Radar failure judgment method, device, equipment, medium and vehicle

Abstract

The present disclosure relates to a radar failure judgment method, apparatus, device, medium, and vehicle, including: obtaining a single-frame radar point cloud collected by a radar; performing ground fitting on the single-frame radar point cloud to determine whether the ground fitting of the single-frame radar point cloud is successful; if the ground fitting of the single-frame radar point cloud fails, obtaining perception information of the vehicle's surrounding environment; based on the perception information, counting the number of obstacles contained in the range of the single-frame radar point cloud; if the number of obstacles is less than a first preset threshold, determining that the radar has failed. The present disclosure assists in the judgment of radar failure by introducing perception information of the vehicle's surrounding environment. Compared with failure judgment based on physical hardware, it can better reflect the degree of influence of the current point cloud quality on the perception algorithm, effectively identify radar failures caused by external factors such as dirty lidar visors, and improve the accuracy of the radar failure judgment method.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a radar failure determination method, apparatus, device, medium, and vehicle. Background Art

[0002] LiDAR (Light Detection and Ranging) is an important environmental perception device in the field of autonomous driving. Determining whether LiDAR has failed or not plays a crucial role in the safety of autonomous driving.

[0003] The existing method of determining whether a LiDAR has failed is based on diagnostic signals from the sensor's own hardware. This method determines whether the LiDAR has failed based on the sensor's own hardware. However, this method can be counterproductive. For example, if there is no problem with the LiDAR's hardware, but the point cloud output by the LiDAR is of poor quality and low precision due to factors such as a dirty LiDAR hood, rain, snow, or foggy weather, the LiDAR will still be deemed to have failed. Consequently, the corresponding safety measures will not be implemented, and subsequent point cloud data fusion and other tasks will proceed normally. However, in this case, not only will accurate data not be generated, but a large amount of processing resources will also be wasted. The point cloud generated by such a LiDAR will inevitably affect the effectiveness of autonomous driving and may even lead to serious problems such as driving safety. Summary of the Invention

[0004] In order to solve the above technical problems, the present disclosure provides a radar failure judgment method, device, equipment, medium and vehicle to improve the accuracy of radar failure judgment.

[0005] In a first aspect, an embodiment of the present disclosure provides a radar failure determination method, including:

[0006] Obtain a single-frame radar point cloud collected by radar;

[0007] Performing ground fitting on the single-frame radar point cloud, and determining whether the ground fitting of the single-frame radar point cloud is successful;

[0008] If the ground fitting of the single-frame radar point cloud fails, obtaining perception information of the vehicle's surrounding environment;

[0009] Based on the perception information, counting the number of obstacles contained in the range of a single-frame radar point cloud;

[0010] If the number of obstacles is less than a first preset threshold, it is determined that the radar has failed.

[0011] In some embodiments, performing ground fitting based on the single-frame radar point cloud and determining whether the ground fitting of the single-frame radar point cloud is successful includes:

[0012] Dividing the single-frame radar point cloud into a preset number of grid cells;

[0013] For each grid cell, performing plane fitting based on the point cloud within the grid cell;

[0014] If the proportion of grid cells with successful plane fitting to the preset number of grid cells is greater than a second preset threshold, it is determined that the ground fitting of the single-frame radar point cloud is successful; or

[0015] If the proportion of grid cells with successful plane fitting to the preset number of grid cells is less than or equal to a second preset threshold, it is determined that the ground fitting of the single-frame radar point cloud has failed.

[0016] In some embodiments, performing plane fitting based on the point cloud within the grid unit includes:

[0017] If the average distance between each point cloud in the grid unit and the plane obtained by plane fitting is less than a third preset threshold, it is determined that the plane fitting of the grid unit is successful.

[0018] In some embodiments, if the average distance between each point cloud in the grid unit and the plane obtained by plane fitting is less than a third preset threshold, determining that the grid unit plane fitting is successful includes:

[0019] If the average distance between each point cloud in the grid unit and the plane obtained by plane fitting is less than a third preset threshold, the number of point clouds in the grid unit is greater than or equal to a fourth preset threshold, and the number of noise points in the grid unit is less than or equal to a fifth preset threshold, then it is determined that the grid unit plane fitting is successful.

[0020] In some embodiments, the method further comprises:

[0021] If the ground fitting of the single-frame radar point cloud is successful, then the proportion of low-intensity points in the single-frame radar point cloud is counted;

[0022] If the proportion of low-intensity points in the single-frame radar point cloud is greater than a sixth preset threshold, it is determined that the radar has failed; or,

[0023] If the proportion of low-intensity points in the single-frame radar point cloud is less than or equal to a sixth preset threshold, it is determined that the radar is normal.

[0024] In some embodiments, the method further comprises:

[0025] If the number of obstacles is greater than or equal to a first preset threshold, obtaining a radar state judgment result corresponding to the historical radar point cloud of the single-frame radar point cloud;

[0026] The radar state judgment result of the historical radar point cloud is used as the radar state judgment result of the single-frame radar point cloud.

[0027] In a second aspect, an embodiment of the present disclosure provides a radar failure determination device, including:

[0028] A first acquisition module is used to acquire a single-frame radar point cloud collected by the radar;

[0029] a ground fitting module, configured to perform ground fitting on the single-frame radar point cloud and determine whether the ground fitting of the single-frame radar point cloud is successful;

[0030] The second acquisition module is used to obtain perception information of the vehicle's surrounding environment when the ground fitting of a single-frame radar point cloud fails;

[0031] A first statistical module is configured to count the number of obstacles contained in a single-frame radar point cloud range based on the perception information;

[0032] The first determining module is configured to determine that the radar fails when the number of obstacles is less than a first preset threshold.

[0033] In a third aspect, an embodiment of the present disclosure provides an electronic device, including:

[0034] Memory;

[0035] processor; and

[0036] computer programs;

[0037] The computer program is stored in the memory and is configured to be executed by the processor to implement the method as described in the first aspect.

[0038] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the method described in the first aspect.

[0039] In a fifth aspect, an embodiment of the present disclosure provides a vehicle comprising the device, electronic device or computer-readable storage medium as described above.

[0040] The radar failure judgment method, apparatus, device, medium, and vehicle provided by the embodiments of the present disclosure assist in the judgment of radar failure by introducing perception information of the vehicle's surrounding environment, thereby avoiding misjudgment of the ground fitting quality of a single-frame radar point cloud when there are many obstacles around the vehicle. Compared with failure judgment based on physical hardware, it can better reflect the impact of the current point cloud quality on the perception algorithm, effectively identify radar failures caused by external factors such as dirty lidar hood, and improve the accuracy of the radar failure judgment method. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0042] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0043] Figure 1 A flow chart of a radar failure determination method provided in an embodiment of the present disclosure;

[0044] Figure 2 A flow chart of a method for determining ground fitting results of a single-frame radar point cloud provided by another embodiment of the present disclosure;

[0045] Figure 3 A flow chart of a radar failure determination method provided by another embodiment of the present disclosure;

[0046] Figure 4 A schematic structural diagram of a radar failure determination device provided in an embodiment of the present disclosure;

[0047] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0048] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features therein can be combined with each other in the absence of conflict.

[0049] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0050] The embodiments of the present disclosure provide a radar failure determination method, which is described below in conjunction with specific embodiments.

[0051] Figure 1 This is a flow chart of the radar failure determination method provided in an embodiment of the present disclosure. This method can be applied to in-vehicle devices, including vehicle computers, smartphones, PDAs, tablet computers, laptops, all-in-one computers, and intelligent driving devices. It is understood that the radar failure determination method provided in an embodiment of the present disclosure can also be applied in other scenarios.

[0052] Below Figure 1 The radar failure judgment method shown in FIG. 1 is introduced, and the specific steps of the method are as follows:

[0053] S101: Acquire a single-frame radar point cloud collected by radar.

[0054] Point cloud data can be collected by on-board equipment using a LiDAR device installed on the vehicle. When a laser beam emitted by the LiDAR hits the surface of an object, the laser reflected by the object will reveal information such as the object's position, distance, and reflectivity. If the laser beam is scanned along a certain trajectory, a large number of laser points can be obtained, forming radar point cloud data. While the vehicle is driving, the LiDAR device installed on the vehicle continuously collects point cloud data of objects around the vehicle at a certain frequency. The point cloud data obtained from a single acquisition is called a single-frame radar point cloud.

[0055] S102: Perform ground fitting on the single-frame radar point cloud, and determine whether the ground fitting of the single-frame radar point cloud is successful.

[0056] A single-frame radar point cloud contains multiple ground points, corresponding to laser reflection points on the road surface in a real environment. The point cloud information obtained from a single-frame radar point cloud is often dense. Before ground fitting, the ground points must be extracted from the single-frame radar point cloud.

[0057] Optionally, ground point extraction can be performed based on point clouds below a preset frame height in a single-frame radar point cloud. Since ground points generally appear in the lower half of a single-frame radar point cloud, extracting ground points based on point clouds below a preset frame height can effectively reduce the computational complexity of ground point extraction.

[0058] In one possible implementation, the ground fitting process can include selecting three ground points from a single-frame radar point cloud, calculating a sampling plane formed by these three points, and then iteratively fitting the remaining ground points to the sampling plane until a stopping condition is met, thereby obtaining the final fitted target plane. The stopping condition includes, but is not limited to, a maximum number of iterations.

[0059] Furthermore, the quality of the target plane is evaluated. If the quality of the target plane is poor, it is determined that the ground fitting of the single-frame radar point cloud has failed; if the quality of the target plane is good, it is determined that the ground fitting of the single-frame radar point cloud has succeeded.

[0060] S103: If the ground fitting of the single-frame radar point cloud fails, acquiring perception information of the vehicle's surrounding environment.

[0061] The vehicle is equipped with sensors for collecting information about the vehicle's surrounding environment, such as cameras, visual sensors, etc., of course, not limited to these. During the vehicle's driving process, the on-board equipment obtains real-time information about the vehicle's surrounding environment through the above sensors.

[0062] In some embodiments, visual sensor information such as images may be acquired through a visual sensor, and the visual sensor information may be used as perception information of the vehicle's surrounding environment;

[0063] In some embodiments, the single-frame radar point cloud obtained in the above steps can also be directly used as the perception information of the vehicle's surrounding environment;

[0064] In some embodiments, visual sensing information can also be obtained through a visual sensor, and the visual sensing information is fused with the single-frame radar point cloud obtained in the above steps to obtain fused data, and the fused data is used as perception information of the vehicle's surrounding environment.

[0065] Environmental information can be used to identify various types of obstacle information around the vehicle, where obstacle information includes but is not limited to: the number, location, size, distance, etc. of obstacles.

[0066] When ground fitting of a single-frame radar point cloud fails, the system obtains sensory information about the vehicle's surroundings to eliminate poor point cloud data acquisition quality caused by the vehicle's surroundings. Specifically, the sensory information about the vehicle's surroundings can include images or videos of the vehicle's surroundings captured by a camera mounted on the vehicle.

[0067] S104: Based on the perception information, count the number of obstacles contained in the range of a single-frame radar point cloud.

[0068] The vehicle-mounted device analyzes the collected perception information and identifies multiple obstacles around the vehicle. Specifically, the perception information can be input into a pre-trained obstacle recognition model to obtain obstacle information in the perception information.

[0069] In some embodiments, visual sensing information can be input into a pre-trained obstacle recognition model to obtain obstacle information, and then the number of obstacles contained in the range of a single-frame radar point cloud can be counted;

[0070] In some embodiments, a single-frame radar point cloud may be input into a pre-trained obstacle recognition model to obtain obstacle information therein, and then the number of obstacles contained in the range of the single-frame radar point cloud may be counted.

[0071] In some embodiments, the fused data may be input into a pre-trained obstacle recognition model to obtain obstacle information therein, and then the number of obstacles contained in the range of a single-frame radar point cloud may be counted.

[0072] According to the above different perception information, different obstacle recognition models can be set to identify obstacle information.

[0073] In some embodiments, if the information about the obstacle cannot be obtained through the perception information, or the number of obstacles cannot be counted, the radar is determined to be failed, and a corresponding warning message is output to inform the user.

[0074] Since obstacle information can reflect the orientation, distance and other information of an object, the obstacle recognition results based on the perception information can count the number of obstacles contained in the range of a single-frame radar point cloud.

[0075] S105: If the number of obstacles is less than a first preset threshold, determine that the radar is inoperative.

[0076] When the number of obstacles within a single-frame radar point cloud exceeds a first preset threshold, the vehicle is considered to be surrounded by numerous obstacles, obstructing a large area of the ground around it. This can result in a small number of ground points in the single-frame radar point cloud, leading to poor ground fitting quality or even ground fitting failure, even though the radar itself may not be faulty. The first preset threshold can be set in advance by the user.

[0077] In some embodiments, the first preset threshold may be an average value of historical obstacle threshold experience values for successfully determining whether the radar is faulty; or,

[0078] When using a single-frame radar point cloud or fused data as perception information and the number of obstacles contained in the range of the single-frame radar point cloud can be obtained, the first preset threshold can be set in advance by the user, or the number of obstacles contained in the range of the single-frame radar point cloud can be obtained by inputting visual sensing information into a pre-trained obstacle recognition model.

[0079] For example, when a vehicle is traveling on a road with heavy traffic, other vehicles around it will obstruct a large area of the ground. Even if the radar is operating normally, the ground fitting of the single-frame radar point cloud collected will fail due to the scarcity of ground points in the collected single-frame radar point cloud. Therefore, when there are many obstacles around the vehicle, even if the ground fitting of the single-frame radar point cloud fails, it cannot be directly concluded that the radar has failed. Conversely, if the number of obstacles around the vehicle is less than a first preset threshold, it is considered that there are few obstacles around the vehicle. In this case, the poor ground fitting quality caused by the scarcity of ground points can be ruled out. The main cause of the ground fitting failure in this case is the poor quality of the collected point cloud itself, such as bad weather, a dirty lidar optics, or even a problem with the radar hardware itself. Therefore, a radar failure judgment needs to be made and the vehicle management system needs to be notified in a timely manner.

[0080] The disclosed embodiment obtains a single-frame radar point cloud collected by a radar; performs ground fitting on the single-frame radar point cloud and determines whether the ground fitting of the single-frame radar point cloud is successful; if the ground fitting of the single-frame radar point cloud fails, obtains perception information of the vehicle's surrounding environment; based on the perception information, counts the number of obstacles contained in the single-frame radar point cloud range; if the number of obstacles is less than a first preset threshold, determines that the radar has failed. By introducing the perception information of the vehicle's surrounding environment to assist in the judgment of radar failure, it avoids misjudgment of the ground fitting quality of the single-frame radar point cloud when there are many obstacles around the vehicle. Compared with failure judgment based on physical hardware, it can better reflect the degree of influence of the current point cloud quality on the perception algorithm, effectively identify radar failure caused by external factors such as dirty lidar mask, and improve the accuracy of the radar failure judgment method.

[0081] Figure 2 This is a flow chart of a method for determining the ground fitting result of a single-frame radar point cloud provided by another embodiment of the present disclosure. In some embodiments, the above step S102 can be specifically performed as follows: Figure 2 The process shown is implemented. Figure 2 As shown in Figure 1, the method for judging the ground fitting results of a single-frame radar point cloud includes the following steps:

[0082] S201: Acquire a single-frame radar point cloud collected by radar.

[0083] S202: Divide the single-frame radar point cloud into a preset number of grid units.

[0084] S203 : For each grid unit, perform plane fitting based on the point cloud within the grid unit.

[0085] A single-frame radar point cloud contains a large number of points, and directly performing ground fitting on the entire frame would result in an excessively large amount of computation. Therefore, dividing the single-frame radar point cloud into a preset number of grid cells and performing plane fitting on the point cloud within each grid cell can effectively reduce the complexity of plane fitting. Optionally, a region of interest can be identified in the single-frame radar point cloud, and the radar point cloud within the region of interest can be divided into several grid cells, with plane fitting performed on the point cloud within each grid cell.

[0086] For example, the region of interest (ROI) is defined as 40 meters in front of the LiDAR and 4 meters to the left and right. Within the ROI, the region is divided into 16 grid cells, each 10 meters long and 2 meters wide. A plane is fitted to the ground points within each grid cell, and the corresponding plane parametric equation is calculated, such as Ax + By + Cz + D = 0, to represent the fitted plane.

[0087] S204: Determine whether the ratio of the grid cells with successfully plane fitting to the preset number of grid cells is greater than a second preset threshold. If so, execute S205; if not, execute S206.

[0088] During the plane fitting process, since the point cloud quality within each grid cell varies, the quality of the plane fitted to each grid cell varies. If the plane fitting quality is poor, the plane fitting for that grid cell is considered a failure; if the plane fitting quality is good, the plane fitting for that grid cell is considered a success. The second preset threshold can be manually set in advance. The following describes two possible implementations for determining the success or failure of a grid cell plane fitting.

[0089] In some embodiments, if the average distance between each point cloud in the grid unit and the plane obtained by plane fitting is less than a third preset threshold, it is determined that the plane fitting of the grid unit is successful.

[0090] As described above, after plane fitting is performed on the grid unit, the plane parameters corresponding to the grid unit will be obtained. Combined with the position of each point in the grid unit in space, the distance between each point in the grid unit and the fitted plane can be calculated, and the average of multiple distances is calculated to obtain the average distance between each point cloud in the grid unit and the plane obtained by plane fitting. If the average distance is less than the third preset threshold, it is considered that most of the point clouds in the grid unit are close to the plane, and the plane is closer to the ground in the real environment. It is considered that the plane fitting quality of the grid unit is good, and the plane fitting of the grid unit is determined to be successful. Among them, the third preset threshold can be set artificially in advance. On the contrary, if the average distance is greater than or equal to the third preset threshold, it may be due to errors or omissions in the point cloud data collected by the radar, resulting in poor quality of plane fitting. In this case, it is determined that the plane fitting of the grid unit has failed.

[0091] In some embodiments, determining whether the grid unit plane is successfully fitted may also be based on the number of point clouds in the grid unit and the number of noise points in the grid unit.

[0092] If the average distance between each point cloud in the grid unit and the plane obtained by plane fitting is less than a third preset threshold, the number of point clouds in the grid unit is greater than or equal to a fourth preset threshold, and the number of noise points in the grid unit is less than or equal to a fifth preset threshold, then it is determined that the grid unit plane fitting is successful.

[0093] If the number of point clouds in the grid unit is less than a fourth preset threshold, or the number of noise points in the grid unit is greater than a fifth preset threshold, it is determined that the plane fitting of the grid unit fails.

[0094] If the number of point clouds within the grid cell is less than the fourth preset threshold, plane fitting cannot be performed and a plane fitting failure judgment is given. If the number of point clouds within the grid cell is greater than or equal to the fourth preset threshold, it is necessary to further determine whether the grid cell plane fitting is successful based on the plane fitting quality.

[0095] Affected by the characteristics of the instrument, the surrounding environment, and the scanned target itself, some noise points are inevitable in the point cloud data. Noise points mainly include drift points, isolated points, redundant points, etc. Drift points are sparse, scattered points that are far away from the target body and float above the point cloud; isolated points are small and dense points that are far away from the center of the point cloud; redundant points are redundant points that exceed the predetermined scanning area. When the number of noise points is too large and greater than the fifth preset threshold, it will have a serious impact on the plane fitting process within the grid unit, and may even cause the fitting to fail. At this time, a judgment of plane fitting failure will be given. If the number of noise points in the grid is less than or equal to the fifth preset threshold, it is necessary to further judge whether the grid unit plane fitting is successful based on the plane fitting quality. Among them, the fourth preset threshold and the fifth preset threshold can be set manually in advance.

[0096] S205: Determine whether the ground fitting of the single-frame radar point cloud is successful.

[0097] S206: Determine that the ground fitting of the single-frame radar point cloud fails.

[0098] The disclosed embodiment divides a single-frame radar point cloud into grid cells and comprehensively judges whether the ground fitting of the single-frame radar point cloud is successful based on the plane fitting results of each grid cell. This greatly reduces the amount of data and calculations in the fitting process, effectively saving memory and computing resources. It also improves the robustness of the judgment of the ground fitting results of the single-frame radar point cloud.

[0099] Based on the above embodiment, if the ground fitting of the single-frame radar point cloud is successful, the proportion of low-intensity points in the single-frame radar point cloud is counted; if the proportion of low-intensity points in the single-frame radar point cloud is greater than a sixth preset threshold, it is determined that the radar has failed; or, if the proportion of low-intensity points in the single-frame radar point cloud is less than or equal to the sixth preset threshold, it is determined that the radar is normal.

[0100] In practice, slippery surfaces can also affect the quality of the collected point cloud, such as waterlogging caused by inclement weather like rain and snow. This situation can be identified by counting the proportion of low-intensity points in a single-frame radar point cloud. The point cloud intensity value range is 0-255, and the point cloud intensity threshold can be determined in advance based on empirical values. Points with intensity values below the intensity threshold are considered low-intensity points. In rainy weather, slippery roads, and other conditions, point cloud imaging is poor, and the proportion of collected point cloud points with low intensity values is high. On sunny days, when conditions are better, if the ground fitting of a single-frame radar point cloud is successful, the proportion of low-intensity points in the single-frame radar point cloud is further counted. If the proportion of low-intensity points in the single-frame radar point cloud is high and greater than a sixth preset threshold, it is determined that the ground point cloud imaging effect is poor under the current circumstances, and a radar failure determination result needs to be promptly issued. If the proportion of low-intensity points in the single-frame radar point cloud is low and less than or equal to the sixth preset threshold, the ground point cloud imaging effect under the current circumstances is considered to be within the normal range, and a radar normal determination result can be issued based on the successful ground fitting of the single-frame radar point cloud. The sixth preset threshold may be set manually in advance.

[0101] The disclosed embodiment further counts the proportion of low-intensity points in a single-frame radar point cloud based on the successful ground fitting of the single-frame radar point cloud, thereby achieving more accurate detection of the point cloud imaging quality, and giving a radar failure judgment result when the point cloud imaging quality is poor, further improving the accuracy of the radar failure judgment method.

[0102] Based on the above embodiment, if the number of obstacles is greater than or equal to the first preset threshold, the radar state judgment result corresponding to the historical radar point cloud of the single-frame radar point cloud is obtained; and the radar state judgment result of the historical radar point cloud is used as the radar state judgment result of the single-frame radar point cloud.

[0103] As described above, when the number of obstacles within a single-frame radar point cloud is greater than or equal to a first preset threshold, the vehicle is considered to be surrounded by numerous obstacles, obstructing a large area of the ground surrounding it. This can result in a small number of ground points in the single-frame radar point cloud, leading to poor ground fitting quality or even ground fitting failure, even though the radar itself may not be faulty. In this case, it is impossible to determine the radar status based on the ground fitting quality of the single-frame radar point cloud. Therefore, the radar status determination results of the historical radar point cloud are used as the radar status determination results for the single-frame radar point cloud. This avoids misjudgments of the radar status corresponding to the single-frame radar point cloud and further improves the accuracy of the radar failure determination method.

[0104] Figure 3 This is a flow chart of a radar failure determination method provided by another embodiment of the present disclosure. Figure 3 As shown, the method includes the following steps:

[0105] S301: Acquire a single-frame radar point cloud collected by radar.

[0106] S302: Perform ground fitting on the single-frame radar point cloud to determine whether the ground fitting is successful. If so, execute S305; if not, execute S303.

[0107] Specifically, the single-frame radar point cloud is divided into a preset number of grid units; for each grid unit, plane fitting is performed based on the point cloud within the grid unit; it is determined whether the proportion of grids with successful plane fitting to the preset number of grid units is greater than a second preset threshold; if so, it is determined that the ground fitting of the single-frame radar point cloud is successful; if not, it is determined that the ground fitting of the single-frame radar point cloud is successful.

[0108] S303: Acquire perception information of the vehicle's surrounding environment, and based on the perception information, count the number of obstacles contained in a single-frame radar point cloud range.

[0109] S304: Determine whether the number of obstacles is less than a first preset threshold. If so, execute S308; if not, execute S309.

[0110] S305: Count the proportion of low-intensity points in the single-frame radar point cloud.

[0111] S306: Determine whether the proportion of the low-intensity points is greater than a sixth preset threshold. If so, execute S308; if not, execute S307.

[0112] S307: Determine whether the radar is normal.

[0113] S308: Determine that the radar fails.

[0114] S309: Using the radar state judgment result corresponding to the historical radar point cloud of the single-frame radar point cloud as the radar state judgment result of the single-frame radar point cloud.

[0115] The disclosed embodiment assists in determining radar failure by introducing perceptual information about the vehicle's surrounding environment, thereby avoiding misjudgments of the ground fitting quality of a single-frame radar point cloud when there are many obstacles around the vehicle. Compared to failure determination based on physical hardware, it is more able to reflect the degree of influence of the current point cloud quality on the perception algorithm, and effectively identifies radar failures caused by external factors such as a dirty lidar visor, thereby improving the accuracy of the radar failure determination method. Furthermore, by dividing the single-frame radar point cloud into grid cells, the success of the ground fitting of the single-frame radar point cloud is comprehensively determined based on the plane fitting results of each grid cell. This greatly reduces the amount of data and computation required for the fitting process, effectively saving memory and computing resources, and also improves the robustness of the ground fitting result judgment of the single-frame radar point cloud.

[0116] Figure 4 This is a schematic diagram of the structure of the radar failure judgment device provided in the embodiment of the present disclosure. The radar failure judgment device can be the vehicle-mounted device described in the above embodiment, or the radar failure judgment device can be a component or assembly in the vehicle-mounted device. The radar failure judgment device provided in the embodiment of the present disclosure can execute the processing flow provided in the embodiment of the radar failure judgment method, such as Figure 4 As shown, the radar failure judgment device 40 includes: a first acquisition module 41, a ground fitting module 42, a second acquisition module 43, a first statistics module 44, and a first determination module 45; wherein the first acquisition module 41 is used to obtain a single-frame radar point cloud collected by the radar; the ground fitting module 42 is used to perform ground fitting on the single-frame radar point cloud and determine whether the ground fitting of the single-frame radar point cloud is successful; the second acquisition module 43 is used to obtain perception information of the vehicle's surrounding environment when the ground fitting of the single-frame radar point cloud fails; the first statistics module 44 is used to count the number of obstacles contained in the range of the single-frame radar point cloud based on the perception information; and the first determination module 45 is used to determine that the radar has failed when the number of obstacles is less than a first preset threshold.

[0117] Optionally, the ground fitting module 42 includes a grid division unit 421, a plane fitting unit 422, a first determination unit 423, and a second determination unit 424; the grid division unit 421 is used to divide the single-frame radar point cloud into a preset number of grid units; the plane fitting unit 422 is used to perform plane fitting for each grid unit based on the point cloud within the grid unit; the first determination unit 423 is used to determine that the ground fitting of the single-frame radar point cloud is successful when the proportion of grids with successful plane fitting to the preset number of grid units is greater than a second preset threshold; the second determination unit 424 is used to determine that the ground fitting of the single-frame radar point cloud fails when the proportion of grids with successful plane fitting to the preset number of grid units is less than or equal to the second preset threshold.

[0118] Optionally, the first determining unit 423 is configured to determine that the plane fitting of the grid unit is successful when an average distance between each point cloud in the grid unit and a plane obtained by plane fitting is less than a third preset threshold.

[0119] Optionally, the second determining unit 424 is configured to determine that the plane fitting of the grid cell fails when the number of point clouds in the grid cell is less than a fourth preset threshold, or the number of noise points in the grid cell is greater than a fifth preset threshold.

[0120] Optionally, the radar failure judgment device 40 also includes a second statistical module 46 and a second determination module 47; the second statistical module 46 is used to count the proportion of low-intensity points in the single-frame radar point cloud when the ground fitting of the single-frame radar point cloud is successful; the second determination module 47 is used to determine that the radar has failed when the proportion of low-intensity points in the single-frame radar point cloud is greater than a sixth preset threshold; or to determine that the radar is normal when the proportion of low-intensity points in the single-frame radar point cloud is less than or equal to the sixth preset threshold.

[0121] Optionally, the radar failure judgment device 40 also includes an acquisition module 48, which is used to obtain the radar status judgment result corresponding to the historical radar point cloud of the single-frame radar point cloud when the number of obstacles is greater than or equal to a first preset threshold, and use the radar status judgment result of the historical radar point cloud as the radar status judgment result of the single-frame radar point cloud.

[0122] Figure 4 The radar failure judgment device of the illustrated embodiment can be used to implement the technical solution of the above-mentioned method embodiment. Its implementation principle and technical effects are similar and will not be repeated here.

[0123] In addition, an embodiment of the present disclosure also provides a vehicle, which includes the radar failure judgment device as described in the above embodiment.

[0124] Figure 5 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present disclosure. The electronic device may be the vehicle-mounted device described in the above embodiment. The electronic device provided in the embodiment of the present disclosure may execute the processing flow provided in the embodiment of the radar failure judgment method, such as Figure 5 As shown, the electronic device 50 includes: a memory 51, a processor 52, a computer program and a communication interface 53; wherein the computer program is stored in the memory 51 and is configured so that the processor 52 executes the radar failure judgment method as described above.

[0125] In addition, an embodiment of the present disclosure further provides a computer-readable storage medium on which a computer program is stored. The computer program is executed by a processor to implement the radar failure judgment method described in the above embodiment.

[0126] In addition, an embodiment of the present disclosure further provides a computer program product, which includes a computer program or instructions, and when the computer program or instructions are executed by a processor, the radar failure judgment method described above is implemented.

[0127] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including, but not limited to, object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0128] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of the boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0129] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0130] The foregoing description is intended only to provide specific embodiments of the present disclosure, intended to enable those skilled in the art to understand and implement the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the embodiments described herein, but rather to be construed in the broadest manner consistent with the principles and novel features disclosed herein.

Claims

1. A radar failure judgment method, characterized in that: The method comprises: Obtain a single-frame radar point cloud collected by radar; Performing ground fitting on the single-frame radar point cloud, and determining whether the ground fitting of the single-frame radar point cloud is successful; If the ground fitting of the single-frame radar point cloud fails, obtaining perception information of the vehicle's surrounding environment; Based on the perception information, counting the number of obstacles contained in the range of a single-frame radar point cloud; If the number of obstacles is less than a first preset threshold, it is determined that the radar has failed.

2. The method according to claim 1, characterized in that The performing ground fitting based on the single-frame radar point cloud and judging whether the ground fitting of the single-frame radar point cloud is successful includes: Dividing the single-frame radar point cloud into a preset number of grid cells; For each grid cell, performing plane fitting based on the point cloud within the grid cell; If the proportion of grid cells with successful plane fitting to the preset number of grid cells is greater than a second preset threshold, it is determined that the ground fitting of the single-frame radar point cloud is successful; or If the proportion of grid cells with successful plane fitting to the preset number of grid cells is less than or equal to a second preset threshold, it is determined that the ground fitting of the single-frame radar point cloud has failed.

3. The method according to claim 2, characterized in that The performing plane fitting based on the point cloud in the grid unit includes: If the average distance between each point cloud in the grid unit and the plane obtained by plane fitting is less than a third preset threshold, it is determined that the plane fitting of the grid unit is successful.

4. The method according to claim 3, characterized in that If the average distance between each point cloud in the grid unit and the plane obtained by plane fitting is less than a third preset threshold, determining that the grid unit plane fitting is successful includes: If the average distance between each point cloud in the grid unit and the plane obtained by plane fitting is less than a third preset threshold, the number of point clouds in the grid unit is greater than or equal to a fourth preset threshold, and the number of noise points in the grid unit is less than or equal to a fifth preset threshold, then it is determined that the grid unit plane fitting is successful.

5. The method according to claim 1, wherein The method further comprises: If the ground fitting of the single-frame radar point cloud is successful, then the proportion of low-intensity points in the single-frame radar point cloud is counted; If the proportion of low-intensity points in the single-frame radar point cloud is greater than a sixth preset threshold, it is determined that the radar has failed; or, If the proportion of low-intensity points in the single-frame radar point cloud is less than or equal to a sixth preset threshold, it is determined that the radar is normal.

6. The method according to claim 1, characterized in that The method further comprises: If the number of obstacles is greater than or equal to a first preset threshold, obtaining a radar state judgment result corresponding to the historical radar point cloud of the single-frame radar point cloud; The radar state judgment result of the historical radar point cloud is used as the radar state judgment result of the single-frame radar point cloud.

7. A radar failure judgment device, characterized in that: The device comprises: A first acquisition module is used to acquire a single-frame radar point cloud collected by the radar; a ground fitting module, configured to perform ground fitting on the single-frame radar point cloud and determine whether the ground fitting of the single-frame radar point cloud is successful; The second acquisition module is used to obtain perception information of the vehicle's surrounding environment when the ground fitting of a single-frame radar point cloud fails; A first statistical module is configured to count the number of obstacles contained in a single-frame radar point cloud range based on the perception information; The first determining module is configured to determine that the radar fails when the number of obstacles is less than a first preset threshold.

8. An electronic device, characterized in that: include: Memory; processor; as well as computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A vehicle comprising: The radar failure determination device according to claim 7; or the electronic device according to claim 8; Or, the computer-readable storage medium of claim 9.

Images