Method for monitoring environmental changes by laser positioning and autonomous vehicle

Abstract

The application provides a method for monitoring environmental changes by laser positioning and an automatic driving vehicle, and belongs to the technical field of vehicle environment perception. In the automatic driving process, laser point cloud data of each grid on a positioning map is collected in real time, and the laser point cloud data of each grid is converted to a positioning map coordinate system to obtain a current matching rate corresponding to each grid. The current matching rate corresponding to the grid is compared with an original matching rate corresponding to the grid to obtain a matching rate change amount of the corresponding grid. If the matching rate change amount of a certain grid is less than a set threshold, it is considered that the driving environment around the grid has changed. According to the application, whether the driving environment near the running route changes can be judged according to the matching rate, so that the efficiency of discovering positioning jump area and processing problems in the automatic driving process is improved, and the manpower and material cost of automatic driving planning is saved.

Description

Technical Field

[0001] This invention belongs to the field of vehicle environmental perception technology and is applicable to autonomous vehicles that need to judge the driving environment. In particular, it relates to a method for monitoring environmental changes through laser positioning and an autonomous vehicle. Background Technology

[0002] Traditional methods for monitoring surface deformation mainly involve using GPS and total stations to continuously observe key locations. However, this method has a limited number of monitoring points, making it impossible to monitor the overall surface topography. Furthermore, setting up control points is difficult, resulting in a large workload and a long time to obtain results, which greatly affects monitoring efficiency.

[0003] Three-dimensional laser scanning technology, which emerged in the 1990s, can rapidly acquire the three-dimensional coordinate information of the surface of a research target with large area and high resolution through high-speed laser scanning measurement. However, it still requires the establishment of multiple measuring stations and the use of high-precision measuring instruments to achieve this effect. Currently, laser monitoring is mainly used in industrial fields such as construction sites, underground mining exploration, and meteorological and environmental monitoring, thereby enabling real-time detection of buildings or building structures.

[0004] Currently, there is no unified accuracy evaluation system for deformation monitoring data and model accuracy. Furthermore, occlusion issues encountered during laser scanning are difficult to handle, affecting the quality of deformation observations. Considering the high safety and reliability requirements of autonomous vehicles and the impact of the vehicle's driving environment on autonomous driving control, it is necessary to monitor changes in the vehicle's driving environment in real time to improve the safety of autonomous driving. However, in the field of autonomous vehicle control, LiDAR is commonly used for vehicle positioning, and no method has yet been found to monitor and judge the vehicle's driving environment using point cloud data acquired by LiDAR; there is a lack of clear measurement methods and judgment criteria. Summary of the Invention

[0005] The purpose of this invention is to provide a method for monitoring environmental changes using laser positioning and an autonomous vehicle, in order to solve the problem of difficulty in determining whether the driving environment of an autonomous vehicle has changed.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] Option 1: A method for monitoring environmental changes using laser positioning, comprising the following steps:

[0008] 1) The location map of the autonomous vehicle's operating route is rasterized; during the autonomous driving process on the operating route, the laser point cloud data of each corresponding grid is collected in real time;

[0009] 2) Convert the laser point cloud data of each grid collected in real time to the positioning map coordinate system to obtain the current matching rate. Compare the current matching rate of the corresponding grid with the original matching rate of the corresponding grid to obtain the change in the matching rate of the corresponding grid. The original matching rate is the matching rate obtained by converting the laser point cloud data collected during a normal autonomous driving process to the positioning map coordinate system.

[0010] 3) If the change in the matching rate of a certain grid is less than the set threshold, it is considered that the driving environment around that grid has changed.

[0011] This invention provides a method for monitoring environmental changes using laser positioning. During autonomous driving, it collects laser point cloud data of each grid cell on a positioning map in real time, converts this data to the positioning map coordinate system to obtain the current matching rate of each grid cell, compares the current matching rate with the original matching rate, and calculates the change in matching rate. If the change in matching rate of a grid cell is less than a set threshold, the driving environment around that grid cell is considered to have changed. Using this invention, the driving environment near the route can be determined based on the matching rate, thereby improving the efficiency of detecting and handling positioning jumps during autonomous driving and saving manpower and material costs in autonomous driving planning.

[0012] The second approach: Based on the first approach, the set threshold is calibrated by the following method: the autonomous vehicle runs multiple times in different path scenarios, and the change in the matching rate of each grid in the corresponding path scenario is collected. The set threshold is determined based on the change in the matching rate of each grid.

[0013] To facilitate the implementation of this invention, a specific method for obtaining the set threshold is proposed. Through experimental calibration, the autonomous vehicle is repeatedly driven multiple times in path scenarios covering different operating paths. During each drive in each path scenario, the change in the matching rate of each grid is collected, and the set threshold is determined based on the change in the matching rate of each grid.

[0014] The third approach: Based on the second approach, the method for determining the set threshold according to the change in the matching rate of each grid is as follows: calculate the average value of the change in the matching rate of each grid under different path scenarios as the set threshold.

[0015] The average value of the matching rate transformation of each grid under different path scenarios is calculated using mathematical statistical methods and used as the set threshold. The calculation is simple and easy to implement.

[0016] Fourth option: Based on the first option, in step 3), if the change in the matching rate of a certain grid is detected to be greater than the set threshold multiple times in a row, then the driving environment around the grid has changed.

[0017] To improve the reliability of this invention, the driving environment around a grid is considered to have changed only when the change in the matching rate of a grid is detected to be greater than a set threshold multiple times in a row.

[0018] Fifth option: Based on the fourth option, if the driving environment around multiple consecutive grids changes, then the degree of change in the driving environment around the corresponding grid is considered to have reached the set level.

[0019] If the driving environment around multiple consecutive grids changes, it is considered that the degree of change in the driving environment around the corresponding grid has reached a significant predetermined level. In this case, it is necessary to promptly alert the staff to conduct an on-site inspection of the driving environment around these grids to obtain real-world environmental information about the route, thereby better planning the autonomous driving process and preventing accidents.

[0020] Sixth option: Based on the first option, if the degree of change in the driving environment around the grid reaches a set level, the current matching rate of the grid is used as the original matching rate of the corresponding grid to update the matching rate of the corresponding grid.

[0021] If it is confirmed that there is a significant change in the driving environment around a grid, the current matching rate of that grid is promptly used as the original matching rate for that grid to update the matching rate of the corresponding grid, thereby improving the reliability of applying the present invention in the next autonomous driving process.

[0022] Seventh option: Based on the fifth option, according to the set update time interval, the latest current matching rate is used as the original matching rate of the corresponding grid to realize the matching rate update of the corresponding grid.

[0023] To ensure the timeliness of the latest environmental data, the matching rate of each grid is updated at certain time intervals. The update method is to use the latest current matching rate at the time of the update as the original matching rate of the corresponding grid.

[0024] Eighth scheme: Based on the first scheme, in step 1), the method for collecting laser point cloud data corresponding to each grid is as follows: each grid includes several feature points. When the autonomous vehicle passes through the feature points in the corresponding grid, the laser point cloud data of the corresponding feature points is collected.

[0025] In step 2), the current matching rate of the corresponding feature point in the same grid is compared with the original matching rate of the corresponding feature point to obtain the change in the matching rate of each feature point. The average value of the change in the matching rate of each feature point is calculated and used as the change in the matching rate of the corresponding grid.

[0026] To reduce the computational load of this invention, feature points are calibrated on the grid. Laser point cloud data of the corresponding feature points are collected only when the vehicle passes through these feature points, thereby obtaining the current matching rate of each feature point. In the same grid, the difference between the current matching rate of the corresponding feature point and the original matching rate of the corresponding feature point is calculated to obtain the change in the matching rate of each feature point. Then, the average value of the change in the matching rate of each feature point is calculated and used as the change in the matching rate of the corresponding grid.

[0027] Ninth option: Based on the eighth option, the feature points are the positioning points of the autonomous vehicle's operating route.

[0028] Considering that positioning points are used to mark the route when planning the autonomous driving operation, the feature points are directly adopted from the positioning points of the operation route, which facilitates the real-time operation of this invention.

[0029] Tenth solution: An autonomous vehicle includes a controller and a lidar for acquiring laser point cloud data, the lidar facing the left and right sides of the autonomous vehicle; the controller executes instructions to implement the method of monitoring environmental changes by laser positioning according to any of the above solutions. Attached Figure Description

[0030] Figure 1 This is a flowchart illustrating the method for monitoring environmental changes using laser positioning in an embodiment of the present invention. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0032] Method Implementation Examples:

[0033] The present invention provides a method for monitoring environmental changes using laser positioning. During autonomous driving, the method acquires real-time scanned laser point cloud data using lidar, converts the current laser point cloud data into a positioning map coordinate system, calculates the current matching rate of the vehicle in each grid, and compares the current matching rate of the grid with the original matching rate of the same grid in the sample using an algorithm. By comparing the difference and threshold multiple times, the method determines whether there are changes in the driving environment around the route, thereby improving the efficiency of detecting positioning jump areas and handling problems during autonomous driving, and saving the human and material costs of autonomous driving planning.

[0034] In the high-definition map, the operating route of the autonomous vehicle is planned. The route includes multiple positioning points, and the autonomous vehicle achieves autonomous driving by continuously finding and reaching the next positioning point. The operating route and its surrounding area are defined as the vehicle's driving area. Positioning map data covering the vehicle's driving area is acquired, a positioning map coordinate system is established, and the positioning map is rasterized, dividing the vehicle's driving area into different grids proportionally. Each grid includes several positioning points. In this embodiment, grids are evenly distributed within the vehicle's driving area at 2-meter intervals.

[0035] When the vehicle can operate autonomously, the location map is rasterized, dividing the vehicle's driving area into different grids proportionally. As the autonomous vehicle passes a location point within each grid, it acquires laser point cloud data of the surrounding environment using LiDAR, which serves as the raw laser point cloud data. This raw laser point cloud data is then converted to the location map coordinate system to obtain the raw matching rate. The raw matching rates for each location point are stored as samples, and these samples are updated using specific update logic, such as setting a certain update period for automatic updates, like every week or month, to ensure the timeliness of the latest environmental data.

[0036] During real-time operation, each time an autonomous vehicle passes through a grid, that grid is designated as a target grid. Using laser positioning to monitor environmental changes, the current matching rate of the vehicle when passing the target grid is calculated and compared with the original matching rate of the same grid in the sample. This change in matching rate is then compared with a set threshold to determine whether the driving environment near the target grid has changed. The driving environment refers to the physical environment outside the vehicle's driving area, such as trees and buildings. The specific judgment process is as follows: Figure 1 As shown, it includes the following steps:

[0037] S1. Using LiDAR, collect in real time the laser point cloud data of the grid to be identified when the autonomous vehicle passes through the grid to be identified.

[0038] In this embodiment, the laser point cloud data acquired when the vehicle passes through the positioning points within the grid to be determined is mainly used as the laser point cloud data of the positioning points. The laser point cloud data of the grid to be determined includes the laser point cloud data of each positioning point. As another implementation method, those skilled in the art should realize that different feature points can be set within the grid instead of positioning points.

[0039] S2. Convert the real-time acquired laser point cloud data of the raster to be identified into the positioning map coordinate system to obtain the current matching rate of the raster to be identified. Then compare the current matching rate of the raster to be identified with the original matching rate of the raster to be identified to obtain the change in the matching rate of the raster to be identified.

[0040] In this embodiment, the laser point cloud data of the positioning points in the grid to be judged is converted to the positioning map coordinate system to obtain the current matching rate of each positioning point. In the same grid, the difference between the current matching rate of the corresponding feature point and the original matching rate of the corresponding feature point is calculated to obtain the change in the matching rate of each feature point. Then, the average value of the change in the matching rate of each feature point is calculated and the average value is used as the change in the matching rate of the grid to be judged.

[0041] S3. If the change in the matching rate of the grid to be judged is less than the set threshold, it is preliminarily determined that there is no change in the driving environment around the grid to be judged.

[0042] The set threshold was calibrated through repeated trials in different scenarios. The specific trial process was as follows: Following different autonomous driving paths, the autonomous vehicle repeatedly drove along those paths multiple times in different scenarios. Each time, adjustments were made only to the traffic conditions along the route, such as changing the density or size of other vehicles around the autonomous vehicle, as a scenario adjustment. During these multiple drives, the change in matching rate for each grid was collected. Then, based on the changes in the matching rate of each grid, a mathematical statistical method was used to calculate the average of the changes in the matching rate of each grid under different path scenarios, which was then used as the set threshold. The set thresholds for other grids were calibrated using the same method.

[0043] To improve the accuracy and reliability of judging whether the surrounding environment of the driving route has changed, and to reduce the human and material costs of autonomous driving planning, considering that when an autonomous vehicle passes through a problematic grid, the presence of other vehicles on both sides can lead to inaccurate laser point cloud data acquired by the LiDAR, resulting in inaccurate final judgments, the number of times a vehicle passes through the same grid and identifies it as a problematic grid is also counted. Only when the matching rate change of a certain grid is detected multiple times consecutively and exceeds a set threshold is it considered that the driving environment around that grid has changed.

[0044] Furthermore, as the autonomous vehicle travels along the route, it assesses whether the driving environment around each grid cell has changed. After evaluating the matching rate of the driving environment around multiple consecutive grid cells, if the driving environment around these grid cells changes significantly, exceeding a certain level of change, the current matching rate of that grid cell is used as the original matching rate for that grid cell, thus updating the matching rate of that grid cell. Moreover, when significant changes occur in the driving environment around the driving route, the system promptly alerts the staff, allowing the driver to conduct on-site inspections of the driving environment around these consecutive problematic grid cells to obtain real-time environmental information about the driving route. This enables better planning of the autonomous driving process and prevents accidents.

[0045] Vehicle Example:

[0046] The present invention provides an autonomous driving vehicle with lidar installed on both sides of the vehicle body. The autonomous driving controller of the autonomous driving vehicle collects laser point cloud data in real time during the autonomous driving process through the lidar and executes instructions to realize the method of monitoring environmental changes by laser positioning in the method embodiment. The implementation of this method has been clearly described in the method embodiment and will not be repeated here.

Claims

1. A method for monitoring environmental changes using laser positioning, characterized in that, Includes the following steps: 1) The location map of the autonomous vehicle's operating route is rasterized; during the real-time operation of the autonomous vehicle on the operating route, each time the vehicle passes through a grid, the grid is used as the grid to be identified, and the laser point cloud data of the grid to be identified is collected when the vehicle passes through the grid to be identified. 2) Convert the laser point cloud data of the grid to be identified in real time to the positioning map coordinate system to obtain the current matching rate of the grid to be identified. Compare the current matching rate of the grid to be identified with the original matching rate of the corresponding grid to be identified to obtain the change in the matching rate of the corresponding grid to be identified. Under normal autonomous driving conditions, when the autonomous vehicle passes through each grid, it acquires the laser point cloud data of the surrounding environment through LiDAR as the raw laser point cloud data. The raw laser point cloud data is converted into the positioning and mapping coordinate system to obtain the raw matching rate. The raw matching rate is automatically updated at a certain update period. 3) If the matching rate change of a certain grid is detected to be greater than the set threshold multiple times in a row, it is considered that the driving environment around the grid has changed. If the driving environment around multiple grids changes in a row, the staff will be notified. In different scenarios, autonomous vehicles are repeatedly driven along the autonomous driving path multiple times, with adjustments made only to the traffic conditions on the route each time. During the multiple drives, the change in the matching rate of each grid is collected, and then the set threshold is obtained by mathematical statistical methods based on the change in the matching rate of each grid.

2. The method for monitoring environmental changes using laser positioning according to claim 1, characterized in that, Adjusting traffic conditions includes changing the density or size of other vehicles around autonomous vehicles.

3. The method for monitoring environmental changes using laser positioning according to claim 1, characterized in that, The method for determining the set threshold based on the change in the matching rate of each grid is as follows: calculate the average value of the change in the matching rate of each grid under different path scenarios as the set threshold.

4. The method for monitoring environmental changes using laser positioning according to claim 1, characterized in that, The mathematical statistical method involves calculating the average value of the change in matching rate of each grid cell under different path scenarios as a set threshold.

5. The method for monitoring environmental changes using laser positioning according to claim 1, characterized in that, If the driving environment around multiple consecutive grids changes, it is considered that the degree of change in the driving environment around the corresponding grid has reached the set level.

6. The method for monitoring environmental changes using laser positioning according to claim 5, characterized in that, If the change in the driving environment around a grid reaches a set level, the current matching rate of that grid is used as the original matching rate for that grid, thus updating the matching rate of the corresponding grid.

7. The method for monitoring environmental changes using laser positioning according to claim 1, characterized in that, According to the set update interval, the latest current matching rate is used as the original matching rate of the corresponding grid to realize the matching rate update of the corresponding grid.

8. The method for monitoring environmental changes by laser positioning according to claim 1, characterized in that, In step 1), the method for collecting laser point cloud data for each grid is as follows: each grid includes several feature points. When the autonomous vehicle passes through the feature points in the corresponding grid, the laser point cloud data of the corresponding feature points is collected. In step 2), the current matching rate of the corresponding feature point in the same grid is compared with the original matching rate of the corresponding feature point to obtain the change in the matching rate of each feature point. The average value of the change in the matching rate of each feature point is calculated and used as the change in the matching rate of the corresponding grid.

9. The method for monitoring environmental changes by laser positioning according to claim 8, characterized in that, The feature points are the location points along the operating route of the autonomous vehicle.

10. An autonomous vehicle, characterized in that, The system includes a controller and a lidar for acquiring laser point cloud data, the lidar being oriented towards the left and right sides of the autonomous vehicle; the controller executes instructions to implement the method for monitoring environmental changes by laser positioning as described in claim 1.

11. The autonomous vehicle according to claim 10, characterized in that, Adjusting traffic conditions includes changing the density or size of other vehicles around autonomous vehicles.

12. The autonomous vehicle according to claim 10, characterized in that, The method for determining the set threshold based on the change in the matching rate of each grid is as follows: calculate the average value of the change in the matching rate of each grid under different path scenarios as the set threshold.

13. The autonomous vehicle according to claim 10, characterized in that, The mathematical statistical method involves calculating the average value of the change in matching rate of each grid cell under different path scenarios as a set threshold.

14. The autonomous vehicle according to claim 10, characterized in that, If the driving environment around multiple consecutive grids changes, it is considered that the degree of change in the driving environment around the corresponding grid has reached the set level.

15. The autonomous vehicle according to claim 14, characterized in that, If the change in the driving environment around a grid reaches a set level, the current matching rate of that grid is used as the original matching rate for that grid, thus updating the matching rate of the corresponding grid.

16. The autonomous vehicle according to claim 10, characterized in that, According to the set update interval, the latest current matching rate is used as the original matching rate of the corresponding grid to realize the matching rate update of the corresponding grid.

17. The autonomous vehicle according to claim 10, characterized in that, In step 1), the method for collecting laser point cloud data for each grid is as follows: each grid includes several feature points. When the autonomous vehicle passes through the feature points in the corresponding grid, the laser point cloud data of the corresponding feature points is collected. In step 2), the current matching rate of the corresponding feature point in the same grid is compared with the original matching rate of the corresponding feature point to obtain the change in the matching rate of each feature point. The average value of the change in the matching rate of each feature point is calculated and used as the change in the matching rate of the corresponding grid.

18. The autonomous vehicle according to claim 17, characterized in that, The feature points are the location points along the operating route of the autonomous vehicle.

Images