Multipath identification method and apparatus, vehicle, and medium

Abstract

A multipath identification method and apparatus, a vehicle, and a medium. The method comprises: preprocessing an echo signal of a millimeter wave radar, so as to obtain static point cloud data and dynamic point cloud data in the echo signal; clustering the dynamic point cloud data by means of improved DBSCAN clustering algorithm, so as to obtain multi-cluster dynamic data; on the basis of an object reflection boundary parameter and cluster center information in the multi-cluster dynamic data, determining at least one matching set, the matching set containing at least two cluster centers, and the object reflection boundary parameter being determined on the basis of the static point cloud data; and for each matching set, performing identification on the cluster centers in the matching set to obtain a target cluster center, and determining the target cluster center and the point cloud data in a cluster where the target cluster center is located as a multipath object in the echo signal. The method can effectively identify a multipath object, thereby improving the accuracy and robustness of vehicle-mounted millimeter wave radar detection.

Description

Multipath identification method, device, vehicle and medium

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on August 3, 2023, with application number 202310978095.3. The entire contents of the above application are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of millimeter wave radar technology, and in particular to a multipath identification method, device, vehicle, and medium. Background Art

[0003] With the rapid development of millimeter-wave radar technology, automotive millimeter-wave radar has been widely used in the field of autonomous driving. Compared with other sensors, millimeter-wave radar offers advantages such as all-day, all-weather operation, high ranging and speed measurement accuracy, and relatively low cost. However, when using automotive millimeter-wave radar to perceive the environment, the electromagnetic wave signals emitted by the radar often produce complex multipath signals due to reflection and diffraction. When processing radar echo data using conventional signal processing methods, multipath signals can produce ghosting, causing false alarms, which interferes with the detection and tracking of real targets. In particular, when automotive millimeter-wave radar operates in scenes such as road barriers, the interference caused by radar multipath signals is particularly severe.

[0004] However, the existing multipath recognition method has a high misjudgment rate of multipath targets, resulting in low recognition accuracy and cannot be better applied in vehicle-mounted millimeter-wave radar detection.

[0005] Summary of the Invention

[0006] The present application provides a multipath identification method, device, vehicle, and medium to effectively identify multipath targets and improve the accuracy and robustness of vehicle-mounted millimeter-wave radar detection.

[0007] According to one aspect of the present application, a multipath identification method is provided. The method is applied to a vehicle equipped with a millimeter-wave radar, and the method includes:

[0008] Preprocessing the echo signal of the millimeter-wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal;

[0009] The improved DBSCAN clustering algorithm is used to cluster the dynamic point cloud data to obtain multiple clusters of dynamic data;

[0010] Determine at least one matching set based on target reflection boundary parameters and cluster center information in the multi-cluster dynamic data, wherein the matching set includes at least two cluster centers, and the target reflection boundary parameters are determined according to the static point cloud data;

[0011] For each matching set, the cluster center in the matching set is identified to obtain the target cluster center, and the target cluster center and the point cloud data in the cluster where the target cluster center is located are determined as the multipath target in the echo signal.

[0012] According to another aspect of the present application, a multipath identification device is provided. The device is configured on a vehicle equipped with a millimeter-wave radar, and the device includes:

[0013] A preprocessing module, used to preprocess the echo signal of the millimeter wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal;

[0014] A clustering module, configured to cluster the dynamic point cloud data using an improved DBSCAN clustering algorithm to obtain multiple clusters of dynamic data;

[0015] a determination module, configured to determine at least one matching set based on target reflection boundary parameters and cluster center information in the multi-cluster dynamic data, wherein the matching set includes at least two cluster centers, and the target reflection boundary parameters are determined according to the static point cloud data;

[0016] The identification module is used to identify the cluster center in each matching set to obtain the target cluster center, and determine the target cluster center and the point cloud data in the cluster where the target cluster center is located as the multipath target in the echo signal.

[0017] According to another aspect of the present application, a vehicle is provided, comprising:

[0018] at least one processor; and

[0019] a memory communicatively connected to the at least one processor; wherein,

[0020] The memory stores a computer program that can be executed by the at least one processor. The computer program is executed by the at least one processor to enable the at least one processor to perform the multipath identification method described in any embodiment of the present application.

[0021] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the multipath identification method described in any embodiment of the present application when executed.

[0022] An embodiment of the present application provides a multipath identification method, device, vehicle and medium, the method being applied to a vehicle equipped with a millimeter-wave radar, the method comprising: preprocessing the echo signal of the millimeter-wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal; clustering the dynamic point cloud data using an improved DBSCAN clustering algorithm to obtain multi-cluster dynamic data; determining at least one matching set based on target reflection boundary parameters and cluster center information in the multi-cluster dynamic data, the matching set comprising at least two cluster centers, the target reflection boundary parameters being determined based on the static point cloud data; for each matching set, identifying the cluster center in the matching set to obtain a target cluster center, and determining the target cluster center and the point cloud data in the cluster where the target cluster center is located as a multipath target in the echo signal. Utilizing the above technical solution, by adopting the improved DBSCAN clustering algorithm to cluster the dynamic point cloud data in the echo signal, and determining the matching set based on the target reflection boundary parameters and cluster center information, it can provide a basis for the subsequent identification of the cluster center in the matching set, thereby effectively identifying multipath targets and improving the accuracy and robustness of vehicle-mounted millimeter-wave radar detection.

[0023] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0025] FIG1 is a flow chart of a multipath identification method provided according to Embodiment 1 of the present application;

[0026] FIG2 is a schematic diagram of a vehicle and a millimeter-wave radar according to Embodiment 1 of the present application;

[0027] FIG3 is a flow chart of a multipath identification method provided according to Embodiment 2 of the present application;

[0028] FIG4 is a flow chart of another multipath identification method provided according to the second embodiment of the present application;

[0029] FIG5 is a schematic diagram of a multipath identification scenario provided according to the second embodiment of the present application;

[0030] FIG6 is a schematic diagram of a radar point cloud provided according to the second embodiment of the present application;

[0031] FIG7 is a schematic diagram of another radar point cloud provided according to the second embodiment of the present application;

[0032] FIG8 is a schematic diagram of a static point cloud provided according to the second embodiment of the present application;

[0033] FIG9 is a schematic diagram of a static point cloud provided according to the second embodiment of the present application;

[0034] FIG10 is a schematic diagram of a multipath target provided according to Embodiment 2 of the present application;

[0035] FIG11 is a schematic diagram of a processing result provided according to Example 2 of the present application;

[0036] FIG12 is a schematic structural diagram of a multipath identification device provided according to Embodiment 3 of the present application;

[0037] Figure 13 is a structural schematic diagram of a vehicle provided according to Example 4 of the present application. DETAILED DESCRIPTION

[0038] In order to enable people in the field of this embodiment to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of this application.

[0039] It should be noted that the terms "target", "original" etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0040] Example 1

[0041] FIG1 is a flowchart of a multipath identification method provided according to a first embodiment of the present application. This embodiment is applicable to situations where multipath targets are identified. The method can be performed by a multipath identification device, which can be implemented in the form of hardware and / or software and can be configured in a vehicle.

[0042] It can be considered that when the vehicle-mounted millimeter-wave radar emits electromagnetic waves to sense the environment, it often generates complex multipath signals due to phenomena such as reflection and diffraction of electromagnetic waves, and the generation of multipath signals will interfere with the subsequent processing of radar signals, such as target detection and target tracking. In particular, when the vehicle-mounted millimeter-wave radar works in scenes such as road fences, the interference caused by radar multipath signals is particularly serious. Based on this, the embodiment of the present application provides a multipath identification method that effectively identifies and eliminates multipath targets caused by reflective surfaces such as road boundaries and walls from the radar point cloud level, thereby improving the accuracy and robustness of vehicle-mounted millimeter-wave radar detection. As shown in Figure 1, the method includes:

[0043] S110 , preprocessing the echo signal of the millimeter-wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal.

[0044] Static point cloud data may refer to relevant data of a static point cloud, and dynamic point cloud data may refer to relevant data of a dynamic point cloud. For example, the relevant data may include coordinate data, speed information, and / or energy information.

[0045] In this embodiment, a millimeter-wave radar can be installed on the vehicle, such as a vehicle-mounted MIMO millimeter-wave radar, wherein the transmitting antenna of the millimeter-wave radar can radiate a linear frequency modulated continuous wave (LFMCW) signal into the scene and receive a corresponding original echo signal. The original echo signal can obtain an intermediate frequency signal after preliminary mixing and filtering. The intermediate frequency signal is sampled using an ADC to obtain original ADC data, such as an echo signal. The echo signal not only contains the direct path echo signal of the target object in the scene, but also contains multipath echo signals generated by reflection, diffraction, and other paths of electromagnetic waves between different objects in the scene, as well as other noise signals. Therefore, a series of preprocessing is required for the echo signal to obtain static point cloud data and dynamic point cloud data in the echo signal. The specific process of preprocessing is not limited and can be determined according to actual conditions. For example, the echo signal can be preliminarily processed first to obtain the radar point cloud of the echo signal in the Cartesian coordinate system, and then the static point cloud data and dynamic point cloud data can be determined based on the radar point cloud. The echo signal can also be directly input into the training model to obtain the static point cloud data and dynamic point cloud data in the echo signal. This embodiment does not limit this.

[0046] S120 , clustering the dynamic point cloud data using an improved DBSCAN clustering algorithm to obtain multiple clusters of dynamic data.

[0047] The improved DBSCAN clustering algorithm can be considered as an improved algorithm based on the traditional DBSCAN clustering algorithm. For example, the improved DBSCAN clustering algorithm can add some parameters and update some specific calculations.

[0048] In one embodiment, the parameters of the improved DBSCAN clustering algorithm include at least one of cluster arc radius, cluster arc angle, angle deviation, and speed information.

[0049] During the clustering process, new parameters can be added to the traditional DBSCAN clustering algorithm parameters as needed. For example, at least one of the clustering arc radius, clustering arc angle, angular deviation, and velocity information can be added based on actual conditions. For example, when clustering dynamic point cloud data, velocity information can be added to the traditional DBSCAN clustering algorithm parameters as a consideration to improve clustering rationality. Multi-cluster dynamic data is the result of clustering, with each cluster containing multiple dynamic data.

[0050] After obtaining the static point cloud data and dynamic point cloud data in the echo signal, the improved DBSCAN clustering algorithm can be used to cluster the dynamic point cloud data to obtain multiple clusters of dynamic data for subsequent target recognition.

[0051] S130 . Determine at least one matching set based on target reflection boundary parameters and cluster center information in the multi-cluster dynamic data, wherein the matching set includes at least two cluster centers, and the target reflection boundary parameters are determined according to the static point cloud data.

[0052] The target reflection boundary parameters characterize the parameters of the target reflection boundary. For example, the target reflection boundary can be considered as the road boundary in scenarios such as road barriers. The target reflection boundary parameters can be determined from the static point cloud data in the echo signal. A matching set can be understood as a collection of cluster centers, that is, a matching set can contain at least two cluster centers.

[0053] Cluster center information may refer to information related to the cluster center in each cluster of dynamic data, such as the coordinate information of the cluster center, the speed of the cluster center, the number of dynamic point clouds in the cluster where the cluster center is located, etc. The cluster center information may be determined based on information about the dynamic point clouds in the cluster where the cluster center is located. For example, the speed of the cluster center may be the average speed of all dynamic point clouds in the cluster where the cluster center is located.

[0054] Specifically, this embodiment can determine at least one matching set based on the target reflection boundary parameters and the cluster center information in the multi-cluster dynamic data. The determination method can be, for example, traversing all cluster centers and matching multiple cluster centers to obtain at least one matching set, or it can be simply dividing multiple cluster centers based on the target reflection boundary parameters to obtain at least one matching set. This embodiment does not limit this.

[0055] In one embodiment, determining at least one matching set based on the target reflection boundary parameter and cluster center information in the multi-cluster dynamic data includes:

[0056] According to the target reflection boundary parameters and the cluster center information in the multi-cluster dynamic data, each cluster center in the multi-cluster dynamic data is matched to obtain at least one matching set.

[0057] In this step, the cluster centers in the multi-cluster dynamic data can be matched according to the target reflection boundary parameters and the cluster center information in the multi-cluster dynamic data to obtain at least one matching set. The specific means of matching are not limited. For example, at least two adjacent cluster centers can be considered as a matching set based on the coordinate information of the cluster centers, or the cluster centers can be matched according to the size of the speed represented by the speed information.

[0058] For example, multipath ghosts caused by reflective surfaces such as road fences and walls are theoretically located at the mirror image position of the real target with respect to the reflective boundary. Therefore, in this embodiment, for each cluster center, such as cluster center a, the mirror image point b of cluster center a with respect to the reflective boundary is first determined. A search area is then set with mirror image point b as the center. Cluster center matching is then performed by searching for other cluster centers that fall within the search area, thereby obtaining at least one matching set. The search area can be set by any means; for example, the search area can be a circle with mirror image point b as the center and a set radius. The set radius can be determined empirically, such as 3 meters.

[0059] S140 . For each matching set, identify the cluster center in the matching set to obtain a target cluster center, and determine the target cluster center and the point cloud data in the cluster where the target cluster center is located as a multipath target in the echo signal.

[0060] The target cluster center may refer to the cluster center of the multipath targets.

[0061] After obtaining at least one matching set through the above steps, the cluster center in the matching set can be identified for each matching set to obtain the target cluster center. The specific identification method is not limited. For example, the target cluster center can be identified by calculating the target parameters of each cluster center, or the target cluster center can be identified by comparing the cluster center information of each cluster center in the matching set, such as identifying the cluster center with an obviously too small speed or the highest energy in the cluster center information as the target cluster center; after all target cluster centers are identified, the target cluster center and all point cloud data in the cluster where the target cluster center is located can be determined as multipath targets in the echo signal, and the identification of all multipath targets in the echo signal has been completed.

[0062] In one embodiment, identifying the cluster centers in the matching set to obtain the target cluster centers includes:

[0063] Traversing each cluster center in the matching set, determining a sum of target parameters corresponding to the cluster center, the target parameter including at least one of distance information, speed information, capability information, and the number of point clouds;

[0064] Whether the cluster center is a target cluster center is determined based on the sum of the target parameters.

[0065] The target parameter may be understood as a pre-set parameter used to determine whether the cluster center is the target cluster center. For example, the target parameter may include at least one of distance information, speed information, capability information, and the number of point clouds.

[0066] Specifically, the process of identifying the cluster center can be, for example, traversing each cluster center in the matching set, and determining the sum of the target parameters corresponding to each cluster center. For example, the target parameters corresponding to each cluster center can be calculated first, and then the sum of the calculated target parameters can be determined. After determining the sum of the target parameters corresponding to the cluster center, whether the cluster center is the target cluster center can be judged based on the determined sum of the target parameters. For example, the judgment can be made based on the specific size of the sum of the target parameters. It can also be judged whether the cluster center is the target cluster center by comparing the sum of the target parameters with a preset threshold. The preset threshold can be determined according to the actual situation of the target parameters. For example, when the target parameters include distance information, speed information, capability information and the number of point clouds, the preset threshold can be 4, so that when the sum of the target parameters is equal to 4, the cluster center can be judged to be the target cluster center; when the sum of the target parameters is less than 4, it can be judged that the cluster center is not the target cluster center.

[0067] A multipath identification method provided in a first embodiment of the present application is applied to a vehicle equipped with a millimeter-wave radar. The method comprises: preprocessing the echo signal of the millimeter-wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal; clustering the dynamic point cloud data using an improved DBSCAN clustering algorithm to obtain multiple clusters of dynamic data; determining at least one matching set based on target reflection boundary parameters and cluster center information in the multiple clusters of dynamic data, the matching set comprising at least two cluster centers, the target reflection boundary parameters being determined based on the static point cloud data; and for each matching set, identifying the cluster center in the matching set to obtain the target cluster center, and determining the target cluster center and the point cloud data in the cluster containing the target cluster center as the multipath target in the echo signal. Using this method, by clustering the dynamic point cloud data in the echo signal using the improved DBSCAN clustering algorithm and determining the matching set based on the target reflection boundary parameters and cluster center information, a basis is provided for subsequent identification of the cluster centers in the matching set, thereby effectively identifying multipath targets and improving the accuracy and robustness of vehicle-mounted millimeter-wave radar detection.

[0068] In one embodiment, the preprocessing of the echo signal of the millimeter wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal includes:

[0069] Extracting point cloud data corresponding to the radar point cloud in the echo signal;

[0070] The radar point cloud is separated into static and dynamic point cloud data based on the point cloud data to obtain static point cloud data and dynamic point cloud data in the echo signal.

[0071] The point cloud data may be data related to the radar point cloud, such as coordinate data, speed information, etc.

[0072] In one embodiment, the echo signal preprocessing process may include, for example, first extracting point cloud data corresponding to the radar point cloud in the echo signal. For example, the data information of each detection point in the echo signal is first determined, and then the detection points are converted to obtain a radar point cloud in a Cartesian coordinate system. Based on the detection point data, the point cloud data of the corresponding radar point cloud can be determined. After the point cloud data is extracted, the radar point cloud can be separated into static and dynamic point cloud data based on the extracted point cloud data. The static and dynamic separation method can be based on the specific size of the point cloud data, such as velocity information or energy information, to separate static and dynamic point cloud data.

[0073] In one embodiment, the point cloud data includes speed information, and the performing of static and dynamic point cloud separation on the radar point cloud based on the point cloud data to obtain static point cloud data and dynamic point cloud data in the echo signal includes:

[0074] Determine a target radial velocity, where the target radial velocity is the radial velocity of the static point cloud relative to the millimeter-wave radar;

[0075] The radar point cloud is separated into static and dynamic point cloud according to the radial velocity of the target and the velocity information of the radar point cloud, so as to obtain static point cloud data and dynamic point cloud data in the echo signal.

[0076] The target radial velocity can be considered as the radial velocity of the static point cloud relative to the millimeter-wave radar.

[0077] In one embodiment, the radial velocity of the static point cloud relative to the millimeter wave radar may be determined, and then the static point cloud data and the dynamic point cloud data in the radar point cloud may be separated according to the determined target radial velocity.

[0078] For example, FIG2 is a schematic diagram of a vehicle and a millimeter-wave radar according to the first embodiment of the present application. As shown in FIG2 , the target may be a static point cloud. Assuming that the target is at the same height as the millimeter-wave radar, the absolute radial velocity (i.e., the target radial velocity) v between the static point cloud and the millimeter-wave radar is D It can be expressed as: D =-(cos(a)(v C，x -wy′)+sin(a)(v C，y +wx′)), where v C，x , v C，y , w represents the lateral velocity component, longitudinal velocity component and rotation speed of the vehicle respectively, (x′, y′) is the position of the millimeter-wave radar in the vehicle coordinate system, and a is the angle between the radar normal and the vehicle coordinate system.

[0079] Then, the target radial velocity can be used as a reference value. If the velocity of the radar point cloud differs significantly from the reference velocity, it can be said that the point cloud is moving relative to the ground, that is, it is a dynamic point cloud (that is, based on the target radial velocity and the velocity information of the radar point cloud, the radar point cloud is separated into static and dynamic point cloud data in the echo signal). In specific processing, a reference value threshold Δv can be set to make a static and dynamic point cloud judgment on the radar point cloud. Specifically, the judgment can be made according to the following formula: Among them, S and D represent the static point cloud set and the dynamic point cloud set respectively, and the reference value threshold Δv is taken as the empirical value 0.8m / s, v radar is the velocity of the radar point cloud.

[0080] Example 2

[0081] FIG3 is a flow chart of a multipath identification method according to Embodiment 2 of the present application, which is an optimization of the above-described embodiments. In this embodiment, the process prior to determining at least one matching set based on target reflection boundary parameters and cluster center information in the multi-cluster dynamic data is further refined to include performing a linear estimation of the static point cloud data using a Hough transform algorithm to determine the target reflection boundary parameters.

[0082] For details not yet provided in this embodiment, please refer to the first embodiment.

[0083] As shown in FIG3 , the method includes:

[0084] S210 : Preprocess the echo signal of the millimeter-wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal.

[0085] S220 , clustering the dynamic point cloud data using an improved DBSCAN clustering algorithm to obtain multiple clusters of dynamic data.

[0086] S230: Perform linear estimation on the static point cloud data using a Hough transform algorithm to determine target reflection boundary parameters.

[0087] The Hough transform algorithm can be used to parse defined features, such as lines, and is robust to noise. In this embodiment, the Hough transform algorithm can be used to perform linear estimation on static point cloud data to determine the target reflection boundary parameters. For example, linear estimation can be performed directly on the static point cloud data to determine the target reflection boundary parameters. Alternatively, the static point cloud data can be preliminarily processed and then linearly estimated on the processed point cloud data to determine the target reflection boundary parameters. The preliminary processing can, for example, include filtering out interfering point cloud data or clustering the static point cloud data.

[0088] In one embodiment, the linear estimation of the static point cloud data using the Hough transform algorithm to determine the target reflection boundary parameters includes:

[0089] The static point cloud data is clustered using an improved DBSCAN clustering algorithm to obtain multiple clusters of static data;

[0090] For each cluster of static data, a Hough transform algorithm is used to perform linear estimation on the static data to determine the reflection boundary parameters corresponding to the static data;

[0091] Invalid boundary parameters among the plurality of reflection boundary parameters are determined, and target reflection boundary parameters are obtained based on the invalid boundary parameters.

[0092] The multi-cluster static data is the multiple clusters obtained after clustering, each cluster containing multiple static data. The invalid boundary parameter is the invalid boundary parameter among the multiple reflection boundary parameters.

[0093] In one embodiment, an improved DBSCAN clustering algorithm can be used to cluster static point cloud data to obtain multiple clusters of static data. Then, for each cluster of static data, a Hough transform algorithm can be used to perform linear estimation on the static data contained in the cluster to determine the reflection boundary parameters corresponding to the cluster of static data, thereby obtaining all reflection boundary parameters. Then, by determining invalid boundary parameters among the multiple reflection boundary parameters, the final target reflection boundary parameters are obtained, such as by filtering the invalid boundary parameters among the multiple reflection boundary parameters to obtain the final target reflection boundary parameters. The specific process of linear estimation and determination of invalid boundary parameters is not further elaborated here, as long as the linear estimation and determination of invalid boundary parameters can be completed.

[0094] S240 . Determine at least one matching set based on target reflection boundary parameters and cluster center information in the multi-cluster dynamic data, wherein the matching set includes at least two cluster centers, and the target reflection boundary parameters are determined according to the static point cloud data.

[0095] S250 . For each matching set, identify the cluster center in the matching set to obtain a target cluster center, and determine the target cluster center and the point cloud data in the cluster where the target cluster center is located as a multipath target in the echo signal.

[0096] A second embodiment of the present application provides a multipath identification method, applied to a vehicle equipped with a millimeter-wave radar. The method comprises: preprocessing the echo signal of the millimeter-wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal; clustering the dynamic point cloud data using an improved DBSCAN clustering algorithm to obtain multiple clusters of dynamic data; linearly estimating the static point cloud data using a Hough transform algorithm to determine target reflection boundary parameters; determining at least one matching set based on the target reflection boundary parameters and cluster center information in the multiple clusters of dynamic data, the matching set comprising at least two cluster centers, the target reflection boundary parameters being determined based on the static point cloud data; and identifying the cluster centers in each matching set to obtain the target cluster center, and determining the target cluster center and the point cloud data in the cluster containing the target cluster center as the multipath target in the echo signal. Using this method, by linearly estimating the static point cloud data in the echo signal using the Hough transform algorithm, the target reflection boundary parameters can be accurately determined, providing a basis for subsequent effective identification of multipath targets and improving the robustness of vehicle-mounted millimeter-wave radar detection.

[0097] In one embodiment, determining an invalid boundary parameter among the plurality of reflection boundary parameters includes:

[0098] Determine the boundary length and slope parameters corresponding to each reflection boundary parameter;

[0099] An invalid boundary parameter among the plurality of reflection boundary parameters is determined based on each boundary length and each slope parameter.

[0100] In one embodiment, invalid boundary parameters can be determined by boundary length and slope parameters. For example, the corresponding boundary length and slope parameter can be determined for each reflection boundary parameter. After obtaining each boundary length and each slope parameter, the invalid boundary parameters can be selected according to the specific sizes of each boundary length and each slope parameter. For example, a reflection boundary parameter with a boundary length that is too short or a slope parameter that differs too much from the vehicle's driving direction can be determined as an invalid boundary parameter.

[0101] FIG4 is a flow chart of another multipath identification method provided in accordance with the second embodiment of the present application. As shown in FIG4 , after obtaining the ADC data (i.e., the echo signal) of the millimeter wave radar, the ADC The data is preprocessed to generate corresponding point cloud data, including static point cloud and dynamic point cloud (i.e., the echo signal of the millimeter wave radar is preprocessed to obtain static point cloud data and dynamic point cloud data in the echo signal); the obtained static point cloud can be used to complete the estimation of the road boundary (i.e., the Hough transform algorithm is used to linearly estimate the static point cloud data to determine the target reflection boundary parameters), and the estimated road boundary information is used to complete the matching of the moving target of the dynamic point cloud to obtain a matching set (i.e., at least one matching set is determined based on the target reflection boundary parameters and the cluster center information in the multi-cluster dynamic data); multipath judgment is performed based on the distance, speed, energy and point cloud quantity information of the dynamic point cloud to identify the multipath ghost in the matching set and obtain a multipath target recognition result (i.e., for each matching set, the cluster center in the matching set is identified to obtain the target cluster center, and the target cluster center and the point cloud data in the cluster where the target cluster center is located are determined as the multipath target in the echo signal), thereby eliminating the obtained multipath target.

[0102] The following is an exemplary description of the steps of the multipath identification method:

[0103] (1) Signal preprocessing: After the raw data of the millimeter wave radar is processed by channel virtualization, the ADC data can be represented by a 3D data matrix. The three dimensions can be represented as fast time dimension, slow time dimension and antenna dimension respectively. Taking a frame of data as an example, the data matrix can be recorded as Where M is the number of chirps in a frame; N is the number of sampling points for each chirp; and L is the number of antennas in the antenna array after virtualization in azimuth or elevation.

[0104] First, a fast Fourier transform (FFT) can be performed on the ADC data along the fast time dimension to obtain a range spectrum, such as S′(r, n, l) = FFT[S(m, n, l)], where FFT[·] represents the fast Fourier transform operator. The range spectrum can reflect the distance of objects in the scene relative to the radar.

[0105] Next, a fast Fourier transform can be performed on the range spectrum along the slow time dimension to obtain a range Doppler spectrum, such as RDM(r, v, l) = FFT[S′(r, n, l)]. The range Doppler spectrum can reflect the distance of objects in the scene relative to the radar and their mirror speed relative to the radar.

[0106] Furthermore, a constant false alarm rate (CFAR) detection can be performed on the range Doppler spectrum. There are many types of CFAR detectors. To reduce the impact of multi-target interference on target detection, this embodiment can use an OS-CFAR detector, such as P = OS-CFAR[RDM], where OS-CFAR[·] represents the CFAR operator. is the test result, p i Represents a detection, N o After CFAR detection, the targets that pass the threshold can be extracted, and their corresponding distance, speed, and energy amplitude information can be extracted.

[0107] Furthermore, the information between different virtual receiving antennas can be used to estimate the Direction of Arrival (DOA). To achieve high-resolution angle measurement, this embodiment can use a Minimum Variance Distortionless Response (MVDR) algorithm for angle measurement.

[0108] For example, the optimal weight vector w0 of the MVDR beamformer can be: in, is the correlation matrix, E{·} represents the mathematical expectation, Indicates detection p i Index on the range-Doppler spectrum, R -1 represents the inverse matrix of R; a(θ0) is the direction of the steering vector; a H (θ0) is the conjugate transpose of a(θ0).

[0109] The obtained MVDR angle spectrum P MVDR (θ0) can be estimated as In particular, after obtaining the angle spectrum, the spectrum can be normalized, and then the angle information corresponding to the peak target within -3dB can be extracted.

[0110] After processing the raw radar data through the above steps, we can obtain detection points representing the moving and static objects in the scene. Each detection point includes distance, angle, speed and energy information. Further, through coordinate transformation, we can obtain the radar point cloud in the Cartesian coordinate system. in, Represent the detection point p i The relative radar distance and angle, Cartesian coordinates representing the point cloud.

[0111] The generated radar point cloud set is recorded as o i Can represent a point cloud, N o is the number of point clouds, each containing the corresponding spatial coordinate position, velocity, and amplitude information (i.e., the point cloud data corresponding to the radar point cloud in the echo signal is extracted). The generated radar point cloud contains not only point clouds representing static objects in the scene, but also point clouds representing moving targets in the scene.

[0112] Finally, based on the speed and steering angle information of the radar carrier, that is, the car equipped with the radar, the radar point cloud can be separated into static and dynamic point clouds to obtain static point clouds and dynamic point clouds (that is, the radar point cloud is separated into static and dynamic point clouds based on the point cloud data to obtain static point cloud data and dynamic point cloud data in the echo signal).

[0113] (2) Road Boundary Estimation: In scenarios such as road fences, radar can often obtain a static point cloud that clearly represents the road boundary. Therefore, processing the radar’s static point cloud and estimating the road boundary can assist in the identification and elimination of multipath ghosts.

[0114] Specifically, the improved DBSCAN algorithm can be used to cluster the static point cloud S. The corresponding radar point cloud on the roadside often has a strong linear distribution feature in spatial distribution. In order to better cluster the point cloud generated by the road boundary, this embodiment can improve the traditional DBSCAN algorithm and add three new parameters based on the parameters of the traditional DBSCAN clustering algorithm, namely the radius ρ of the desired clustering arc, the angle θ between the desired clustering arc and the positive direction of the x-axis, and the desired angle deviation Δθ. At this time, the neighborhood (two arcs) formed by the new three parameters can be superimposed on the neighborhood of the traditional DBSCAN algorithm. Let the improved DBSCAN algorithm be IDBSCAN, then Φ=IDBSCAN(S), IDBSCAN(·) is the improved DBSCAN algorithm operator, Represents a collection of clusters, clusteri represents the i-th cluster, N a Indicates the number of clusters (ie, the static point cloud data is clustered using the improved DBSCAN clustering algorithm to obtain multiple clusters of static data).

[0115] After clustering, the Hough transform algorithm can be used to perform linear parameter estimation (such as straight line estimation) on each cluster to achieve parameter estimation of the road boundary. Assume that the road boundary is: y = kx + b, where k represents the slope of the line and b represents the intercept. The Hough transform algorithm converts the Cartesian coordinate information of the point cloud into the Hesse normal form: r = x a cosθ+y a sinθ, where a ,y a ) is a point A in the Cartesian coordinate system, r represents the perpendicular distance from the origin to the line passing through point A, and θ represents the rotation angle of the perpendicular line. Through the above conversion, all the lines passing through point A in the Cartesian coordinate system are converted into a curve in the (r, θ) reference coordinate system. Therefore, if multiple corresponding curves intersect, it means that there is a straight line passing through all the above points. In this embodiment, the point with the most intersections can be taken as is the result of straight line parameter estimation. Further, the Hesse normal parameters are converted into slope-intercept parameters: That is, the road boundary parameters are obtained (that is, for each cluster of static data, the Hough transform algorithm is used to perform linear estimation on the static data to determine the reflection boundary parameters corresponding to the static data).

[0116] After estimation, the length of the reflective boundary can be determined based on the boundary point cloud of each cluster. Furthermore, invalid reflective boundaries that are too short or whose line slopes differ significantly from the vehicle's direction of travel are eliminated (i.e., the boundary length and slope parameters corresponding to each reflective boundary parameter are determined; and invalid boundary parameters among the multiple reflective boundary parameters are determined based on the boundary lengths and slopes).

[0117] (3) Motion target matching: Generally, in the vehicle-mounted MIMO millimeter-wave radar system, targets such as cars on the road are all extended targets, that is, a moving target is represented by multiple discretely distributed point clouds. Based on this, the DBSCAN algorithm is used to cluster the obtained dynamic point cloud D. In particular, for the rigid body motion target scene, speed information can be taken into consideration during clustering to improve the rationality of point cloud clustering. Take the cluster center (x c ,y c ), representing a moving target Among them, N is the number of point clouds in a cluster, (x c ,y c ) represents the spatial position of the target.

[0118] After clustering, the speed and signal amplitude information that characterize each cluster center can be further extracted. In this embodiment, the average speed of all point clouds in the cluster is taken as the speed v of the cluster center. c , take the maximum signal energy of all point clouds in the cluster as the signal energy SNR of the cluster center c The above cluster centers and related information can be integrated into a set, expressed as T = {x c ,y c , r c , v c , SNR c , PN}, where PN represents the number of point clouds in the cluster corresponding to the cluster center. In a single frame of data, the clustering result is Indicates the distance from the cluster center to the radar, N at Indicates the number of clusters after dynamic point cloud clustering.

[0119] It can be considered that after the dynamic point cloud D is clustered, not only the center representing the moving target can be extracted, but also some sparse noise and multipath ghosts in the moving point cloud can be effectively identified as noise and removed.

[0120] Generally, multipath ghosts caused by reflective surfaces such as road fences and walls are theoretically located at the mirror image position of the real target with respect to the reflective boundary. Assuming that a road boundary is obtained as L1: y = kx + b, for example, the mirror image point of a moving target T0: (x0, y0) with respect to the reflective boundary can be expressed as (x mirr ,y mirr ),in,

[0121] Due to the influence of factors such as target measurement accuracy and road boundary estimation error, the multipath ghost is not absolutely located at the mirror position of the real target with respect to the estimated road boundary. Therefore, a search threshold ΔR can be set to fall on the mirror center (x mirr ,y mirr ) is the center of the circle, and the center of motion within the area with a radius of ΔR matches the original moving target T0. In this embodiment, ΔR is taken as an empirical value of 3m.

[0122] Traverse all cluster centers in the scene (i.e., dynamic point cloud cluster centers), complete cluster center matching based on the estimated road boundary, and obtain the matching result as Γ={{T i , T j}, ... {T m , T n}},(T i , T j , T m , T n∈T) (i.e., matching each cluster center in the multi-cluster dynamic data according to the target reflection boundary parameter and the cluster center information in the multi-cluster dynamic data to obtain at least one matching set).

[0123] (4) Multipath ghost identification and elimination: Based on the information of the moving target, i.e., the distance, velocity, energy amplitude information of the extracted cluster center, and the number of point clouds, a multipath target decision maker can be set to process the generated matching set. Assume that there is a pair of points to be decided in the matching set, T1 = {r1, v1, SNR1, NP1} and T2 = {r2, v2, SNR2, NP2}. The decision maker refers to the following four factors:

[0124] Factor δ1: distance. The relative distance between the multipath ghost and the radar is greater than the relative distance between the real target and the radar. At long distances, due to measurement errors, the relative distance between the multipath ghost and the radar is close to the relative distance between the real target and the radar.

[0125] Factor δ2: Speed. The multipath ghost speed has a special relationship with the true target speed.

[0126] Factor δ3: SNR. The reflection of electromagnetic waves will attenuate the echo energy, so the echo energy of the real target will be greater than the echo energy of the multipath ghost.

[0127] Factor δ4: number of point clouds. The number of point clouds representing the real target is not less than the number of point clouds of multipath ghosts corresponding to the real target. In this embodiment, the number of point clouds corresponding to the target is the number of point clouds in the cluster corresponding to the cluster center representing the target.

[0128] In summary, the multipath target decision device can be written as Among them, ε v They are speed thresholds, which can be taken as an empirical value of 1m / s in this embodiment.

[0129] By using a decision maker to judge each target in the matching set, multipath ghost identification can be achieved (i.e., traversing each cluster center in the matching set, determining the sum of the target parameters corresponding to the cluster center, the target parameter including at least one of distance information, speed information, capability information, and the number of point clouds; and judging whether the cluster center is the target cluster center based on the sum of the target parameters). If a cluster center is identified as a multipath ghost, the point clouds in its corresponding cluster are all marked as multipath ghosts (i.e., the target cluster center and the point cloud data in the cluster where the target cluster center is located are determined as multipath targets in the echo signal).

[0130] Figure 5 is a schematic diagram of a multipath identification scenario according to Example 2 of the present application. As shown in Figure 5 , a typical road fence and wall scene is depicted, with a moving car. The left side shows the scene view from the left camera, with a continuous fence on one side of the road. The right side shows the scene view from the right camera, with a continuous wall on the other side of the road. A millimeter-wave radar radiates electromagnetic wave signals into the scene to detect the target, which is constantly approaching the millimeter-wave radar.

[0131] Figure 6 is a schematic diagram of a radar point cloud provided according to Example 2 of the present application. As shown in Figure 6, it is an accumulation diagram of the radar point cloud after preprocessing of radar multi-frame data. It can be found that the distribution of radar point clouds caused by road fences and walls is relatively obvious.

[0132] Figure 7 is a schematic diagram of another radar point cloud provided according to Example 2 of the present application. As shown in Figure 7, the cumulative results of multiple frames of dynamic point clouds are displayed. It can be seen from the figure that in addition to the point cloud trajectory of the moving target, the multipath ghosts caused by the fences and walls on both sides of the road are also very obvious.

[0133] Figure 8 is a schematic diagram of a static point cloud provided according to Example 2 of the present application. As shown in Figure 8, the radar point cloud of a single frame data is separated into dynamic and static point clouds, and the static point cloud separation result is the static point cloud; Figure 9 is a schematic diagram of a static point cloud provided according to Example 2 of the present application. As shown in Figure 9, after linear estimation of the static point cloud data using the Hough transform algorithm, the estimation result of the road boundary, that is, the target reflection boundary parameter, can be determined. It can be seen that the boundaries of the fences and walls on both sides of the road can be effectively estimated.

[0134] Figure 10 is a schematic diagram of a multipath target provided according to Example 2 of the present application. As shown in Figure 10, the multipath ghost recognition results accumulated over multiple frames are given. It can be seen from the figure that some sparse noise or multipath ghosts can be filtered out through clustering processing, and multipath ghosts that cannot be eliminated by clustering processing can also be well identified.

[0135] Figure 11 is a schematic diagram of a processing result provided according to Example 2 of the present application. As shown in Figure 11, it is the final processing result of the signal after eliminating the multipath target. It can be seen from the figure that the cluttered multipath ghosts caused by the fences and walls on both sides of the road are effectively eliminated, and the motion point cloud of the real target is effectively highlighted.

[0136] Example 3

[0137] FIG12 is a schematic diagram of the structure of a multipath identification device provided according to Embodiment 3 of the present application. As shown in FIG12 , the device includes:

[0138] A preprocessing module 310 is used to preprocess the echo signal of the millimeter wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal;

[0139] A clustering module 320 is configured to cluster the dynamic point cloud data using an improved DBSCAN clustering algorithm to obtain multiple clusters of dynamic data;

[0140] a determination module 330 for determining at least one matching set based on target reflection boundary parameters and cluster center information in the multi-cluster dynamic data, the matching set including at least two cluster centers, the target reflection boundary parameters being determined based on the static point cloud data;

[0141] The identification module 340 is used to identify the cluster center in each matching set to obtain the target cluster center, and determine the target cluster center and the point cloud data in the cluster where the target cluster center is located as the multipath target in the echo signal.

[0142] A multipath identification device is provided in a third embodiment of the present application. A preprocessing module preprocesses the echo signal of a millimeter-wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal. A clustering module clusters the dynamic point cloud data using an improved DBSCAN clustering algorithm to obtain multiple clusters of dynamic data. A determination module determines at least one matching set based on target reflection boundary parameters and cluster center information in the multiple clusters of dynamic data, the matching set comprising at least two cluster centers, the target reflection boundary parameters being determined based on the static point cloud data. An identification module identifies the cluster center in each matching set to obtain a target cluster center, and identifies the target cluster center and the point cloud data in the cluster containing the target cluster center as a multipath target in the echo signal. This device clusters the dynamic point cloud data in the echo signal using the improved DBSCAN clustering algorithm and determines the matching set based on the target reflection boundary parameters and cluster center information, providing a basis for subsequent identification of the cluster centers in the matching set. This effectively identifies multipath targets and improves the accuracy and robustness of vehicle-mounted millimeter-wave radar detection.

[0143] Optional pre-processing modules include:

[0144] an extraction unit, configured to extract point cloud data corresponding to the radar point cloud in the echo signal;

[0145] A separation unit is used to separate the radar point cloud from static point cloud based on the point cloud data to obtain static point cloud data and dynamic point cloud data in the echo signal.

[0146] Optionally, the point cloud data includes speed information, and the separation unit is specifically configured to:

[0147] Determine a target radial velocity, where the target radial velocity is the radial velocity of the static point cloud relative to the millimeter-wave radar;

[0148] The radar point cloud is separated into static and dynamic point cloud according to the radial velocity of the target and the velocity information of the radar point cloud, so as to obtain static point cloud data and dynamic point cloud data in the echo signal.

[0149] Optionally, the multipath identification device provided in the third embodiment of the present application further includes:

[0150] A linear estimation module is used to use a Hough transform algorithm to perform linear estimation on the static point cloud data to determine the target reflection boundary parameters before determining at least one matching set based on the target reflection boundary parameters and the cluster center information in the multi-cluster dynamic data.

[0151] Optionally, the linear estimation module includes:

[0152] A clustering unit, configured to cluster the static point cloud data using an improved DBSCAN clustering algorithm to obtain multiple clusters of static data;

[0153] a linear estimation unit, configured to perform linear estimation on each cluster of static data using a Hough transform algorithm to determine reflection boundary parameters corresponding to the static data;

[0154] The determining unit is configured to determine an invalid boundary parameter among a plurality of reflection boundary parameters, and obtain a target reflection boundary parameter based on the invalid boundary parameter.

[0155] Optionally, the determining unit is specifically configured to:

[0156] Determine the boundary length and slope parameters corresponding to each reflection boundary parameter;

[0157] An invalid boundary parameter among the plurality of reflection boundary parameters is determined based on each boundary length and each slope parameter.

[0158] Optionally, the parameters of the improved DBSCAN clustering algorithm include at least one of a clustering arc radius, a clustering arc angle, an angle deviation, and speed information.

[0159] Optionally, the determining module 330 is specifically configured to:

[0160] According to the target reflection boundary parameters and the cluster center information in the multi-cluster dynamic data, each cluster center in the multi-cluster dynamic data is matched to obtain at least one matching set.

[0161] Optionally, the identification module 340 is specifically configured to:

[0162] Traversing each cluster center in the matching set, determining a sum of target parameters corresponding to the cluster center, the target parameter including at least one of distance information, speed information, capability information, and the number of point clouds;

[0163] Whether the cluster center is a target cluster center is determined based on the sum of the target parameters.

[0164] The multipath identification device provided in the embodiments of the present application can execute the multipath identification method provided in any embodiment of the present application, and has the corresponding functional modules and beneficial effects of the execution method.

[0165] Example 4

[0166] FIG13 is a schematic diagram of the structure of a vehicle provided according to Example 4 of the present application. The vehicle is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The vehicle can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or required herein.

[0167] As shown in FIG13 , vehicle 10 includes at least one processor 11 and memory, such as a read-only memory (ROM) 12 and a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor, and processor 11 can perform various appropriate actions and processes based on the computer programs stored in ROM 12 or loaded from storage unit 18 into RAM 13. RAM 13 can also store various programs and data required for the operation of vehicle 10. Processor 11, ROM 12, and RAM 13 are interconnected via bus 14. An input / output (I / O) interface 15 is also connected to bus 14.

[0168] Various components in the vehicle 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the vehicle 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0169] The processor 11 may be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any other suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the multipath identification method.

[0170] In some embodiments, the multipath identification method can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on vehicle 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the multipath identification method described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the multipath identification method in any other suitable manner (e.g., via firmware).

[0171] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0172] Computer programs for implementing the methods of the present application may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0173] In the context of the present application, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. A computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0174] To provide interaction with a user, the systems and techniques described herein can be implemented in a vehicle having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the vehicle. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0175] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0176] A computing system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services.

[0177] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this application can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of this application can be achieved. This is not limited herein.

[0178] The above specific embodiments do not constitute a limitation on the scope of protection of this application. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application shall be included within the scope of protection of this application.

Claims

1. A multipath identification method, applied to a vehicle, wherein a millimeter wave radar is installed on the vehicle, and the method comprises: Preprocessing the echo signal of the millimeter wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal; The dynamic point cloud data is clustered using an improved DBSCAN clustering algorithm to obtain multiple clusters of dynamic data; Determine at least one matching set based on target reflection boundary parameters and cluster center information in the multi-cluster dynamic data, wherein the matching set includes at least two cluster centers, and the target reflection boundary parameters are determined according to the static point cloud data; For each matching set, the cluster center in the matching set is identified to obtain the target cluster center, and the target cluster center and the point cloud data in the cluster where the target cluster center is located are determined as the multipath target in the echo signal.

2. The method according to claim 1, wherein: The preprocessing of the echo signal of the millimeter wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal includes: Extracting point cloud data corresponding to the radar point cloud in the echo signal; The radar point cloud is separated into static and dynamic point cloud based on the point cloud data to obtain static point cloud data and dynamic point cloud data in the echo signal.

3. The method according to claim 2, wherein: The point cloud data includes speed information, and the step of separating the radar point cloud from static and dynamic point cloud based on the point cloud data to obtain static point cloud data and dynamic point cloud data in the echo signal includes: Determine a target radial velocity, where the target radial velocity is a radial velocity of the static point cloud relative to the millimeter wave radar; According to the radial velocity of the target and the velocity information of the radar point cloud, the radar point cloud is separated into static and dynamic point cloud data to obtain static point cloud data and dynamic point cloud data in the echo signal.

4. The method according to claim 1, wherein: Before determining at least one matching set based on the target reflection boundary parameter and the cluster center information in the multi-cluster dynamic data, the method further includes: The Hough transform algorithm is used to perform linear estimation on the static point cloud data to determine the target reflection boundary parameters.

5. The method according to claim 4, wherein: The method of using the Hough transform algorithm to linearly estimate the static point cloud data to determine the target reflection boundary parameters includes: The improved DBSCAN clustering algorithm is used to cluster the static point cloud data to obtain multiple clusters of static data; For each cluster of static data, a Hough transform algorithm is used to perform linear estimation on the static data to determine reflection boundary parameters corresponding to the static data; Invalid boundary parameters among the plurality of reflection boundary parameters are determined, and target reflection boundary parameters are obtained based on the invalid boundary parameters.

6. The method according to claim 5, wherein: The determining of the invalid boundary parameters among the plurality of reflection boundary parameters comprises: Determine the boundary length and slope parameter corresponding to each reflection boundary parameter; An invalid boundary parameter among a plurality of reflection boundary parameters is determined based on each boundary length and each slope parameter.

7. The method according to any one of claims 1 to 5, wherein: The parameters of the improved DBSCAN clustering algorithm include at least one of a clustering arc radius, a clustering arc angle, an angle deviation, and speed information.

8. The method according to claim 1, wherein: The determining at least one matching set based on the target reflection boundary parameter and the cluster center information in the multi-cluster dynamic data comprises: According to the target reflection boundary parameter and the cluster center information in the multi-cluster dynamic data, each cluster center in the multi-cluster dynamic data is matched to obtain at least one matching set.

9. The method according to claim 1, wherein: The identifying the cluster center in the matching set to obtain the target cluster center includes: Traversing each cluster center in the matching set, determining the sum of target parameters corresponding to the cluster center, wherein the target parameter includes at least one of distance information, speed information, capability information, and the number of point clouds; Whether the cluster center is a target cluster center is determined based on the sum of the target parameters.

10. A multipath identification device, configured on a vehicle, wherein a millimeter wave radar is installed on the vehicle, and the device comprises: A preprocessing module is configured to preprocess the echo signal of the millimeter wave radar to obtain static point cloud data and dynamic point cloud data in the echo signal; A clustering module is configured to cluster the dynamic point cloud data using an improved DBSCAN clustering algorithm to obtain multiple clusters of dynamic data; a determination module, configured to determine at least one matching set based on target reflection boundary parameters and cluster center information in the multi-cluster dynamic data, wherein the matching set includes at least two cluster centers, and the target reflection boundary parameters are determined according to the static point cloud data; The recognition module is configured to identify the cluster center in each matching set, obtain the target cluster center, and identify the target cluster center and the point cloud data in the cluster where the target cluster center is located. is determined as a multipath target in the echo signal.

11. A vehicle comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the multipath identification method according to any one of claims 1 to 9.

12. A computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a processor to implement the multipath identification method according to any one of claims 1 to 9 when executed.

Images