Method and System for Estimating Driving Obstacle State Based on Millimeter-Wave Radar

Abstract

The present invention relates to the technical field of target detection, and provides a method and system for estimating the state of driving obstacles based on a millimeter-wave radar. The method includes obtaining point cloud data based on the millimeter-wave radar; clustering the point cloud data to obtain at least one driving obstacle point cloud set; determining a state estimation result of the driving obstacle according to the driving obstacle point cloud set; the state estimation result includes a bounding box of the driving obstacle; the bounding box is a rotated box with the minimum fitting cost; the rotated box is obtained by rotating a reference box; the reference box is a geometric figure of a set type, and the reference box is the minimum circumscribed box of the driving obstacle point cloud set at a reference angle; the fitting cost is calculated according to the driving obstacle point cloud set and the rotated box. The present invention can make the bounding box have a higher information density, and at the same time can reflect the angular orientation of the obstacle, thereby providing a basis for reducing the computational resource requirements in the subsequent autonomous driving decision-making process.

Description

Technical Field

[0001] The present invention relates to the technical field of target detection, and particularly to a method and system for estimating the state of driving obstacles based on a millimeter-wave radar. Background Art

[0002] Autopilot, also known as driverless, computer driving or wheeled mobile robot, is a cutting-edge technology that relies on computer and artificial intelligence technologies to complete complete, safe and effective driving without human operation.

[0003] In the 21st century, due to the continuous increase in car users, problems such as traffic congestion and safety accidents faced by road traffic have become increasingly serious. With the support of vehicle networking technology and artificial intelligence technology, autopilot technology can coordinate travel routes and planned times, thereby greatly improving travel efficiency and reducing energy consumption to a certain extent. Autopilot can also help avoid safety hazards such as drunk driving and fatigue driving, reduce driver errors, and enhance safety. Autopilot has thus become a research and development focus in various countries in recent years.

[0004] As an automated vehicle, an autonomous vehicle can sense its environment and navigate without human operation. As a feasible hardware for perceiving the autopilot environment, an in-vehicle millimeter-wave radar can collect the point cloud data of obstacles during driving. Further, based on the point cloud data, the state of the obstacles can be analyzed.

[0005] In the prior art, the information of obstacles collected by an in-vehicle millimeter-wave radar is usually represented by a bounding box with speed, direction and other information. Since the millimeter-wave radar point cloud is relatively sparser than other sensors and has limited penetration, the prior art often directly uses the circumscribed rectangle of the radar point cloud as the bounding box.

[0006] However, it is found in practice that the information density of the bounding box obtained by using the above prior art method is relatively low, and it cannot accurately reflect the obstacle information (such as the orientation of the obstacle), which in turn leads to a sharp increase in the computing resources required for the subsequent driving decision-making process based on the obstacle information, and is not conducive to the realization of high-performance autopilot.

[0007] Therefore, how to more accurately estimate the state of obstacles has become a technical problem that needs to be solved urgently in the industry. Summary of the Invention

[0008] The present invention provides a method and system for estimating the state of driving obstacles based on a millimeter-wave radar, so as to solve the defect of relatively low information density of the bounding box in the prior art and realize a more accurate estimation of the obstacle state.

[0009] The present invention provides a method for estimating the state of driving obstacles based on a millimeter-wave radar, including:

[0010] Obtaining point cloud data based on the millimeter-wave radar;

[0011] Clustering the point cloud data to obtain at least one driving obstacle point cloud set;

[0012] Determining the state estimation result of the driving obstacle according to the driving obstacle point cloud set;

[0013] The state estimation result includes the bounding box of the driving obstacle; the bounding box is a rotated box with the minimum fitting cost; the rotated box is obtained by rotating a reference box; the reference box is a geometric figure of a set type, and the reference box is the minimum circumscribed box of the driving obstacle point cloud set at a reference angle; the fitting cost is calculated according to the driving obstacle point cloud set and the rotated box.

[0014] According to the method for estimating the state of driving obstacles based on a millimeter-wave radar provided by the present invention, the rotated box is the minimum circumscribed box of the driving obstacle point cloud set after rotating the reference box.

[0015] According to the method for estimating the state of driving obstacles based on a millimeter-wave radar provided by the present invention, the rotated box is the minimum circumscribed box of the driving obstacle point cloud set after excluding outliers after rotating the reference box;

[0016] The outlier is a point whose fitting cost difference of the driving obstacle point cloud set before and after exclusion is greater than a set threshold.

[0017] According to the method for estimating the state of driving obstacles based on a millimeter-wave radar provided by the present invention, the fitting cost refers to the sum of the distances from the points in the driving obstacle point cloud set to the rotated box;

[0018] The distance is the distance from the point to the straight line where the nearest side of the rotated box is located.

[0019] According to the method for estimating the state of driving obstacles based on a millimeter-wave radar provided by the present invention, the bounding box is the rotated box with the minimum fitting cost in the set of rotated boxes;

[0020] The set of rotated boxes includes rotated boxes obtained by rotating a rectangular reference box at set angles within a rotation angle range; the rotation angle range refers to 0 degrees to 180 degrees, or -90 degrees to 90 degrees.

[0021] According to the method for estimating the state of driving obstacles based on a millimeter-wave radar provided by the present invention, the step of clustering the point cloud data to obtain at least one driving obstacle point cloud set includes:

[0022] Cluster the point cloud data through DBSCAN, and obtain at least one point cloud cluster based on density as the driving obstacle point cloud set.

[0023] The present invention also provides a driving obstacle state estimation system based on a millimeter-wave radar, including:

[0024] An acquisition module for acquiring point cloud data based on the millimeter-wave radar;

[0025] A clustering module for clustering the point cloud data to obtain at least one driving obstacle point cloud set;

[0026] An estimation module for determining the state estimation result of the driving obstacle according to the driving obstacle point cloud set;

[0027] The state estimation result includes the bounding box of the driving obstacle; the bounding box is a rotated box with the minimum fitting cost; the rotated box is obtained by rotating a reference box; the reference box is a geometric figure of a set type, and the reference box is the minimum circumscribed box of the driving obstacle point cloud set at a reference angle; the fitting cost is calculated according to the driving obstacle point cloud set and the rotated box.

[0028] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of any one of the above-mentioned driving obstacle state estimation methods based on a millimeter-wave radar are implemented.

[0029] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of any one of the above-mentioned driving obstacle state estimation methods based on a millimeter-wave radar are implemented.

[0030] The present invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, the steps of any one of the above-mentioned driving obstacle state estimation methods based on a millimeter-wave radar are implemented.

[0031] The driving obstacle state estimation method and system based on a millimeter-wave radar provided by the present invention introduce a fitting cost, select a bounding box at a suitable angle, so that the bounding box has a higher information density, and at the same time can reflect the angular orientation of the obstacle, thereby providing a basis for reducing the computational resource requirements in the subsequent autonomous driving decision-making process. Description of the Drawings

[0032] To more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the accompanying drawings required for use in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0033] Figure 1 It is a schematic flowchart of a method for estimating the state of driving obstacles based on a millimeter-wave radar provided by the present invention;

[0034] Figure 2 It is a schematic structural diagram of a system for estimating the state of driving obstacles based on a millimeter-wave radar provided by the present invention;

[0035] Figure 3 It is a schematic structural diagram of an electronic device provided by the present invention;

[0036] Figure 4 It is a schematic diagram of an example structure of a rotation box in an embodiment of the present invention.

[0037] Reference numerals:

[0038] 1: Acquisition module; 2: Clustering module; 3: Estimation module;

[0039] 310: Processor; 320: Communication interface; 330: Memory;

[0040] 340: Communication bus. Detailed implementation manners

[0041] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments in the present invention belong to the scope of protection of the present invention.

[0042] The following combines Figure 1 to describe the method for estimating the state of driving obstacles based on a millimeter-wave radar of the present invention.

[0043] As Figure 1 shown, an embodiment of the present invention provides a method for estimating the state of driving obstacles based on a millimeter-wave radar, including:

[0044] Step 101, obtaining point cloud data based on a millimeter-wave radar;

[0045] Step 103, clustering the point cloud data to obtain at least one point cloud set of driving obstacles;

[0046] Step 105: Determine the state estimation result of the driving obstacle according to the driving obstacle point cloud set.

[0047] The state estimation result includes the bounding box of the driving obstacle; the bounding box is the rotated box with the minimum fitting cost; the rotated box is obtained by rotating the reference box; the reference box is a geometric figure of a set type, and the reference box is the minimum circumscribed box of the driving obstacle point cloud set at a reference angle; the fitting cost is calculated according to the driving obstacle point cloud set and the rotated box.

[0048] In this embodiment, the reference angle is a set geometric figure angle.

[0049] For example, for the reference box, rotated box, and bounding box of a two-dimensional ellipse, a set judgment vector on the straight line where the major axis of the ellipse is located can be selected as the angle determination criterion. Define the ellipse with the judgment vector in the same direction as the positive y-axis of the set Cartesian coordinate system as the ellipse at the reference angle, that is, define the judgment vector pointing to the positive y-axis as the reference vector, and the included angle between the judgment vector and the reference vector is the rotation angle of the rotated box (or bounding box) relative to the reference box.

[0050] Furthermore, for geometric figures of a certain type, there may be cases where the rotation angles of the rotated boxes are different but the geometric figures are identical. In the above example of the two-dimensional ellipse, when the included angle between the judgment vector and the reference vector is x degrees and (180 + x) degrees, the two-dimensional ellipses are congruent figures. Therefore, in this case, the rotation range of the rotated box can be set to 0 to 180 degrees to further save computing resources.

[0051] In a preferred embodiment:

[0052] The point cloud data obtained in step 101 is the preprocessed original acquisition data of the millimeter-wave radar.

[0053] Alternatively, the point cloud data obtained in step 101 is the original acquisition data of the millimeter-wave radar, and step 103 clusters and preprocesses the point cloud data to obtain at least one driving obstacle point cloud set.

[0054] The preprocessing can be any one or a combination of removing clutter, removing multi-reflection point clouds, removing point clouds outside the drivable area, and removing multipath point clouds.

[0055] In this embodiment, in the state estimation result obtained in step 105, the bounding box can be either a two-dimensional geometric figure, such as a rectangular box (in this case, the bounding box is the minimum circumscribed box of the point cloud data projected onto a certain set plane, such as the ground); or a three-dimensional geometric figure, such as a cube box (in this case, the bounding box is the minimum circumscribed box of the point cloud data in three-dimensional space).

[0056] The geometric types of the reference box, the rotated box, and the bounding box are the same in a specific obstacle estimation task. For example, if the geometric figure of the set type is a rectangle, then the reference box, the rotated box, and the bounding box are all rectangular boxes, but the sizes and rotation postures of the rectangular boxes of the rotated box and the bounding box may be different from those of the reference box.

[0057] The beneficial effect of this embodiment is as follows:

[0058] By introducing the fitting cost and selecting the bounding box at an appropriate angle, the bounding box has a higher information density and can reflect the angular orientation of the obstacle, thereby providing a basis for reducing the computational resource requirements in the subsequent autonomous driving decision-making process.

[0059] According to the above embodiment, in this embodiment:

[0060] The rotated box is the minimum bounding box of the driving obstacle point cloud set after rotating the reference box.

[0061] That is to say, on the basis of the above embodiment, after rotating the reference box in this embodiment, the size is adjusted to the minimum bounding box of the driving obstacle point cloud set to form a rotated box. This operation is beneficial to further improving the information density of the rotated box and the bounding box;

[0062] Figure 4 In, a two-dimensional rectangular box is selected as the geometric type of the reference box, the rotated box, and the bounding box, showing a construction example of the rotated box in this embodiment.

[0063] Such as Figure 4 On the left, the reference box is the minimum bounding rectangle of the point cloud at the reference angle (that is, a set vector on the straight line where the long side of the rectangle is located is selected as the judgment vector, and the direction with the judgment vector vertically upward is used as the reference vector);

[0064] Figure 4 The middle shows the positional relationship between the rotated reference box and the point cloud. On this basis, without changing the angle of the judgment vector, the size of the rectangular box is further adjusted to obtain the rotated minimum bounding rectangle as the rotated box ( Figure 4 On the right), and the fitting cost is calculated based on this rotated box.

[0065] Furthermore, in a preferred implementation manner, the rotated box is the minimum bounding box of the driving obstacle point cloud set after excluding the outliers after rotating the reference box;

[0066] The outlier is a point whose difference in fitting cost between the driving obstacle point cloud sets before and after exclusion is greater than a set threshold.

[0067] The beneficial effect of this embodiment is as follows:

[0068] By restricting the size of the rotated bounding box (i.e., making the size of the rotated bounding box the minimum circumscribed bounding box of the driving obstacle point cloud set), the information density in the finally obtained bounding box is improved;

[0069] By excluding outliers, the influence of acquisition errors not excluded in preprocessing on the obstacle state is further avoided.

[0070] According to any of the above embodiments, in this embodiment:

[0071] The fitting cost refers to the sum of the distances from the points in the driving obstacle point cloud set to the rotated bounding box;

[0072] The distance refers to the distance from the point to the straight line where the nearest side of the rotated bounding box is located.

[0073] The bounding box is the rotated bounding box with the minimum fitting cost in the set of rotated bounding boxes;

[0074] The set of rotated bounding boxes includes the rotated bounding boxes obtained by rotating a rectangular reference bounding box at set angles within the rotation angle range; the rotation angle range refers to 0 degrees to 180 degrees, or -90 degrees to 90 degrees.

[0075] It should be noted that if the reference bounding box, rotated bounding box, and bounding box are square boxes, the rotation angle range refers to 0 degrees to 90 degrees, or -90 degrees to 0 degrees.

[0076] The step of clustering the point cloud data to obtain at least one driving obstacle point cloud set includes:

[0077] Clustering the point cloud data by DBSCAN to obtain at least one point cloud cluster as the driving obstacle point cloud set based on density.

[0078] The beneficial effect of this embodiment lies in:

[0079] This embodiment further provides a feasible calculation method for the fitting cost for the most common rectangular reference bounding box, rotated bounding box, and bounding box. In addition, it also provides the rotation angle range of the rotated bounding box and the clustering step based on DBSCAN. Through the additional limitations of this embodiment, the driving obstacle state can be estimated more accurately and efficiently within the framework of the prior art.

[0080] According to any of the above embodiments, from the perspective of the dynamic process, the steps of a complete method for estimating the driving obstacle state based on a millimeter-wave radar will be provided below:

[0081] Step 101, obtaining point cloud data based on a millimeter-wave radar;

[0082] Step 103, clustering the point cloud data by DBSCAN to obtain at least one driving obstacle point cloud set;

[0083] Step 1051: Determine a reference bounding box according to the driving obstacle point cloud set;

[0084] Step 1052: Rotate the reference bounding box at a set angle within the range of rotation angles;

[0085] Step 1053: Adjust the size of the rotated reference bounding box to obtain the minimum bounding box of the driving obstacle point cloud set at this angle as the rotated bounding box;

[0086] Step 1054: For suspected outlier points, calculate the fitting cost of the rotated bounding box before and after exclusion (i.e., the minimum bounding box before and after excluding suspected outlier points at this angle), and determine the suspected outlier points with the difference between the fitting costs before and after exclusion greater than the set threshold as outlier points;

[0087] The suspected outlier points refer to the points with a distance less than the set judgment threshold relative to the rotated bounding box;

[0088] Step 1055: After excluding the outlier points determined in Step 1054, determine the rotated bounding box at this angle and calculate the fitting cost;

[0089] Step 1056: Repeat Steps 1052 to 1055 until a specified number of rotation angles are traversed.

[0090] Step 1057: Take the rotated bounding box with the minimum fitting cost as the reference bounding box and add it to the state estimation result of the driving obstacle;

[0091] The beneficial effect of this embodiment is as follows:

[0092] By introducing the fitting cost and selecting the bounding box at a suitable angle, the bounding box has a higher information density and can reflect the angular orientation of the obstacle, thereby providing a basis for reducing the computational resource requirements in the subsequent autonomous driving decision-making process.

[0093] Next, the driving obstacle state estimation device based on millimeter-wave radar provided by the present invention will be described. The driving obstacle state estimation device based on millimeter-wave radar described below can be mutually referred to the driving obstacle state estimation method based on millimeter-wave radar described above.

[0094] As Figure 2 shown, the embodiment of the present invention further provides a driving obstacle state estimation system based on millimeter-wave radar, including:

[0095] Acquisition module 1, configured to acquire point cloud data based on a millimeter-wave radar;

[0096] Clustering module 2, configured to cluster the point cloud data to obtain at least one driving obstacle point cloud set;

[0097] An estimation module 3, configured to determine a state estimation result of the driving obstacle according to the set of driving obstacle point clouds;

[0098] The state estimation result includes a bounding box of the driving obstacle; the bounding box is a rotated box with the minimum fitting cost; the rotated box is obtained by rotating a reference box; the reference box is a geometric figure of a set type, and the reference box is the minimum circumscribed box of the set of driving obstacle point clouds at a reference angle; the fitting cost is calculated according to the set of driving obstacle point clouds and the rotated box.

[0099] The clustering module 2 further includes:

[0100] A DBSCAN unit, configured to cluster the point cloud data by DBSCAN and obtain at least one point cloud cluster as the set of driving obstacle point clouds based on density.

[0101] The beneficial effects of this embodiment are as follows:

[0102] By introducing the fitting cost and selecting a bounding box at an appropriate angle, the bounding box has a higher information density and can reflect the angular orientation of the obstacle, thereby providing a basis for reducing the computational resource requirements in the subsequent autonomous driving decision-making process.

[0103] Figure 3 An example of a schematic physical structure diagram of an electronic device is shown in Figure 3 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340. Among them, the processor 310, the communication interface 320, and the memory 330 complete mutual communication through the communication bus 340. The processor 310 can call the logical instructions in the memory 330 to execute a method for estimating the state of a driving obstacle based on a millimeter-wave radar. The method includes: acquiring point cloud data based on the millimeter-wave radar; clustering the point cloud data to obtain at least one set of driving obstacle point clouds; determining a state estimation result of the driving obstacle according to the set of driving obstacle point clouds; the state estimation result includes a bounding box of the driving obstacle; the bounding box is a rotated box with the minimum fitting cost; the rotated box is obtained by rotating a reference box; the reference box is a geometric figure of a set type, and the reference box is the minimum circumscribed box of the set of driving obstacle point clouds at a reference angle; the fitting cost is calculated according to the set of driving obstacle point clouds and the rotated box.

[0104] In addition, when the logical instructions in the above-mentioned memory 330 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0105] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the method for estimating the state of driving obstacles based on a millimeter-wave radar provided by the above-mentioned various methods. The method includes: obtaining point cloud data based on the millimeter-wave radar; clustering the point cloud data to obtain at least one point cloud set of driving obstacles; determining a state estimation result of the driving obstacles according to the point cloud set of driving obstacles; the state estimation result includes a bounding box of the driving obstacles; the bounding box is a rotated box with the minimum fitting cost; the rotated box is obtained by rotating a reference box; the reference box is a geometric figure of a set type, and the reference box is the minimum circumscribed box of the point cloud set of driving obstacles at a reference angle; the fitting cost is calculated according to the point cloud set of driving obstacles and the rotated box.

[0106] On another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the method for estimating the state of driving obstacles based on a millimeter-wave radar provided by the above-mentioned various methods. The method includes: obtaining point cloud data based on the millimeter-wave radar; clustering the point cloud data to obtain at least one point cloud set of driving obstacles; determining a state estimation result of the driving obstacles according to the point cloud set of driving obstacles; the state estimation result includes a bounding box of the driving obstacles; the bounding box is a rotated box with the minimum fitting cost; the rotated box is obtained by rotating a reference box; the reference box is a geometric figure of a set type, and the reference box is the minimum circumscribed box of the point cloud set of driving obstacles at a reference angle; the fitting cost is calculated according to the point cloud set of driving obstacles and the rotated box.

[0107] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative work.

[0108] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0109] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for estimating the state of driving obstacles based on millimeter-wave radar, characterized in that, Including: Obtaining point cloud data based on a millimeter-wave radar; Clustering the point cloud data to obtain at least one driving obstacle point cloud set; Determining a state estimation result of the driving obstacle according to the driving obstacle point cloud set; The state estimation result includes a bounding box of the driving obstacle; The bounding box is a rotated box with the minimum fitting cost; The rotated box is obtained by rotating a reference box; The reference box is a geometric figure of a set type, and the reference box is the minimum circumscribed box of the driving obstacle point cloud set at a reference angle; The fitting cost is calculated according to the driving obstacle point cloud set and the rotated box; The rotated box is the minimum circumscribed box of the driving obstacle point cloud set after rotating the reference box; The fitting cost refers to the sum of the distances from the points in the driving obstacle point cloud set to the rotated box; The distance is the distance from the point to the straight line where the nearest side of the rotated box is located.

2. The method for estimating the state of driving obstacles based on millimeter-wave radar according to claim 1, characterized in that The rotated box is the minimum circumscribed box of the driving obstacle point cloud set after excluding outliers after rotating the reference box; The outlier is a point whose difference in fitting cost between the driving obstacle point cloud sets before and after exclusion is greater than a set threshold.

3. The method for estimating the state of driving obstacles based on a millimeter-wave radar according to any one of claims 1 to 2, characterized in that, The bounding box is the rotated box with the minimum fitting cost in the set of rotated boxes; The set of rotated boxes includes rotated boxes obtained by rotating a rectangular reference box at set angles within a rotation angle range; the rotation angle range refers to 0 degrees to 180 degrees, or -90 degrees to 90 degrees.

4. The method for estimating the state of driving obstacles based on millimeter-wave radar according to any one of claims 1 to 2, characterized in that, The step of clustering the point cloud data to obtain at least one driving obstacle point cloud set includes: Clustering the point cloud data by DBSCAN to obtain at least one point cloud cluster as a driving obstacle point cloud set based on density.

5. A driving obstacle state estimation system based on millimeter-wave radar, characterized in that, Including: An acquisition module for obtaining point cloud data based on a millimeter-wave radar; A clustering module for clustering the point cloud data to obtain at least one driving obstacle point cloud set; An estimation module for determining a state estimation result of the driving obstacle according to the driving obstacle point cloud set; The state estimation result includes a bounding box of the driving obstacle; The bounding box is a rotated box with the minimum fitting cost; The rotated box is obtained by rotating a reference box; The reference box is a geometric figure of a set type, and the reference box is the minimum circumscribed box of the driving obstacle point cloud set at a reference angle; The fitting cost is calculated according to the driving obstacle point cloud set and the rotated box; The rotated box is the minimum circumscribed box of the driving obstacle point cloud set after rotating the reference box; The fitting cost refers to the sum of the distances from the points in the driving obstacle point cloud set to the rotated box; the distance is the distance from the point to the straight line where the nearest side of the rotated box is located.

6. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method for estimating the state of a driving obstacle based on a millimeter-wave radar according to any one of claims 1 to 4.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method for estimating the state of a driving obstacle based on a millimeter-wave radar according to any one of claims 1 to 4.

8. A computer program product comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method for estimating the state of a driving obstacle based on a millimeter-wave radar according to any one of claims 1 to 4.

Images