a detection speed limit value of the next frame, a blurring speed of the to-be-confirmed point in the next frame is determined; Based on the blurring speed of the to-be-confirmed point in the next frame, the Doppler index of the to-be-confirmed point in the next frame is determined.

5. The method of claim 4, wherein, The step of determining the blurring speed of the to-be-confirmed point in the next frame based on the speed of the to-be-confirmed point obtained after deblurring of the current frame and the detection speed limit value of the next frame comprises the following steps: An integer part and a decimal part of a quotient of the speed of the to-be-confirmed point obtained after deblurring of the current frame divided by the detection upper limit speed of the next frame are taken; If the integer part is odd, the blurring speed of the to-be-confirmed point in the next frame is calculated based on the decimal part and a first relationship.

6. The method of claim 5, wherein, The method further comprises the following steps: If the integer part is even, the blurring speed of the to-be-confirmed point in the next frame is calculated based on the decimal part and a second relationship.

7. The method of claim 4, wherein, The step of determining the Doppler index of the to-be-confirmed point in the next frame based on the blurring speed of the to-be-confirmed point in the next frame comprises the following steps: Based on the blurring speed of the to-be-confirmed point in the next frame, a target Doppler index is determined; The target Doppler index is taken as a Doppler index, and a distance index of the to-be-confirmed point is taken as a distance index to obtain a reference point; From a preset range of the reference point in a range-Doppler graph of the next frame, the Doppler index of the to-be-confirmed point in the next frame is searched.

8. The method of claim 7, wherein, The step of searching the Doppler index of the to-be-confirmed point in the next frame from the preset range of the reference point in the range-Doppler graph of the next frame comprises the following steps: A Doppler index corresponding to a maximum value in the preset range of the reference point in the range-Doppler graph of the next frame is determined as the Doppler index of the to-be-confirmed point in the next frame.

9. The method of claim 7, wherein, The step of determining the target Doppler index based on the blurring speed of the to-be-confirmed point in the next frame comprises the following steps: An initial Doppler index is obtained by rounding a quotient of the blurring speed of the to-be-confirmed point in the next frame divided by a speed resolution of the next frame; If the to-be-confirmed point in the next frame has a blurring speed less than 0, the target Doppler index is determined as a sum of the initial Doppler index and a maximum value of the next frame Doppler index.

10. The method of claim 9, wherein, The method further comprises: If the to-be-confirmed point in the next frame has a blurring speed greater than or equal to 0, the target Doppler index is determined as the initial Doppler index plus 1.

11. The method of claim 1, wherein, The coherent accumulation of the data of the multiple antenna channels of the point corresponding to the to-be-confirmed point in the next frame based on the angle of the to-be-confirmed point in the current frame comprises: Phase compensation is performed on the data of the multiple antenna channels of the point corresponding to the to-be-confirmed point in the next frame to obtain phase-compensated data; Based on the angle of the to-be-confirmed point in the current frame, weighting coefficients of the multiple antenna channels are calculated; Based on the weighting coefficients of the multiple antenna channels, coherent accumulation is performed on the phase-compensated data to obtain the coherent accumulation value of the point corresponding to the to-be-confirmed point in the next frame.

12. The method according to any one of claims 1 to 11, characterized in that, The to-be-confirmed point is a point in a long-distance range that has a value exceeding a pre-detection threshold.

13. The method of claim 12, wherein, The method further comprises: Based on the noise floor estimation value of each distance unit in the long-distance range of the current frame and a detection threshold factor of the current frame, a pre-detection threshold of each distance unit in the long-distance range of the current frame is determined. Points in the long-distance range of the current frame that exceed the pre-detection threshold of the distance unit are determined as to-be-confirmed points.

14. The method of claim 13, wherein, The method further comprises: Noise floor estimation is performed on Doppler data of each distance unit in the long-distance range of the current frame to obtain noise floor estimation values of each distance unit in the long-distance range of the current frame.

15. The method of claim 13, wherein, The method further comprises: Non-coherent accumulation is performed on a range-Doppler matrix of the current frame to obtain a non-coherent accumulation matrix of the current frame; Noise floor estimation is performed on Doppler data of each distance unit in the long-distance range of the non-coherent accumulation matrix of the current frame to obtain noise floor estimation values of each distance unit in the long-distance range of the current frame.

16. The method of claim 15, wherein, The noise floor estimation of the Doppler data of each distance unit in the long-distance range of the non-coherent accumulation matrix of the current frame comprises: For each distance unit in the long-distance range of the non-coherent accumulation matrix of the current frame, a mean value of the Doppler data of the distance unit is calculated; The maximum value of the Doppler data of the distance unit is replaced by the mean value for multiple times until a difference between two adjacent mean values is less than a preset value, and adjusted Doppler data of the distance unit is obtained; Based on a variance of the adjusted Doppler data of the distance unit, a noise floor estimation value of the distance unit is determined.

17. The method of claim 13, wherein, The method further comprises: Based on signal-to-noise ratios of first-type points and second-type points in to-be-confirmed points of a previous frame, a detection threshold factor of the current frame is determined.

18. The method of claim 17, wherein, The determination of the detection threshold factor of the current frame based on the signal-to-noise ratios of the first-type points and the second-type points in the to-be-confirmed points of the previous frame comprises: A minimum value in the signal-to-noise ratios of the first-type points in the to-be-confirmed points of the previous frame is counted to obtain a first signal-to-noise ratio; A maximum value in the signal-to-noise ratios of the second-type points in the to-be-confirmed points of the previous frame is counted to obtain a second signal-to-noise ratio; Based on the first signal-to-noise ratio and the second signal-to-noise ratio, the detection threshold factor of the current frame is determined.

19. The method of claim 18, wherein, The method further comprises: The smaller value of the first signal-to-noise ratio and the second signal-to-noise ratio is determined as the detection threshold factor of the current frame.

20. The method of claim 17, wherein, The method further comprises: The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame.

21. The method of claim 20, wherein, The method further comprises: The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame.

22. The method of claim 21, wherein, The method further comprises: The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame.

23. The method according to any one of claims 17-22, characterized by, The method further comprises: The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame.

24. The method of claim 23, wherein, The method further comprises: The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame.

25. The method of claim 24, wherein, The method further comprises: The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame. The method further comprises:

26. The method of claim 23, wherein, The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame. The method further comprises:

27. A target detection method characterized by, The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame. The method further comprises:

28. The method of claim 27, wherein, The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame. The method further comprises:

29. A phase coherent accumulation device, characterized by The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame. The method further comprises: The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame.

30. A target detection apparatus characterized by comprising: The method further comprises: The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame. The method further comprises: The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame. The method further comprises: The points of the previous frame are classified based on the coherent accumulation gain of the points of the previous frame to be confirmed to obtain the first type of points and the second type of points of the previous frame; the coherent accumulation gain of the points of the previous frame to be confirmed is determined based on the coherent accumulation value of the points of the previous frame to be confirmed corresponding to the points of the current frame. Target detection is performed based on the data of each frame after phase correlation accumulation; wherein the data of the point corresponding to the last frame to be confirmed point in the data of each frame after phase correlation accumulation is the phase correlation accumulation value of the point; the phase correlation accumulation value is obtained based on the method of any one of claims 1-26.

31. An electronic device, comprising: comprising a memory and a processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, so that the processor executes the method of any one of claims 1-28.

32. A radar system, characterized by comprising a control unit; the control unit is configured to execute the method of any one of claims 1-28.

33. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are executed by the processor to implement the method of any one of claims 1-28.

34. A computer program product, characterised in that, comprising a computer program, which, when executed by a processor, implements the method of any one of claims 1-28.

Images