Multi-laser fusion perception method for outdoor cleaning vehicle

Abstract

The application provides a multi-laser fusion sensing method for an outdoor cleaning vehicle, comprising the following steps: setting the obstacle detected by laser radar 1 as accurate, observing whether the radar point cloud of the obstacle detected by laser radar 2 coincides with the obstacle detected by laser radar 1; performing radar parameter calibration, manually modifying the offset value of laser radar 2, so that the radar point cloud of the obstacle detected by laser radar 2 coincides with the obstacle detected by laser radar 1; writing the above parameters into a system file, reading the parameters in the system file into the driving node of laser radar 2 each time the driving node is started, and converting the data in the driving node, so that the data of laser radar 1 and laser radar 2 coincide; the data of laser radar 1 and laser radar 2 does not need to be processed, and is directly supplied to the navigation node of the outdoor cleaning vehicle, so as to control the obstacle avoidance and movement function of the outdoor cleaning vehicle. The method does not need post-processing of laser data, has low cost, and is simple to deploy.

Description

Technical Field

[0001] This invention relates to a multi-laser fusion sensing method for outdoor cleaning vehicles, applicable to the field of unmanned outdoor cleaning vehicles. Background Technology

[0002] As is well known, LiDAR is an optical remote sensing technology that measures distance and spatial information by emitting pulsed lasers towards a target. It is widely used in fields such as autonomous driving, mapping, and construction engineering. On unmanned sanitation vehicles, LiDAR is used for mapping and navigation. Due to the structural limitations of the sanitation vehicle itself, a single LiDAR cannot detect the entire surrounding environment. Therefore, multiple LiDARs need to be installed in different locations on the sanitation vehicle to jointly detect the surrounding environment. To ensure the accuracy of the detection, the data from multiple LiDARs needs to be fused.

[0003] Currently, the method of fusing LiDAR data in unmanned cleaning and sanitation vehicles that has been developed and produced is quite complex. After acquiring data from different LiDAR sources, the data undergoes coordinate transformation and other data processing based on the LiDAR locations before being fused together and distributed to other nodes, such as navigation nodes. LiDAR needs to be installed at both the vehicle and roadside, resulting in high costs and complex deployment; furthermore, the subsequent processing of the LiDAR data is also very complex.

[0004] Therefore, it is necessary to design a new multi-laser fusion sensing method for outdoor cleaning vehicles to overcome the above problems. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a multi-laser fusion sensing method for outdoor cleaning vehicles. By using manually calibrated laser radar position parameters, the data of multiple laser radars are fused, eliminating the need for laser data post-processing, resulting in lower costs and simpler deployment.

[0006] This invention is implemented as follows: This invention provides a multi-laser fusion sensing method for outdoor cleaning vehicles, comprising the following steps: Step 1: Set the obstacle detected by LiDAR 1 to be accurate, and observe whether the radar point cloud of the obstacle detected by LiDAR 2 overlaps with that of the obstacle detected by LiDAR 1. If the radar point clouds do not overlap, it proves that the parameters of LiDAR 2 are incorrect, and the position parameters of LiDAR 2 relative to the outdoor cleaning vehicle need to be modified and calibrated. Proceed to the following steps. Step 2: Perform radar parameter calibration. Manually modify the offset values ​​x, y, z and offset angles r, p, θ of LiDAR 2 relative to the outdoor cleaning vehicle in the X, Y, and Z axes. Adjust the above six degrees of freedom so that the radar point cloud of the obstacle detected by LiDAR 2 coincides with that of the obstacle detected by LiDAR 1. Step 3: Write the above parameters into the system file. Each time the driver node of LiDAR 2 starts, it reads the parameters from the system file and converts the data in the driver node. At this time, the data of LiDAR 1 and LiDAR 2 are the same. Step 4: The data from LiDAR 1 and LiDAR 2 do not need to be processed and are directly supplied to the navigation nodes of the outdoor cleaning vehicle to control the obstacle avoidance and movement functions of the outdoor cleaning vehicle.

[0007] Furthermore, in step one, if the obstacle detected by lidar 2 completely overlaps with the radar point cloud of the obstacle detected by lidar 1, then there is no need to calibrate the radar parameters, and the radar data can be used directly.

[0008] Furthermore, after the radar parameters are calibrated, the radar point cloud data from multiple lidars can be added and fused together to form a single data frame, which can be used by navigation nodes for obstacle avoidance or path planning.

[0009] Furthermore, in step three, the driving node of lidar 2 is an application that reads data from lidar.

[0010] Furthermore, if multiple lidars are included, the parameter calibration process of lidar 2 is repeated for each lidar.

[0011] Furthermore, The present invention has the following beneficial effects: The multi-laser fusion sensing method for outdoor cleaning vehicles provided by this invention manually calibrates the position of the lidar, records relevant parameters after calibration, and reads the parameters during program execution. It can directly output the lidar data without subsequent fusion processing. The data from multiple lidars can also ensure 360-degree obstacle detection without blind spots. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 A schematic diagram of obstacles detected by lidar 1 and lidar 2 according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the laser data flow provided in an embodiment of the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] This invention provides a multi-laser fusion sensing method for outdoor cleaning vehicles, applied to outdoor cleaning vehicles. The outdoor cleaning vehicle has multiple lidars, defined here as lidar 1 and lidar 2; wherein the obstacle detection of lidar 1 is considered accurate, and the data detected by lidar 2 is used to compare with the data detected by lidar 1; if multiple other lidars are also included, each lidar repeats the parameter calibration process of lidar 2.

[0016] Specifically, the multi-laser fusion sensing method of this outdoor cleaning vehicle includes the following steps: Step 1: Assume the obstacle detected by LiDAR 1 is accurate. Observe whether the radar point clouds of the obstacle detected by LiDAR 2 overlap with those of the obstacle detected by LiDAR 1. If the radar point clouds do not overlap, it proves that the parameters of LiDAR 2 are incorrect (this parameter assumes that the position of LiDAR 1 relative to the outdoor cleaning vehicle is accurate, and at this time, the position parameters of LiDAR 2 relative to the vehicle are incorrect). The position parameters of LiDAR 2 relative to the outdoor cleaning vehicle need to be modified and calibrated before proceeding to Step 2. If the radar point clouds of the obstacle detected by LiDAR 2 completely overlap with those of the obstacle detected by LiDAR 1, then there is no need to calibrate the radar parameters, and the radar data can be used directly.

[0017] Step 2: Perform radar parameter calibration. Manually modify the offset values ​​x, y, z and offset angles r, p, θ of LiDAR 2 relative to the outdoor cleaning vehicle in the X, Y, and Z axes. Adjust the above six degrees of freedom (these six degrees of freedom can change the position of the LiDAR point cloud in space, and these six degrees of freedom are techniques known to those skilled in the art) so that the obstacles detected by LiDAR 2 coincide with the obstacles detected by LiDAR 1, that is, the radar point clouds detected by the two coincide. Step 3: Write the above parameters into the system file (control system of the outdoor cleaning vehicle). The drive node of LiDAR 2 (the drive node of LiDAR 2 is the application that reads data from the LiDAR) reads the parameters from the system file each time it starts and converts the data in the drive node. At this time, the data of LiDAR 1 and LiDAR 2 are the same. Step 4: The data from LiDAR 1 and LiDAR 2 do not require processing and are directly supplied to the navigation node of the outdoor cleaning vehicle to control its obstacle avoidance and movement functions. The fused laser data is output to the motion planning module through the vehicle's built-in sensing and positioning modules, thereby completing the vehicle's obstacle avoidance and movement functions.

[0018] Specifically, after the radar parameters are calibrated, the radar point cloud data from multiple lidars can be added and fused together to form a single data frame, which can then be used by the navigation node (the navigation control system built into the outdoor cleaning vehicle) for obstacle avoidance or path planning.

[0019] The following is a brief explanation using examples, such as... Figure 1 As shown, lidar1 represents lidar 1, and lidar2 represents lidar 2. The obstacles detected by lidar 1 and lidar 2 do not overlap. Lidar 2 needs to be moved by distances s1 and s2 along the X-axis and Z-axis respectively to make them overlap. s1 and s2 are the calibration parameters. If the offset values ​​x, y, and z in the X-axis, Y-axis, and Z-axis directions, and the offset angles r, p, and θ in the Z-axis directions need to be modified to merge the two, then they are all calibration parameters; otherwise, the corresponding parameters are 0. The calibration process is as follows: the data detected by LiDAR 1 is defined as the raw data of LiDAR 1, and the data detected by LiDAR 2 is defined as the raw data of LiDAR 2. The two do not overlap. The raw data of LiDAR 2 is calibrated. After calibration and conversion, the fused laser data of the two is obtained. Then, it is output to the motion planning module through the vehicle's built-in perception module, positioning module, etc., thereby completing the vehicle's obstacle avoidance and motion functions.

[0020] After calibration, point cloud data from multiple LiDARs can be simply added and fused to form a single data frame, which can then be used by navigation nodes for obstacle avoidance or path planning. The entire data flow is as follows: Figure 2 As shown.

[0021] In summary, the multi-laser fusion sensing method for outdoor cleaning vehicles provided by this invention manually calibrates the position of the lidar, records relevant parameters after calibration, and reads these parameters during program execution. It can directly output lidar data without subsequent fusion processing. Data from multiple lidars can ensure 360-degree obstacle detection without blind spots. It is low-cost and relatively simple to deploy.

[0022] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-laser fusion sensing method for outdoor cleaning vehicles, characterized in that, Includes the following steps: Step 1: Set the obstacle detected by LiDAR 1 to be accurate, and observe whether the radar point cloud of the obstacle detected by LiDAR 2 overlaps with that of the obstacle detected by LiDAR 1. If the radar point clouds do not overlap, it proves that the parameters of LiDAR 2 are incorrect, and the position parameters of LiDAR 2 relative to the outdoor cleaning vehicle need to be modified and calibrated. Proceed to the following steps. Step 2: Perform radar parameter calibration. Manually modify the offset values ​​x, y, z and offset angles r, p, θ of LiDAR 2 relative to the outdoor cleaning vehicle in the X, Y, and Z axes. Adjust the above six degrees of freedom so that the radar point cloud of the obstacle detected by LiDAR 2 coincides with that of the obstacle detected by LiDAR 1. Step 3: Write the above parameters into the system file. Each time the driver node of LiDAR 2 starts, it reads the parameters from the system file and converts the data in the driver node. At this time, the data of LiDAR 1 and LiDAR 2 are the same. Step 4: The data from LiDAR 1 and LiDAR 2 do not need to be processed and are directly supplied to the navigation nodes of the outdoor cleaning vehicle to control the obstacle avoidance and movement functions of the outdoor cleaning vehicle.

2. The multi-laser fusion sensing method for outdoor cleaning vehicles as described in claim 1, characterized in that: In step one, if the obstacle detected by lidar 2 completely overlaps with the radar point cloud of the obstacle detected by lidar 1, then there is no need to calibrate the radar parameters, and the radar data can be used directly.

3. The multi-laser fusion sensing method for outdoor cleaning vehicles as described in claim 1, characterized in that: After the radar parameters are calibrated, the radar point cloud data from multiple lidars can be added and fused together to form a single data frame, which can be used by navigation nodes for obstacle avoidance or path planning.

4. The multi-laser fusion sensing method for outdoor cleaning vehicles as described in claim 1, characterized in that: In step three, the driving node of LiDAR 2 is the application that reads data from the LiDAR.

5. The multi-laser fusion sensing method for outdoor cleaning vehicles as described in claim 1, characterized in that: If multiple lidars are included, then each lidar will repeat the parameter calibration process of lidar 2.

Images