Multi-target detection method and system based on millimeter wave radar

Abstract

The invention discloses a multi-target detection method and system based on a millimeter wave radar, and the method comprises the steps: obtaining original echo data collected by the millimeter wave radar, and the original echo data comprise target data points of a plurality of targets; selecting a preset number of data points from the target data points of the plurality of targets, and taking the selected data points as effective data points; and detecting the plurality of targets based on the valid data points. The non-strict threshold value screening constant false alarm rate algorithm is combined with the dual index determination method and the spatial distance judgment algorithm, so that the influence of the multipath effect on result data can be reduced while multi-target detection is realized, the accuracy of target point detection is improved, the output data volume of frequency domain data processing is greatly reduced, and the accuracy of target point detection is improved. The operation amount of logic processing is reduced, and implementation of the algorithm on edge end mobile equipment is facilitated.

Description

Technical Field

[0001] The present invention relates to the technical field of radar signal processing, and in particular to a multi-target detection method and system based on millimeter wave radar. Background Art

[0002] With the rapid growth of the autonomous driving and intelligent transportation industries, there is a strong demand for high-precision and real-time environmental perception capabilities, especially for terminal devices such as driverless cars and drones. Millimeter-wave radar has been widely used in these fields due to its excellent stability and reliable detection performance. However, millimeter-wave radar algorithms face problems in data processing, such as high computational complexity, strict real-time requirements, and traditional algorithms that are not friendly to hardware configuration.

[0003] For example, the patent publication number CN113253228B discloses a multi-size target detection method based on a radar condensation algorithm, including determining a radar condensation threshold; reserving spatial variables; if there is a target, forming a detection point cluster with points exceeding the detection threshold in each frame at the angle where the target is located; respectively counting the number of target points that are not output under the distance unit where the detection point cluster is located, the average angle of the target points, and the maximum amplitude of the target points, and numbering them; analyzing the output variables and outputting the condensed target points; if a new frame of detection data is reported, returning to detection until all frames are scanned.

[0004] The above patents all have the problems raised by this background technology: the traditional radar algorithm is not fully adapted to the hardware characteristics of FPGA, and because the traditional radar algorithm has high computational complexity and requires a large amount of computing resources, it is not conducive to implementation on edge mobile devices; to solve the above problems, the present invention proposes a multi-target detection method and system based on millimeter-wave radar. Summary of the invention

[0005] In view of the shortcomings of the prior art, the main purpose of the present invention is to provide a multi-target detection method and system based on millimeter wave radar, which can effectively solve the problems in the background technology. The specific technical solutions of the present invention are as follows:

[0006] A multi-target detection method based on millimeter wave radar, comprising:

[0007] Acquire original echo data collected by the millimeter wave radar, wherein the original echo data includes target data points of multiple targets;

[0008] Selecting a preset number of data points from the target data points of the plurality of targets, and using the selected data points as valid data points;

[0009] The multiple targets are detected based on the valid data points.

[0010] Specifically, selecting a preset number of data points from the target data points of the multiple targets and using the selected data points as valid data points includes:

[0011] Selecting a preset number of data points from the target data points of the multiple targets using a constant false alarm rate algorithm with non-strict threshold screening to obtain a first data point;

[0012] Using a dual index determination method to screen the first data point to obtain a second data point;

[0013] The second data points are screened using a spatial distance judgment algorithm to obtain valid data points.

[0014] Specifically, the constant false alarm rate algorithm using non-strict threshold screening selects a preset number of data points from the target data points of the multiple targets to obtain the first data point, including:

[0015] Performing FIR calculation on the target data points of the multiple targets to obtain environmental interference data around each target data point;

[0016] Removing the environmental interference data from the target data point data to obtain interference-free data;

[0017] The interference-removed data are sorted, and a preset number of data point indexes are selected from large to small to obtain a first data point.

[0018] Specifically, removing the environmental interference data from the target data point data to obtain interference-free data includes:

[0019] Performing a fast Fourier transform of a preset amount of data on the target data point data, and performing modulus calculation to obtain a first modulus calculation result;

[0020] Modulo the environmental interference data to obtain a second modulo result;

[0021] The first modulo result is subtracted from the second modulo result to obtain interference-free data.

[0022] Specifically, the using the dual index determination method to screen the first data point to obtain the second data point includes:

[0023] According to the multi-dimensional information points of the first data points, data points in the first data points that exist in both the distance dimension information points and the speed dimension information points are screened out as second data points.

[0024] Specifically, the using of a spatial distance judgment algorithm to screen the second data points to obtain valid data points includes:

[0025] By comparing the sizes of the distance dimension indexes and the speed dimension indexes of the second data points, a plurality of target points pointing to the same physical target to be detected are obtained, wherein the plurality of target points pointing to the same physical target to be detected are points whose distance dimension index and speed dimension index are close to each other;

[0026] Among the multiple target points, a point with a larger value in the fast Fourier result data in the angle dimension is selected as a valid data point.

[0027] Specifically, the detecting the multiple targets based on the valid data points includes:

[0028] According to the valid data points, the information of each target is calculated from multiple dimensions to obtain the dynamic information of multiple targets, thereby realizing multi-target detection based on millimeter-wave radar data.

[0029] A system for implementing the multi-target detection method based on millimeter wave radar, comprising:

[0030] Millimeter wave radar, CPU, acceleration core, DDR, FPGA and display terminal;

[0031] The millimeter wave radar is used to collect echo data;

[0032] The central processing unit is used to implement the small point calculation and logical judgment required by the multi-target detection method;

[0033] The acceleration core is used to implement the digital signal processing operations required in the multi-target detection method;

[0034] The DDR utilizes double data transmission to increase the data transmission rate;

[0035] The FPGA is used to integrate the central processing unit, the acceleration core and the DDR;

[0036] The display terminal is used to display the calculation results of the multi-target detection method.

[0037] Specifically, the acceleration core is used to implement the digital signal processing operations required in the multi-target detection method, including:

[0038] The acceleration core realizes the digital signal processing operations required in the multi-target detection method through one-dimensional sequence FFT operations, one-dimensional sequence FIR operations, one-dimensional sequence point multiplication operations, one-dimensional sequence modulus operations and one-dimensional sequence addition and subtraction operations.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] The present invention uses a constant false alarm rate algorithm with non-strict threshold screening, selects a certain number of valid data points as algorithm output according to the millimeter wave radar acquisition configuration, so as to greatly reduce the output data volume of frequency domain data processing and reduce the calculation amount of subsequent logic processing. In addition, the certain number of output data volumes can reduce the complexity of algorithm task deployment, which is beneficial to the implementation of the algorithm on edge mobile devices. A dual index determination method of a constant false alarm rate algorithm with non-strict threshold screening in different dimensions and a combination of a spatial distance algorithm that only compares the dimensional indexes can reduce the influence of multipath effects on result data. FPGA hardware equipped with both a DSP acceleration core and a CPU is selected on the system. By performing segmentation during task deployment, batch digital signal calculations are calculated using the DSP acceleration core, while other calculations such as logic judgment are calculated by the CPU. This calculation method has higher scalability and flexibility, and supports improvement of the algorithm in subsequent practice without changing or replacing the hardware platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A flowchart of a multi-target detection method based on millimeter wave radar of the present invention;

[0042] Figure 2 This is a schematic diagram of the algorithm task flow deployment of Example 1 of the present invention;

[0043] Figure 3 Schematic diagram of the constant false alarm rate algorithm in Example 1 of the present invention;

[0044] Figure 4 This is a schematic diagram of a constant false alarm rate algorithm for non-strict threshold screening in Example 1 of the present invention;

[0045] Figure 5 Schematic diagram of a dual index determination method in Embodiment 1 of the present invention;

[0046] Figure 6 Schematic diagram of a spatial distance determination algorithm in Embodiment 1 of the present invention;

[0047] Figure 7 This is a diagram of the hardware platform architecture in Example 2 of the present invention. DETAILED DESCRIPTION

[0048] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0049] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0050] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0051] Example 1

[0052] This embodiment provides a multi-target detection method based on millimeter wave radar, and the specific scheme is as follows: Figure 1 As shown, the multi-target detection method based on millimeter wave radar includes:

[0053] S101, acquiring original echo data collected by a millimeter wave radar, wherein the original echo data includes target data points of multiple targets;

[0054] S102, selecting a preset number of data points from the target data points of the multiple targets, and using the selected data points as valid data points;

[0055] S103: Detect the multiple targets based on the valid data points.

[0056] In this embodiment, Figure 2 After performing three-dimensional fast Fourier transform on the echo data collected by the millimeter-wave radar, a hardware-friendly non-strict threshold screening constant false alarm rate algorithm is used to realize multi-target data detection, a dual index determination method is used to further screen data points, and a spatial distance judgment algorithm is used to obtain target point data information. Since the strict threshold screening constant false alarm rate algorithm will screen out different amounts of target data information due to different radar collected data, the present invention uses a hardware-friendly non-strict threshold screening constant false alarm rate algorithm to obtain a certain amount of multi-target data information, and then further screens the target data information through a spatial distance judgment algorithm.

[0057] Specifically, the time domain signal is converted into a frequency domain signal by performing a multidimensional fast Fourier transform on the original echo data. The multidimensional fast Fourier transform includes a distance fast Fourier transform, a velocity fast Fourier transform and an angle fast Fourier transform. Performing a fast Fourier transform can improve the signal processing speed and separate information of different dimensions; using a constant false alarm rate algorithm with non-strict threshold screening to process the frequency domain information data, the detection threshold can be adaptively set to keep the false alarm rate constant, while maximizing the detection probability, being able to adapt to different background noise environments, improving the detection sensitivity and reducing the false alarm rate; using a dual index determination method to determine the first effective signal The interest points are screened using distance dimension information and speed dimension information. Combining information from multiple dimensions can improve detection accuracy, further reduce false alarms, and improve the reliability of target detection. The spatial distance judgment algorithm is used to obtain target point data information, separate multiple targets, and further screen the remaining data point information to improve the accuracy of target recognition. According to the target point data information, the dynamic information of multiple physical targets to be detected is calculated. Based on the target point information obtained in the previous steps, the dynamic parameters such as the position, speed, and acceleration of the target are calculated to achieve multi-target tracking, thereby finally realizing the detection of multiple targets and the acquisition of dynamic information.

[0058] The present invention uses a constant false alarm rate algorithm with non-strict threshold screening, selects a certain number of valid data points as algorithm output according to the millimeter wave radar acquisition configuration, so as to greatly reduce the output data volume of frequency domain data processing and reduce the calculation amount of subsequent logic processing. In addition, the certain number of output data volumes can reduce the complexity of algorithm task deployment, which is beneficial to the implementation of the algorithm on edge mobile devices. A dual index determination method of a constant false alarm rate algorithm with non-strict threshold screening in different dimensions and a combination of a spatial distance algorithm that only compares the dimensional indexes can reduce the influence of multipath effects on result data. FPGA hardware equipped with both a DSP acceleration core and a CPU is selected on the system. By performing segmentation during task deployment, batch digital signal calculations are calculated using the DSP acceleration core, while other calculations such as logic judgment are calculated by the CPU. This calculation method has higher scalability and flexibility, and supports improvement of the algorithm in subsequent practice without changing or replacing the hardware platform.

[0059] Furthermore, the selecting a preset number of data points from the target data points of the plurality of targets and using the selected data points as valid data points includes:

[0060] S201, selecting a preset number of data points from the target data points of the multiple targets using a constant false alarm rate algorithm with non-strict threshold screening to obtain a first data point;

[0061] S202, screening the first data point using a dual index determination method to obtain a second data point;

[0062] S203: Screen the second data points using a spatial distance judgment algorithm to obtain valid data points.

[0063] In this embodiment, the frequency domain information data is processed using a constant false alarm rate algorithm with non-strict threshold screening, and the detection threshold can be adaptively set to keep the false alarm rate constant while maximizing the detection probability, which can adapt to different background noise environments, improve detection sensitivity, and reduce false alarm rate; the first data point is screened using a dual index determination method, and information of at least two dimensions is used for screening. Combining information of multiple dimensions can improve detection accuracy, further reduce false alarms, and improve the reliability of target detection; target point data information is obtained using a spatial distance judgment algorithm to separate multiple targets, and the remaining data point information is further screened to improve the accuracy of target recognition; dynamic information of multiple physical targets to be detected is calculated based on the target point data information, and dynamic parameters such as position, velocity, and acceleration of the target are calculated based on the target point information obtained in the previous steps to realize information calculation of multiple targets, thereby finally realizing detection of multiple targets and acquisition of dynamic information.

[0064] Further, the constant false alarm rate algorithm using non-strict threshold screening selects a preset number of data points from the target data points of the multiple targets to obtain the first data point, including:

[0065] S301, performing FIR operation on the target data points of the multiple targets to obtain environmental interference data around each target data point;

[0066] S302, removing the environmental interference data from the target data point data to obtain interference-free data;

[0067] S303 , sorting the interference-removed data, and selecting a preset number of data point indexes from large to small to obtain a first data point.

[0068] In this embodiment, the background noise of the signal is estimated through FIR operation, and the data within a certain range around each data point is analyzed to obtain the environmental interference data around each data point. The frequency domain information data is subtracted from the environmental interference data to remove the estimated environmental interference, thereby extracting a purer signal and eliminating the influence of noise on the target signal, which helps to improve the contrast between the signal and the noise and enhance the detectability of the effective signal; the processed interference-removed data is sorted from large to small to facilitate the subsequent selection of effective information points, and the first N (fixed number) largest values ​​are selected from the sorted interference-removed data, where N can be set according to the calculation needs, so as to obtain the first effective information point. By selecting the first N data points, it can be ensured that the selected signal is the most significant, thereby improving the reliability of target detection.

[0069] Specifically, the constant false alarm rate algorithm with non-strict threshold screening uses a fixed number of FIR operations with pre-input FIR core data to calculate the threshold information in the traditional constant false alarm rate algorithm, and obtains comparison information by subtracting the input constant false alarm rate source data from the calculated threshold information, and then obtains the output result of the constant false alarm rate algorithm with non-strict threshold screening by sorting and screening a predetermined number of data points of the subtraction operation results.

[0070] like Figure 3 , Schematic diagram of the Constant False Alarm Rate (CFAR) algorithm, a schematic diagram of the principle of the most common CFAR algorithm in radar data processing, through which the algorithm determines whether a data point is a valid information point by comparing the value of the point with the data threshold obtained by calculating the value of the surrounding reference units. Obviously, when the millimeter-wave radar works in different environments or faces different detection targets, the algorithm will generate different amounts of data results; and under normal circumstances, the amount of result data generated by the algorithm is very large, close to the amount of data collected by the radar. The present invention improves the traditional CFAR algorithm based on this feature.

[0071] like Figure 4 , the schematic diagram of the constant false alarm rate algorithm of the non-strict threshold screening of the present invention, the algorithm has made improvements on the traditional CFAR algorithm, and fully utilized the advantages of the hardware platform in the algorithm deployment stage to implement a hardware-friendly algorithm. The algorithm uses the acquisition configuration of the millimeter wave radar, mainly the size of Samples and Chirps, to obtain the determined FIR operation core data, and calls the DSP acceleration core to perform FIR operation to obtain the environmental interference data around each data point. On this basis, the algorithm performs batch subtraction on the source data point and the environmental interference data of the corresponding point, and sorts the result data, and only selects a fixed number of data point indexes from large to small as the result data of the algorithm. Since the target data points finally screened by the radar algorithm only account for a very small part of the original data volume, and the data points with larger data values ​​also mean that the echo signal is stronger and more likely to be the desired target point, so under this premise, this non-strict threshold screening algorithm will not cause obvious data errors.

[0072] Furthermore, removing the environmental interference data from the target data point data to obtain interference-free data includes:

[0073] S401, performing a fast Fourier transform of a preset amount of data on the target data point data, and performing modulus calculation to obtain a first modulus calculation result;

[0074] S402, modulo the environmental interference data to obtain a second modulo result;

[0075] S403: Subtract the second modulo result from the first modulo result to obtain interference-free data.

[0076] In this embodiment, by performing a fast Fourier transform of a preset data amount on the target data point data and taking the modulus to obtain a first modulo result, the frequency distribution characteristics of the signal can be observed, which is suitable for processing a large amount of data; the environmental interference data is modulo to obtain a second modulo result, and the frequency spectrum amplitude of the interference is obtained by taking the modulus of the interference signal to obtain the frequency characteristics of the interference; the first modulo result is subtracted from the second modulo result to obtain interference-free data, and the interference signal is subtracted in the frequency domain, which is equivalent to filtering out the interference component, which can effectively remove background interference in the environment and retain the main characteristics of the target signal.

[0077] Further, the using the dual index determination method to screen the first data point to obtain the second data point includes:

[0078] According to the multi-dimensional information points of the first data points, data points in the first data points that exist in both the distance dimension information points and the speed dimension information points are screened out as second data points.

[0079] In this embodiment, the dual index determination method screens the constant false alarm rate algorithm results according to the non-strict thresholds of different dimensions, and compares and determines whether the data points are considered valid data points in the results of the two dimensions, so as to increase the credibility of the data points. The constant false alarm rate algorithm of the non-strict threshold screening performs an operation in the distance dimension and the speed dimension respectively and obtains two sets of data information. By comparing the corresponding indexes of the two sets of data points in the millimeter wave radar echo data tensor, the multipath effect is reduced to a certain extent by determining the dual index. After this stage of operation, the design further screens the remaining data point information through the spatial distance algorithm, compares the indexes of the data points and the credible target points in the data tensor in the distance dimension and the speed dimension, and excludes the data points whose indexes are too close to further reduce the impact of the multipath effect on the result data, so as to obtain the credible target points.

[0080] Specifically, for each first valid information point, its validity in different dimensions (such as distance, angle, and speed) is checked. Only information points that are considered valid in both distance and speed dimensions will be retained. Through multi-dimensional verification, the false detection rate is greatly reduced, and points that appear to be valid in a single dimension but may actually be noise or interference can be effectively filtered out. Requiring information points to be valid in multiple dimensions significantly reduces the detection of false targets and helps to separate and identify multiple targets in complex scenes.

[0081] like Figure 5The present invention discloses a method for determining a dual index of a constant false alarm rate algorithm with a non-strict threshold screening. The algorithm performs a non-strict threshold screening constant false alarm rate algorithm on the distance dimension and the Doppler dimension of the frequency domain data tensor. If a data point appears in the result of the constant false alarm rate algorithm with a non-strict threshold screening in the distance dimension, it is proved that it is considered as a valid information point in the distance dimension. If the data point also appears in the result of the constant false alarm rate algorithm with a non-strict threshold screening in the Doppler dimension, it is considered as a valid information point in both dimensions. This screening step can further enhance the credibility of the algorithm result and control the multipath effect to a certain extent. Thanks to the fact that the result data amount of the constant false alarm rate algorithm with a non-strict threshold screening is very small compared to the result data amount of the traditional constant false alarm rate algorithm, the actual amount of calculation required for this screening step is very small.

[0082] Furthermore, the using of a spatial distance judgment algorithm to screen the second data points to obtain valid data points includes:

[0083] S501, obtaining multiple target points pointing to the same physical target to be detected by comparing the sizes of the distance dimension indexes and the sizes of the speed dimension indexes of the second data points, wherein the multiple target points pointing to the same physical target to be detected are points whose distance dimension indexes and speed dimension indexes are close;

[0084] S502: Select a point with a larger value in the fast Fourier transform result data in the angle dimension from among the multiple target points as a valid data point.

[0085] In this embodiment, the algorithm judges the spatial position index corresponding to the target information in the result of the constant false alarm rate algorithm based on the non-strict threshold, and selects the information points according to the distance size of the effective information points in the data space. By comparing the size between the distance dimension index and the size between the speed dimension index of the second effective information point, the difference in distance and speed between different effective information points is calculated. If the distance and speed of multiple points are similar, it can be reasonably inferred that they may point to the same physical target to be detected, and multiple observation points of the same physical target can be effectively identified, thereby improving the utilization rate of data. Among the multiple identified target points, it is necessary to further analyze their performance in the angle dimension, select points with larger values ​​in the fast Fourier results of the angle dimension, and usually these points represent more obvious target signals. Among the multiple target points, select points with larger values ​​in the fast Fourier result data of the angle dimension to obtain target point data information. The angle information can be used to further verify the target position and enhance the ability to confirm the target. By screening out points with large values, the amount of data for subsequent processing is reduced, thereby improving the processing efficiency of the system.

[0086] like Figure 6The spatial distance judgment algorithm of the present invention is a method for determining the double index of a constant false alarm rate algorithm with non-strict threshold screening. The result data generated by this method is used as input, and the index information of the valid information point is judged by the three-dimensional spatial index of the corresponding data point in the echo data tensor. By comparing the size between the distance dimension indexes and the size between the speed dimension indexes of the valid information points, it is judged whether the two valid information points point to the same physical target to be detected. For two valid information points pointing to the same physical target to be detected, the point with the larger value in the angle fast Fourier result data is selected as the final target, so as to control the multipath effect.

[0087] Further, the detecting the multiple targets based on the valid data points includes:

[0088] According to the target point data information, the dynamic information of the physical target to be detected is calculated from multiple dimensions to obtain the dynamic information of multiple physical targets to be detected, thus realizing multi-target detection based on millimeter-wave radar data.

[0089] In this embodiment, the dynamic information of each physical target to be detected is calculated through the filtered target point data information. Based on the comprehensive calculation of distance, speed and time, the position information of the current target can be obtained. The relative speed of the target can be obtained by using the Doppler effect and combining the speed component of the target point. The distance, speed and angle measurement information of multiple targets can be finally obtained after calculation using the radar principle formula. By integrating data from multiple dimensions, a more comprehensive and accurate dynamic state of the target can be provided, and multiple physical targets can be detected and tracked at the same time to meet the needs of complex scenarios.

[0090] Example 2

[0091] A system for implementing the multi-target detection method based on millimeter wave radar, comprising:

[0092] Millimeter wave radar, CPU, acceleration core, DDR, FPGA and display terminal;

[0093] The millimeter wave radar is used to collect echo data;

[0094] The central processing unit is used to implement the small point calculation and logical judgment required by the multi-target detection method;

[0095] The acceleration core is used to implement the digital signal processing operations required in the multi-target detection method;

[0096] The DDR utilizes double data transmission to increase the data transmission rate;

[0097] The FPGA is used to integrate the central processing unit, the acceleration core and the DDR;

[0098] The display terminal is used to display the calculation results of the multi-target detection method.

[0099] In this embodiment, the millimeter wave radar system transmits millimeter wave signals and receives echo data, detects the position, speed and other characteristics of the object by measuring the reflection change of the signal, and can achieve high-precision detection at short distances and for relatively small objects; the central processing unit is responsible for executing program instructions, performing logical judgment and control system operations, processing basic small-point operations, managing data flows and control signals between modules, coordinating the operation of the entire system, and can execute different algorithms and processing tasks to adapt to different detection requirements; the acceleration core is a computing unit designed specifically for processing digital signals, which can quickly execute complex calculations and processing tasks related to multi-target detection, has parallel processing capabilities, and can simultaneously Handling complex signal processing in multiple data streams and optimizing specific algorithms significantly improves computing speed and shortens target detection latency. DDR memory supports double data transmission and can transmit data simultaneously on both the rising and falling edges of each clock cycle, greatly increasing data transmission rates and providing fast data storage and access capabilities, avoiding data processing bottlenecks and improving overall system performance. FPGA integrates CPU, acceleration core and DDR. Through the parallel processing capabilities of FPGA, it can effectively execute complex signal processing algorithms, realize parallel signal processing and shorten response time. The display terminal is used to present real-time calculation results of multi-target detection and visualize the processed data.

[0100] like Figure 7 The hardware configuration fully considers the digital signal processing requirements of the millimeter wave radar algorithm. It accelerates the large number of DSP operations required by the millimeter wave radar algorithm by configuring the digital signal processing acceleration hardware, and implements the other small-point operations and logical judgments required by the millimeter wave radar algorithm by configuring the CPU. Through the coordinated work of the millimeter wave radar, central processing unit, acceleration core, DDR, FPGA and display terminal, an efficient, flexible and real-time multi-target detection platform is built. The fine coordination between the modules ensures that the system can accurately and quickly detect and track multiple targets in a dynamic environment, thereby improving the overall performance and application effect.

[0101] Furthermore, the acceleration core is used to implement the digital signal processing operations required in the multi-target detection method, including:

[0102] The acceleration core realizes the digital signal processing operations required in the multi-target detection method through one-dimensional sequence FFT operations, one-dimensional sequence FIR operations, one-dimensional sequence point multiplication operations, one-dimensional sequence modulus operations and one-dimensional sequence addition and subtraction operations.

[0103] In this embodiment, the acceleration core implements the digital signal processing required in the multi-target detection method through a variety of one-dimensional sequence operations, converts the time domain signal into a frequency domain representation through a one-dimensional sequence FFT operation, implements digital filtering through a one-dimensional sequence FIR operation using a convolution operation to maintain the time characteristics of the signal and remove environmental interference, multiplies the corresponding elements of two vectors through a one-dimensional sequence point multiplication operation and then sums them, calculates the amplitude of the complex sequence through a one-dimensional sequence modulus operation to obtain signal strength information, and performs element-by-element addition or subtraction operations on two sequences through a one-dimensional sequence addition and subtraction operation to achieve basic signal combination and comparison. By implementing these basic one-dimensional sequence operations in the acceleration core, the system can efficiently process radar signals and achieve fast and accurate multi-target detection. This method not only improves the performance of the system, but also maintains sufficient flexibility to adapt to different application scenarios and algorithm requirements.

[0104] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A multi-target detection method based on millimeter wave radar, characterized in that: include: Acquire original echo data collected by the millimeter wave radar, wherein the original echo data includes target data points of multiple targets; Selecting a preset number of data points from the target data points of the plurality of targets, and using the selected data points as valid data points; The multiple targets are detected based on the valid data points.

2. The multi-target detection method based on millimeter wave radar according to claim 1, characterized in that: The step of selecting a preset number of data points from the target data points of the plurality of targets and using the selected data points as valid data points includes: Selecting a preset number of data points from the target data points of the multiple targets using a constant false alarm rate algorithm with non-strict threshold screening to obtain a first data point; Using a dual index determination method to screen the first data point to obtain a second data point; The second data points are screened using a spatial distance judgment algorithm to obtain valid data points.

3. The multi-target detection method based on millimeter wave radar according to claim 2 is characterized in that: The constant false alarm rate algorithm using non-strict threshold screening selects a preset number of data points from the target data points of the multiple targets to obtain the first data point, including: Performing FIR calculation on the target data points of the multiple targets to obtain environmental interference data around each target data point; Removing the environmental interference data from the target data point data to obtain interference-free data; The interference-removed data are sorted, and a preset number of data point indexes are selected from large to small to obtain a first data point.

4. The multi-target detection method based on millimeter wave radar according to claim 3 is characterized in that: The step of removing the environmental interference data from the target data point data to obtain interference-free data includes: Performing a fast Fourier transform of a preset amount of data on the target data point data, and performing modulus calculation to obtain a first modulus calculation result; Modulo the environmental interference data to obtain a second modulo result; The first modulo result is subtracted from the second modulo result to obtain interference-free data.

5. The multi-target detection method based on millimeter wave radar according to claim 2 is characterized in that: The step of screening the first data point by using the dual index determination method to obtain the second data point includes: According to the multi-dimensional information points of the first data points, data points in the first data points that exist in both the distance dimension information points and the speed dimension information points are screened out as second data points.

6. The multi-target detection method based on millimeter wave radar according to claim 2 is characterized in that: The using of a spatial distance judgment algorithm to screen the second data points to obtain valid data points includes: By comparing the sizes of the distance dimension indexes and the speed dimension indexes of the second data points, a plurality of target points pointing to the same physical target to be detected are obtained, wherein the plurality of target points pointing to the same physical target to be detected are points whose distance dimension index and speed dimension index are close to each other; Among the multiple target points, a point with a larger value in the fast Fourier result data in the angle dimension is selected as a valid data point.

7. The multi-target detection method based on millimeter wave radar according to claim 1, characterized in that: The detecting the multiple targets based on the valid data points includes: According to the valid data points, the information of each target is calculated from multiple dimensions to obtain the dynamic information of multiple targets, thereby realizing multi-target detection based on millimeter-wave radar data.

8. A system for implementing a multi-target detection method based on millimeter wave radar as described in claims 1 to 7, characterized in that: include: Millimeter wave radar, CPU, acceleration core, DDR, FPGA and display terminal; The millimeter wave radar is used to collect echo data; The central processing unit is used to implement the small point calculation and logical judgment required by the multi-target detection method; The acceleration core is used to implement the digital signal processing operations required in the multi-target detection method; The DDR utilizes double data transmission to increase the data transmission rate; The FPGA is used to integrate the central processing unit, the acceleration core and the DDR; The display terminal is used to display the calculation results of the multi-target detection method.

9. A system according to claim 8, characterized in that: The acceleration core is used to implement the digital signal processing operations required in the multi-target detection method, including: The acceleration core realizes the digital signal processing operations required in the multi-target detection method through one-dimensional sequence FFT operations, one-dimensional sequence FIR operations, one-dimensional sequence point multiplication operations, one-dimensional sequence modulus operations and one-dimensional sequence addition and subtraction operations.

Images