Vehicle driving area map drawing method, device, detection method and detection system

Abstract

The present invention belongs to the field of autonomous driving technology and specifically provides a method, device, detection method, and detection system for mapping a vehicle's driving area. The system includes a communication device, a mapping device, and a control device; the communication device is connected to the mapping device; the communication device is used to read vehicle wheel speed information and laser radar point cloud data and transmit the read data to the mapping device; a control module is connected to the mapping device and the communication device respectively; the point cloud is ground segmented, and multiple frames of ground point clouds are superimposed to form a local ground map of a certain range around the vehicle, a non-ground point cloud, and simultaneously measure the distance between the vehicle and obstacles and road boundaries, and provide prompts and send them to the control device. The control device is used to read the grid map drawn by the mapping device, measure the distance between the vehicle and obstacles and road boundaries, and transmit deceleration and braking instructions to the vehicle through the communication device. This improves the safety of vehicle operation.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a vehicle driving area map drawing method, device, detection method, and detection system. Background Art

[0002] LiDAR is a radar system that uses laser beams to detect a target's position, velocity, and other characteristic parameters. Its operating principle involves transmitting a detection signal (a laser beam) toward the target. The received signal (the target echo) reflected from the target is then compared with the transmitted signal. After appropriate processing, relevant information about the target is obtained.

[0003] During the driving process of an unmanned vehicle, the vehicle may encounter obstacles and road boundaries. Especially when the combined navigation positioning is unstable, the vehicle may easily move suddenly in a certain direction, causing the vehicle to collide with other obstacles or run out of the road boundary, affecting the driving safety of the autonomous vehicle.

[0004] The problem to be solved in this application is how to use lidar to map the drivable area during unmanned driving operation and provide real-time warning of obstacles or outside road boundaries. Summary of the Invention

[0005] In order to solve the problem of how to use laser radar to map the drivable area during unmanned driving operation for real-time early warning of obstacles or outside road boundaries, the present invention provides a vehicle driving area map drawing method, device, detection method and detection system.

[0006] The technical solution of the present invention is:

[0007] In a first aspect, the technical solution of the present invention provides a method for mapping a vehicle driving area, comprising the following steps:

[0008] Access the lidar point cloud and vehicle wheel speed information through configuration;

[0009] Divide the point cloud into grids, calculate the height difference between adjacent grids and the point clouds within each grid, and extract the ground point cloud and non-ground point cloud based on the calculated height difference;

[0010] The initial matching value assigned by the vehicle wheel speed information is matched with the non-ground point cloud to obtain a transformation matrix, and the transformation matrix is ​​multiplied by the ground point cloud to obtain a point cloud in the current vehicle coordinate system. The transformed point clouds of multiple frames are superimposed to obtain a superimposed ground point cloud.

[0011] The superimposed ground point cloud is added to the grid map to represent the drivable area; the real-time non-ground point cloud is added to the grid map to represent obstacles, and the grid without point cloud in the grid map is used to represent unknown areas outside the road boundary.

[0012] When a LiDAR radar illuminates areas beyond the road boundary, the point cloud disappears. Each LiDAR beam forms an elliptical point cloud on the ground, separated by a certain distance. Environmental information within this area is unknown. To improve environmental information, point cloud matching is used to overlay multiple point cloud frames and reduce LiDAR blind spots.

[0013] Preferably, before the steps of dividing the point cloud into grids, calculating the height differences between adjacent grids and point clouds within each grid, and extracting the ground point cloud and the non-ground point cloud by the calculated height differences, the method further comprises:

[0014] Preprocess the received point cloud, including:

[0015] Downsample the point cloud;

[0016] The lidar point cloud is converted to the vehicle coordinate system, and noise points and outliers are removed from each frame of the received point cloud.

[0017] Point cloud preprocessing includes downsampling the point cloud, converting the LiDAR point cloud to the vehicle coordinate system, removing the point clouds that are far away from the vehicle body and the point clouds on the vehicle body from each frame of the received point cloud, and removing noise to form a preprocessed point cloud.

[0018] Preferably, the steps of dividing the point cloud into grids, calculating the height differences between adjacent grids and point clouds within each grid, and extracting the ground point cloud and the non-ground point cloud by the calculated height differences specifically include:

[0019] Mesh the pre-processed point cloud;

[0020] Calculate the height difference between adjacent grids and the point clouds within each grid;

[0021] The point clouds with height differences greater than the set height threshold are extracted as ground point clouds through the calculated height differences, and the point clouds with height differences less than the set height threshold are extracted as ground point clouds.

[0022] Preferably, the steps of performing non-ground point cloud matching on the matching initial value assigned by the vehicle wheel speed information, calculating a transformation matrix, multiplying the calculated transformation matrix with the ground point cloud to obtain a point cloud in the current vehicle body coordinate system, and superimposing multiple frames of transformed point clouds to obtain a superimposed ground point cloud include:

[0023] Create a structure to store a frame of ground point cloud and the transformation matrix that realizes the coordinate transformation between the ground point cloud and the vehicle body, and create a container of size m to store m structures. Each structure in the container is an element of the container; that is, m frames of ground point cloud and the corresponding transformation matrix;

[0024] Assigning an initial matching value based on the vehicle wheel speed information, matching the non-ground point cloud with the previous frame point cloud to obtain a transformation matrix between the ground point clouds, and storing the data in a container when the transformation distance is greater than a second threshold;

[0025] Multiply the transformation matrix of each element in the container with the ground point cloud to obtain the point cloud in the current vehicle coordinate system, and superimpose the transformed point clouds of multiple frames to obtain the superimposed ground point cloud.

[0026] Because ground point clouds are mostly distributed on a flat surface, point cloud registration is difficult and prone to errors. During processing, ground point clouds and non-ground point clouds appear in pairs. The coordinate transformation between the ground point clouds is calculated by the coordinate transformation between the non-ground point cloud and the previous frame's point cloud. Wheel speed information is required to provide an initial value for the matching during self-matching. To reduce point cloud density, the data is stored in a container when the transformation distance exceeds a second threshold. When the container size exceeds m, the elements in the first container are pushed out, and the new ground and coordinate transformation data are saved.

[0027] In a second aspect, the technical solution of the present invention provides a vehicle driving area mapping device, comprising a data access module, a point cloud segmentation module, a point cloud overlay module, and a grid map drawing module;

[0028] Data access module, used to access lidar point cloud and vehicle wheel speed information through configuration;

[0029] The point cloud segmentation module is used to divide the point cloud into grids, calculate the height difference between adjacent grids and the point clouds within each grid, and extract the ground point cloud and non-ground point cloud based on the calculated height difference;

[0030] The point cloud overlay module is used to perform non-ground point cloud matching using the initial matching value assigned by the vehicle wheel speed information to obtain a transformation matrix, and multiply the transformation matrix with the ground point cloud to obtain a point cloud in the current vehicle coordinate system. The transformed point clouds of multiple frames are then overlaid to obtain a superimposed ground point cloud.

[0031] The raster map drawing module is used to add the superimposed ground point cloud to the raster map to represent the drivable area; the real-time non-ground point cloud is added to the raster map to represent obstacles, and the grid without point cloud in the raster map is used to represent unknown areas outside the road boundary.

[0032] Preferably, the device further comprises a point cloud pre-processing module;

[0033] The point cloud preprocessing module is used to downsample the point cloud and convert the lidar point cloud into the vehicle coordinate system. It removes noise points and outliers from each frame of the received point cloud to form a preprocessed point cloud that is input into the point cloud segmentation module.

[0034] Preferably, the point cloud segmentation module is specifically used to divide the preprocessed point cloud into grids, calculate the height difference between adjacent grids and the point clouds within each grid, and extract the point clouds with height differences greater than the set height threshold as ground point clouds through the calculated height differences, and extract the point clouds with height differences less than the set height threshold as ground point clouds.

[0035] Preferably, the point cloud overlay module includes a storage creation unit, a ground point cloud transformation processing unit and an overlay unit;

[0036] The storage creation unit is used to create a structure to store a frame of ground point cloud and the transformation matrix that realizes the coordinate transformation between the ground point cloud and the vehicle body, and create a container of size m to store m structures, where each structure in the container is an element of the container;

[0037] A ground point cloud transformation processing unit is used to assign a matching initial value according to the vehicle wheel speed information, calculate the transformation matrix between the ground point clouds by matching the coordinate transformation between the non-ground point cloud and the previous frame point cloud, and store the data in a container when the transformation distance is greater than a second threshold;

[0038] The superposition unit is used to multiply the transformation matrix of each element in the container with the ground point cloud to obtain the point cloud in the current vehicle coordinate system, and superimpose the transformed point clouds of multiple frames to obtain the superimposed ground point cloud.

[0039] In a third aspect, the technical solution of the present invention provides a method for detecting a vehicle driving area, comprising the following steps:

[0040] Read vehicle wheel speed information and lidar point cloud and send the read data;

[0041] Drawing a grid map using the map drawing method of the first aspect;

[0042] Read the drawn grid map, measure the distance between the vehicle and obstacles and road boundaries, and send control instructions to the vehicle.

[0043] According to the relationship between the vehicle body and surrounding obstacles and unknown areas, corresponding control strategies are adopted and control instructions are issued to achieve vehicle deceleration and braking.

[0044] Preferably, the steps of reading the vehicle wheel speed information, the point cloud of the laser radar and sending the read data include:

[0045] Receive the point cloud collected by the LiDAR and analyze the vehicle wheel speed information in the vehicle CAN data;

[0046] The parsed vehicle wheel speed information and the received point cloud are sent down.

[0047] In a fourth aspect, the technical solution of the present invention provides a vehicle driving area detection system, comprising a communication device, a mapping device and a control device;

[0048] The communication device is connected to the mapping device;

[0049] The communication device is used to read the vehicle wheel speed information and the point cloud data of the laser radar and send the read data to the map drawing device;

[0050] The map drawing device is the map drawing device described in the second aspect;

[0051] The control module is connected to the map drawing device and the communication device respectively;

[0052] The control device is used to measure the distance between the vehicle and obstacles and road boundaries by reading the grid map drawn by the map drawing device, and send control instructions to the vehicle through the communication device.

[0053] It can detect obstacles and road boundaries and provide them to the control device. The control device makes corresponding control strategies based on the distance between the vehicle and the obstacles and road boundaries and sends them to the vehicle through the communication device to achieve vehicle deceleration and emergency braking, etc.

[0054] Preferably, the communication device includes a CAN communication data receiving module and a laser radar data receiving module;

[0055] A CAN communication data receiving module is used to parse the vehicle wheel speed information in the vehicle CAN data and input the parsed vehicle wheel speed information into the map drawing device;

[0056] The laser radar data receiving module is used to receive data collected by the laser radar and input the data into the map drawing device.

[0057] As can be seen from the above technical solution, the present invention has the following advantages: using non-ground objects to represent obstacles; creating an overlaid ground point cloud to accurately describe the situation outside the road boundary, and representing the ground, obstacles and outside the road boundary with different numerical values ​​in a grid map. The grid map is provided to the control system for formulating different safety strategies, thereby improving the safety of the unmanned vehicle system when the positioning system is unstable.

[0058] In addition, the present invention has a reliable design principle, a simple structure and a very broad application prospect.

[0059] It can be seen that compared with the prior art, the present invention has outstanding substantial features and significant progress, and the beneficial effects of its implementation are also obvious. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present invention 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.

[0061] Figure 1 It is a schematic flow chart of a map drawing method according to an embodiment of the present invention.

[0062] Figure 2 is a schematic block diagram of a map drawing device according to an embodiment of the present invention.

[0063] Figure 3 It is a schematic flow chart of a detection method according to an embodiment of the present invention.

[0064] Figure 4 It is a schematic block diagram of a detection system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0065] LiDAR is a sensor widely used in autonomous driving that can accurately output point cloud information near the vehicle. Compared with cameras, it is not affected by light and has been widely used in autonomous vehicles. In order to improve the safety of autonomous vehicles during operation, it is necessary to observe obstacles in the vehicle's surrounding environment and information outside the road boundary in real time to provide a safe drivable area for the vehicle. The present invention performs ground segmentation on the point cloud, superimposes multiple frames of ground point clouds to form a local ground map of a certain range around the vehicle, and non-ground point clouds, while measuring the distance between the vehicle and obstacles and outside the road boundary, and prompts and sends them to the control device. The control device performs corresponding deceleration and braking actions according to the distance between the vehicle and obstacles and rivers outside the road boundary, thereby improving the safety of vehicle operation.

[0066] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0067] like Figure 1 As shown, an embodiment of the present invention provides a method for drawing a vehicle driving area map, comprising the following steps:

[0068] S1: Access the lidar point cloud and vehicle wheel speed information through configuration;

[0069] S2: Divide the point cloud into grids, calculate the height difference between adjacent grids and the point clouds within each grid, and extract the ground point cloud and non-ground point cloud based on the calculated height difference;

[0070] The point cloud is meshed and the height difference between a certain grid and the surrounding grids is calculated. If the height difference is greater than the threshold, it is a non-ground point cloud, and if it is less than the threshold, it is a ground point cloud.

[0071] S3: Perform non-ground point cloud matching on the initial matching value assigned by the vehicle wheel speed information to obtain a transformation matrix, and multiply the transformation matrix with the ground point cloud to obtain a point cloud in the current vehicle coordinate system. The transformed point clouds of multiple frames are superimposed to obtain a superimposed ground point cloud.

[0072] By matching the non-ground point clouds in the two frames of point cloud data, the transformation matrix between the ground point clouds is obtained. After obtaining the transformation matrix, one frame of ground point cloud can be superimposed on another ground point cloud.

[0073] S4: The superimposed ground point cloud is added to the grid map to represent the drivable area; the real-time non-ground point cloud is added to the grid map to represent obstacles, and the grid without point cloud in the grid map is used to represent unknown areas outside the road boundary.

[0074] It should be noted that before the step of dividing the point cloud into grids in step S2, calculating the height differences between adjacent grids and point clouds within each grid, and extracting the ground point cloud and the non-ground point cloud by the calculated height differences, the following steps are also included:

[0075] S02: Pre-process the received point cloud, including:

[0076] S021: Downsample the point cloud;

[0077] S022: Convert the LiDAR point cloud to the vehicle coordinate system, and remove noise points and outliers from each frame of the received point cloud.

[0078] The lidar point cloud data is converted to the vehicle coordinate system, and the point clouds far away from the vehicle body and the point clouds on the vehicle body are removed from each frame of the received point cloud to remove noise and form a pre-processed point cloud.

[0079] In some embodiments, the steps of dividing the point cloud into grids in step S2, calculating the height differences between adjacent grids and point clouds within each grid, and extracting ground point clouds and non-ground point clouds based on the calculated height differences specifically include:

[0080] S21: Meshing the pre-processed point cloud;

[0081] The preprocessed point cloud is divided into grids with a grid size of M×N;

[0082] S22: Calculate the height difference between adjacent grids and the point cloud inside each grid;

[0083] Calculate the maximum value Z of the point cloud in each grid G ​​in the direction perpendicular to the ground max , minimum value Z min , and calculate the maximum height difference Z dist ;

[0084] S23: extracting the point cloud with a height difference greater than a set height threshold as the ground point cloud and extracting the point cloud with a height difference less than the set height threshold as the ground point cloud based on the calculated height difference.

[0085] Point clouds that meet the following conditions are considered non-ground point clouds:

[0086] 1) Height difference Z within the same grid dist Greater than the threshold Z th1 ;

[0087] 2) Grid G ​​to be inspected ij With its upper side grid G ij+1 Height difference between |G ij (Z min )-G ij+1 (Z min )|greater than the threshold Z th2 ;

[0088] 3) Grid G ​​to be inspected ij With its lower side grid G ij-1 Height difference between |G ij (Z min )-G ij-1 (Z min )|greater than the threshold Z th3 ;

[0089] 4) Grid G ​​to be inspected ij With its left grid G i-1j Height difference between |G ij (Z min )-G i-1j (Z min )|greater than the threshold Z th4 ;

[0090] 5) Grid G ​​to be inspected ij and its right side grid G i+1j Height difference between |G ij (Z min )-G i+1j (Z min )|greater than the threshold Z th5 ;

[0091] Extract the ground point cloud P1 and non-ground point cloud P2 through the above conditions;

[0092] The pre-processed point cloud is divided into grids, and the height difference between adjacent grids and the point cloud inside each grid is calculated. A threshold is set, and points with a height difference greater than the threshold are considered non-ground point clouds, while points with a height difference less than the threshold are considered ground point clouds. Ground point clouds and non-ground point clouds are extracted. Non-ground point clouds are considered obstacle point clouds.

[0093] In some embodiments, in step S3, performing non-ground point cloud matching on the matching initial value assigned by the vehicle wheel speed information, calculating a transformation matrix, multiplying the calculated transformation matrix with the ground point cloud to obtain a point cloud in the current vehicle coordinate system, and superimposing multiple frames of transformed point clouds to obtain a superimposed ground point cloud includes:

[0094] S31: Create a structure to store a frame of ground point cloud and the transformation matrix that realizes the coordinate transformation between the ground point cloud and the vehicle body, and create a container of size m to store m structures, each structure in the container is an element of the container; that is, m frames of ground point cloud and the corresponding transformation matrix;

[0095] Create a structure {P, T}, where P represents the point cloud and T represents the transformation matrix, which is used to store the ground point cloud of a frame of point cloud and the coordinate transformation between it and the vehicle body. Create a container {{P0, T0}, {P1, T1}, ..., {P m , T m}}Used to store m frames of ground data and coordinate transformation.

[0096] S32: assigning an initial matching value according to the vehicle wheel speed information, matching the non-ground point cloud with the previous frame point cloud to obtain a transformation matrix between the ground point clouds, and storing the data in a container when the transformation distance is greater than a second threshold;

[0097] Since ground point clouds are mostly distributed on a flat surface, point cloud registration is difficult and prone to errors. Ground and non-ground point clouds are processed in pairs. The coordinate transformation between the ground point clouds is calculated by registering the coordinate transformation between the non-ground point cloud and the previous frame's point cloud. Self-matching requires wheel speed information to provide an initial value for the match. To reduce point cloud density, data is stored in a container when the transformation distance exceeds a threshold L. When the container size exceeds m, the elements in the first container are pushed out, and the new ground and coordinate transformation data are saved.

[0098] S33: Multiply the transformation matrix of each element in the container by the ground point cloud to obtain a point cloud in the current vehicle coordinate system, and superimpose the transformed point clouds of multiple frames to obtain a superimposed ground point cloud.

[0099] like Figure 2As shown, an embodiment of the present invention provides a vehicle driving area map drawing device, including a data access module, a point cloud segmentation module, a point cloud overlay module, and a grid map drawing module;

[0100] Data access module, used to access lidar point cloud and vehicle wheel speed information through configuration;

[0101] The point cloud segmentation module is used to divide the point cloud into grids, calculate the height difference between adjacent grids and the point clouds within each grid, and extract the ground point cloud and non-ground point cloud based on the calculated height difference;

[0102] The point cloud overlay module is used to perform non-ground point cloud matching using the initial matching value assigned by the vehicle wheel speed information to obtain a transformation matrix, and multiply the transformation matrix with the ground point cloud to obtain a point cloud in the current vehicle coordinate system. The transformed point clouds of multiple frames are then overlaid to obtain a superimposed ground point cloud.

[0103] The raster map drawing module is used to add the superimposed ground point cloud to the raster map to represent the drivable area; the real-time non-ground point cloud is added to the raster map to represent obstacles, and the grid without point cloud in the raster map is used to represent unknown areas outside the road boundary.

[0104] It should be noted that the device also includes a point cloud pre-processing module;

[0105] The point cloud preprocessing module is used to downsample the point cloud and convert the LiDAR point cloud to the vehicle coordinate system. It then removes noise and outliers from each frame of the received point cloud to form a preprocessed point cloud that is input into the point cloud segmentation module. The point cloud segmentation module is specifically used to grid the preprocessed point cloud, calculate the height difference between adjacent grids and the point clouds within each grid, and extract point clouds with height differences greater than a set height threshold as ground point clouds, and point clouds with height differences less than a set height threshold as ground point clouds based on the calculated height differences.

[0106] In some embodiments, the point cloud overlay module includes a storage creation unit, a ground point cloud transformation processing unit, and an overlay unit;

[0107] The storage creation unit is used to create a structure to store a frame of ground point cloud and the transformation matrix that realizes the coordinate transformation between the ground point cloud and the vehicle body, and create a container of size m to store m structures, where each structure in the container is an element of the container;

[0108] The ground point cloud transformation processing unit is used to assign matching initial values ​​according to the vehicle wheel speed information, obtain the transformation matrix between the ground point clouds by matching the non-ground point cloud with the previous frame point cloud, and store the data in a container when the transformation distance is greater than a second threshold;

[0109] The superposition unit is used to multiply the transformation matrix of each element in the container with the ground point cloud to obtain the point cloud in the current vehicle coordinate system, and superimpose the transformed point clouds of multiple frames to obtain the superimposed ground point cloud.

[0110] like Figure 3 As shown, an embodiment of the present invention provides a vehicle driving area detection method, comprising the following steps:

[0111] Step 1: Read vehicle wheel speed information and lidar point cloud and send the read data;

[0112] Step 2: Draw a raster map using the map drawing method;

[0113] In this step, the method for drawing the grid map is the vehicle driving area map drawing method provided in the above embodiment; it will not be described in detail here;

[0114] Step 3: Read the drawn grid map, measure the distance between the vehicle and obstacles and road boundaries, and send control instructions to the vehicle.

[0115] By receiving information from the grid map, the system can determine the distance between obstacles in the grid map and the road boundary and take corresponding actions such as deceleration and emergency stops.

[0116] In some embodiments, the step of reading the vehicle wheel speed information and the laser radar point cloud and sending the read data in step 1 includes:

[0117] Step 11: Receive the point cloud collected by the lidar and analyze the vehicle wheel speed information in the vehicle CAN data. Those skilled in the art know that the vehicle wheel speed is generally collected by a wheel speed meter.

[0118] Step 12: Send the parsed vehicle wheel speed information and the received point cloud.

[0119] like Figure 4 As shown, an embodiment of the present invention provides a vehicle driving area detection system, including a communication device, a map drawing device and a control device;

[0120] The communication device is connected to the mapping device;

[0121] The communication device is used to read the vehicle wheel speed information and the point cloud data of the laser radar and send the read data to the map drawing device;

[0122] The map drawing device is the map drawing device provided in the above device embodiment;

[0123] The control module is connected to the map drawing device and the communication device respectively;

[0124] The control device is used to measure the distance between the vehicle and obstacles and road boundaries by reading the grid map drawn by the map drawing device, and send control instructions to the vehicle through the communication device.

[0125] In some embodiments, the communication device includes a CAN communication data receiving module and a lidar data receiving module;

[0126] A CAN communication data receiving module is used to parse the vehicle wheel speed information in the vehicle CAN data and input the parsed vehicle wheel speed information into the map drawing device;

[0127] The laser radar data receiving module is used to receive data collected by the laser radar and input the data into the map drawing device.

[0128] Although the present invention has been described in detail with reference to the accompanying drawings and in conjunction with preferred embodiments, the present invention is not limited thereto. Without departing from the spirit and essence of the present invention, a person of ordinary skill in the art may make various equivalent modifications or substitutions to the embodiments of the present invention, and such modifications or substitutions shall be within the scope of the present invention. Any person skilled in the art who can easily conceive of changes or substitutions within the technical scope disclosed in the present invention shall be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be based on the scope of protection of the claims.

Claims

1. A method for drawing a vehicle driving area map, characterized in that: The steps include: Access the lidar point cloud and vehicle wheel speed information through configuration; Divide the point cloud into grids, calculate the height difference between adjacent grids and the point clouds within each grid, and extract the ground point cloud and non-ground point cloud based on the calculated height difference; The initial matching value assigned by the vehicle wheel speed information is matched with the non-ground point cloud to obtain a transformation matrix, and the transformation matrix is ​​multiplied by the ground point cloud to obtain a point cloud in the current vehicle coordinate system. The transformed point clouds of multiple frames are superimposed to obtain a superimposed ground point cloud. The superimposed ground point cloud is added to the grid map to represent the drivable area; the real-time non-ground point cloud is added to the grid map to represent obstacles, and the grid without point cloud in the grid map is used to represent the unknown area, which is the area outside the road boundary; The steps of performing non-ground point cloud matching using the initial matching value assigned by the vehicle wheel speed information, calculating a transformation matrix, multiplying the calculated transformation matrix with the ground point cloud to obtain a point cloud in the current vehicle coordinate system, and superimposing multiple frames of transformed point clouds to obtain a superimposed ground point cloud include: Create a structure to store a frame of ground point cloud and the transformation matrix that realizes the coordinate transformation between the ground point cloud and the vehicle body, and create a container of size m to store m structures, where each structure in the container is an element of the container; According to the vehicle wheel speed information, the initial value of the matching is assigned, and the non-ground point cloud is matched with the previous frame point cloud to obtain the transformation matrix between the ground point cloud. When the transformation distance is greater than the second threshold, the structure {P, T} is stored in a container, where P represents the point cloud and T represents the transformation matrix. It is used to store the ground point cloud of a frame of point cloud and the coordinate transformation between it and the vehicle body; Multiply the transformation matrix of each element in the container with the ground point cloud to obtain the point cloud in the current vehicle coordinate system, and superimpose the transformed point clouds of multiple frames to obtain the superimposed ground point cloud.

2. The vehicle driving area map drawing method according to claim 1, characterized in that: The steps of meshing the point cloud, calculating the height difference between adjacent grids and the point clouds within each grid, and extracting the ground point cloud and the non-ground point cloud by the calculated height difference also include: Preprocess the received point cloud, including: Downsample the point cloud; The lidar point cloud is converted to the vehicle coordinate system, and noise points and outliers are removed from each frame of the received point cloud.

3. A vehicle driving area map drawing device, characterized in that: Including data access module, point cloud segmentation module, point cloud overlay module, and raster map drawing module; Data access module, used to access lidar point cloud and vehicle wheel speed information through configuration; The point cloud segmentation module is used to divide the point cloud into grids, calculate the height difference between adjacent grids and the point clouds within each grid, and extract the ground point cloud and non-ground point cloud based on the calculated height difference; The point cloud overlay module is used to perform non-ground point cloud matching using the initial matching value assigned by the vehicle wheel speed information to obtain a transformation matrix, and multiply the transformation matrix with the ground point cloud to obtain a point cloud in the current vehicle coordinate system. The transformed point clouds of multiple frames are then overlaid to obtain a superimposed ground point cloud. The grid map drawing module is used to add the superimposed ground point cloud to the grid map to represent the drivable area; the real-time non-ground point cloud is added to the grid map to represent obstacles, and the grid without point cloud in the grid map is used to represent unknown areas and outside the road boundary; The point cloud overlay module includes a storage creation unit, a ground point cloud transformation processing unit and an overlay unit; The storage creation unit is used to create a structure to store a frame of ground point cloud and a transformation matrix that realizes the coordinate transformation between the ground point cloud and the vehicle body, and to create a container of size m to store m structures, where each structure in the container is an element of the container; The ground point cloud transformation processing unit is used to assign matching initial values ​​according to the vehicle wheel speed information, obtain the transformation matrix between the ground point cloud by matching the non-ground point cloud with the previous frame point cloud, and store the structure {P, T} into a container when the transformation distance is greater than a second threshold, where P represents the point cloud and T represents the transformation matrix, which is used to store the ground point cloud of a frame point cloud and the coordinate transformation between it and the vehicle body; The superposition unit is used to multiply the transformation matrix of each element in the container with the ground point cloud to obtain the point cloud in the current vehicle coordinate system, and superimpose the transformed point clouds of multiple frames to obtain the superimposed ground point cloud.

4. The vehicle driving area map drawing device according to claim 3, characterized in that: The device also includes a point cloud pre-processing module; The point cloud preprocessing module is used to downsample the point cloud and convert the lidar point cloud into the vehicle coordinate system. It removes noise points and outliers from each frame of the received point cloud to form a preprocessed point cloud that is input into the point cloud segmentation module.

5. A vehicle driving area detection method, characterized in that: The steps include: Read vehicle wheel speed information and lidar point cloud and send the read data; Drawing a grid map by the map drawing method according to any one of claims 1 to 2; Read the drawn grid map, measure the distance between the vehicle and obstacles and road boundaries, and send control instructions to the vehicle.

6. The vehicle driving area detection method according to claim 5, characterized in that: The steps for reading vehicle wheel speed information, LiDAR point cloud, and sending the read data include: Receive the point cloud collected by the LiDAR and analyze the vehicle wheel speed information in the vehicle CAN data; The parsed vehicle wheel speed information and the received point cloud are sent down.

7. A vehicle driving area detection system, characterized in that: including a communication device, a mapping device, and a control device; The communication device is connected to the mapping device; The communication device is used to read the vehicle wheel speed information and the point cloud data of the laser radar and send the read data to the map drawing device; The map drawing device comprises the map drawing device according to any one of claims 3 to 4; The control module is connected to the map drawing device and the communication device respectively; The control device is used to measure the distance between the vehicle and obstacles and road boundaries by reading the grid map drawn by the map drawing device, and send control instructions to the vehicle through the communication device.

8. The vehicle driving area detection system according to claim 7, characterized in that: The communication device includes a CAN communication data receiving module and a laser radar data receiving module; A CAN communication data receiving module is used to parse the vehicle wheel speed information in the vehicle CAN data and input the parsed vehicle wheel speed information into the map drawing device; The laser radar data receiving module is used to receive data collected by the laser radar and input the data into the map drawing device.

Images