A system and method for trajectory planning and edge cleaning of an autonomous cleaning vehicle

Abstract

A kind of automatic driving cleaning car trajectory planning and system and method of edge cleaning, wherein, system includes positioning and map loading module, perception module, trajectory planning module and automatic driving control module.The specific method steps of the automatic driving cleaning car trajectory planning and edge cleaning are as follows: the generation of edge trajectory, the generation of intermediate trajectory, the identification processing of different trajectories and the processing of around obstacle decision planning.The present application adds road edge marker in road edge trajectory section by the way of pre-processing trajectory line data file when mapping, and the system distinguishes road edge trajectory and intermediate trajectory by identifying marker.When the system judges that vehicle is located in different trajectory, the corresponding around obstacle rule is dynamically changed, and the rule is changed in road edge trajectory to adapt to road edge working environment, and the edge cleaning efficiency and vehicle safety are considered.Not need to increase additional sensor to identify curbstone or road edge feature obstacle, cost is low, and environmental applicability is also high.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and more specifically to a system and method for autonomous cleaning vehicle trajectory planning and edge cleaning. Background Technology

[0002] The function of autonomous sweepers is to replace manually driven sweepers and complete road cleaning tasks. To ensure thorough cleaning, in addition to considering the road coverage of the sweeping trajectory, special attention must be paid to the cleaning effect along the curbs, as debris tends to accumulate. However, since curbs are often low, vehicles have difficulty automatically recognizing them, and getting too close poses a collision risk. This necessitates balancing obstacle avoidance and cleaning efficiency in trajectory planning, ensuring both collision avoidance and the safety of autonomous driving, while maximizing cleaning effectiveness.

[0003] Currently, among the publicly available autonomous cleaning vehicle solutions, such as the Chinese invention patent with application publication number CN115248447A, which discloses a curb recognition method and system based on laser point clouds, this patent mainly relies on lidar to identify curbs or roadside obstacles (such as flowers and shrubs) to balance edge-following cleaning efficiency and vehicle safety. However, this recognition method is difficult and costly, and has low environmental applicability. Therefore, we provide a system and method for autonomous cleaning vehicle trajectory planning and edge-following cleaning. Summary of the Invention

[0004] This invention provides a system and method for trajectory planning and edge cleaning of an autonomous cleaning vehicle, which overcomes the limitations of existing autonomous cleaning vehicles that rely on LiDAR to identify curbs or roadside obstacles, which is difficult, costly, and has low environmental applicability.

[0005] The present invention adopts the following technical solution:

[0006] A system for autonomous cleaning vehicle trajectory planning and edge cleaning includes: a positioning and map loading module: providing accurate positioning and basic trajectory for the vehicle; a perception module: identifying obstacles around the vehicle; a trajectory planning module: calculating the optimal path based on the basic trajectory and surrounding obstacles; and an autonomous driving control module: controlling the vehicle to move along the optimal path.

[0007] This invention also provides a method for trajectory planning and edge cleaning of an autonomous cleaning vehicle, using the above system, with the following specific steps:

[0008] S1. Generation of edge-following trajectory: Manually drive the vehicle to sweep around the curb, record the manual edge-following sweeping path as an edge-following vector map and save it. When the cleaning vehicle drives automatically, it uses the pre-recorded edge-following vector map to achieve edge-following sweeping.

[0009] S2. Generation of the intermediate trajectory: The edge - attached vector map includes several pre - recorded trajectory lines arranged at intervals in the road width direction. The distance m between adjacent pre - recorded trajectory lines is less than the vehicle width w. Calculate the number n of intermediate trajectory lines, satisfying the following relationship: n=(p - 2w) / (w - k), where p is the width at the widest part of the road, k is the overlapping width, 0 < k < w, and n is rounded up.

[0010] S3. Identification and processing of different trajectories: By pre - processing the data file of the pre - recorded trajectory lines during mapping, add edge - attached marks to the roadside trajectory segments. The system distinguishes the edge - attached trajectory and the intermediate trajectory by identifying the marks.

[0011] S4. Processing of obstacle - avoidance decision - making and planning: Judge in real time whether the vehicle reaches the edge - attached trajectory through the marks, and dynamically change the obstacle - avoidance decision - making and planning, including the virtual body side margin of the vehicle and the distance between the optimal path and the obstacle, and limit the drivable path on one side of the roadside.

[0012] Preferably, the data of the above - mentioned edge - attached vector map includes the serial numbers and position coordinate information of each point.

[0013] In a preferred implementation, in the above step S3, the pre - processing process of the pre - recorded trajectory line data file includes designing a script, selecting the edge - attached trajectory, adding roadside marks at one key, and writing into the map file.

[0014] In a preferred implementation, in the above step S4, the selection of the optimal path by the vehicle needs to meet the prerequisite L≥h + j, where L is the distance between the selected optimal path and the obstacle, h is the virtual body side margin, and j is the distance allowing the obstacle to approach the virtual body; block the path with L < h + j; when L < h, block all paths, that is, once the obstacle enters the virtual body range, all paths are blocked and the vehicle stops.

[0015] In a preferred implementation, when on the intermediate trajectory, set the virtual body side margin as h1 and the distance allowing the obstacle to approach the virtual body as j1; when on the edge - attached trajectory, set the virtual body side margin as h2 and the distance allowing the obstacle to approach the virtual body as j2; satisfy w / 2 + d≤h2 < h1, 0≤j2 < j1, where w is the actual vehicle width and d is the edge - attached gap.

[0016] When there is no obstacle on the roadside side of the edge - attached section and there is an obstacle on the non - roadside side, normally block the roadside - side path of the edge - attached trajectory, and make the vehicle bypass the obstacle to the non - roadside side.

[0017] As can be seen from the above description of the present invention, compared with the prior art, the present invention has the following advantages:

[0018] This invention preprocesses the trajectory line data file during mapping, adding curb markers to the curb trajectory segments. The system distinguishes between curb trajectories and intermediate trajectories by recognizing these markers. When the system determines that a vehicle is on a different trajectory, it dynamically adjusts the corresponding obstacle avoidance rules, modifying the rules along the curb trajectory to adapt to the curb working environment, balancing edge cleaning efficiency and vehicle safety. It eliminates the need for additional sensors to identify curbs or roadside obstacles, resulting in low cost and high environmental adaptability. Attached Figure Description

[0019] Figure 1 This is a structural block diagram of the present invention.

[0020] Figure 2 This is a flowchart of the present invention.

[0021] Figure 3 This is a vector map recorded from the manual edge-sweeping trajectory of this invention.

[0022] Figure 4 This is a map file preprocessed for the trajectory line data file of this invention.

[0023] Figure 5 This is a schematic diagram illustrating the optimal path selection for the trajectory line of the present invention.

[0024] Figure 6 This is a schematic diagram of the vehicle of the present invention recognizing obstacles on a trajectory line. Detailed Implementation

[0025] The following reference Figure 1 Specific embodiments of the present invention will be described below. Many details are described below to provide a comprehensive understanding of the invention; however, those skilled in the art can implement the invention without these details. Well-known components, methods, and processes will not be described in detail hereafter.

[0026] This invention relates to a system for trajectory planning and edge cleaning of an autonomous cleaning vehicle, with reference to... Figure 1 It includes a positioning and map loading module, a perception module, a trajectory planning module, and an autonomous driving control module. The main functions of each module are as follows:

[0027] (1) Positioning and map loading module: Provides accurate positioning and basic trajectory for vehicles.

[0028] (2) Perception module: Identify obstacles around the vehicle.

[0029] (3) Trajectory planning module: Calculates the optimal path based on the basic trajectory and surrounding obstacles.

[0030] (4) Automatic driving control module: controls the vehicle to move along the optimal path.

[0031] Based on the above system architecture, the present invention also provides a method for trajectory planning and edge cleaning of an autonomous cleaning vehicle, referring to Figure 2 , including:

[0032] Step 1: Generation of the edge trajectory.

[0033] To ensure that the edge cleaning effect is closest to the manual cleaning effect, drive the vehicle manually along the road edge for one circle and record the manual edge cleaning trajectory as a vector map for storage, referring to Figure 3 . The pre-recorded trajectory lines 1 and 2 are the trajectories generated by manual edge cleaning, and the vehicle tracks the trajectory lines during autonomous driving. The vector map data includes information such as the serial number and position coordinates of each point.

[0034] Step 2: Generation of the middle trajectory.

[0035] To ensure that the road surface is comprehensively covered and cleaned and the garbage removal rate is improved, the distance m between each pre-recorded trajectory line should be less than the vehicle width w. Considering that the actual lane widths are not uniformly equal, the number n of the middle trajectory lines depends on the width p of the widest part of the lane, and is related to the vehicle width w and the overlapping width k, satisfying the following relationship: n = (p - 2w) / (w - k), 0 < k < w, and n is rounded up.

[0036] For example: the road width p of the widest part of the lane is 5 meters, the vehicle width w is 1.5 meters, and the overlapping width k is set to 0.4 meters. Then the number of middle trajectories n = (p - 2w) / (w - k) = (5 - 3) / (1.5 - 0.4) ≈ 1.8, rounded up to be 2. As Figure 3 shown, the pre-recorded trajectory lines 3 and 4 are the middle trajectories.

[0037] Step 3: Identification and processing of different trajectories.

[0038] Since the features of the road edge are mostly low road curbs and there are also many types of obstacles on the road edge. The present invention designs a script by preprocessing the data file of the pre-recorded trajectory line during mapping, selects the edge trajectory, adds the road edge mark at one key, and writes it into the map file. As Figure 4 shown, Lane Mode 1 is the edge mark. The system reads the map file and distinguishes the road edge trajectory and the middle trajectory by identifying the mark. When the system determines that the vehicle is located on the edge trajectory, the corresponding obstacle avoidance rule is dynamically changed.

[0039] Step 4: Processing of obstacle avoidance decision planning.

[0040] As Figure 5As shown, assuming each pre-recorded trajectory line has selectable paths ①②③④⑤, the vehicle's selection of the optimal path must satisfy the prerequisite L≥h+j. When there are no obstacles, the middle path ③ is selected by default; paths L<h+j are blocked; and when L<h, all paths are blocked, meaning that once an obstacle enters the virtual vehicle's range, all paths are blocked, and the vehicle stops. Figure 6 In the diagram, L represents the distance from the selected optimal path to the obstacle; h represents the side distance of the virtual vehicle body; j represents the distance at which the obstacle is allowed to approach the virtual vehicle body; w represents the actual vehicle width; and d represents the edge gap.

[0041] In the middle trajectory, since surrounding obstacles are usually dynamic with complex types and trajectories, the autonomous sweeper must maintain a greater safe distance from identified obstacles to minimize collisions. This is achieved by setting a larger virtual vehicle side distance h1 and a larger allowable distance j1 for obstacles to approach the virtual vehicle. Furthermore, since there is generally less garbage accumulation at non-curb areas, obstacles are usually avoided. This obstacle avoidance reduces the cleaning efficiency. Depending on the actual situation, the garbage removal rate can be improved by increasing the cleaning overlap area k of each trajectory.

[0042] In the edge-following trajectory, since the obstacles at the roadside are mostly static, such as flowers, grass, and shrubs, the collision risk is relatively low. A small virtual vehicle side distance h2 is set, and a large allowable distance j2 for obstacles to approach the virtual vehicle is set. To prevent situations where there are no obstacles on one side of the roadside but obstacles on the other side, if the obstacle avoidance rules of the intermediate trajectory are applied, the vehicle will veer towards the roadside, risking hitting the curb. By permanently blocking the roadside path of the edge-following trajectory, the vehicle can only veer towards the non-roadside, thus avoiding this risk. The purpose is to enable the autonomous sweeper to clean along the edge while ensuring safe obstacle avoidance.

[0043] The above conditions must be met: w / 2+d≤h2

[0044] For example: When the vehicle is on the edge-hugging trajectory line 1: The side paths ④ and ⑤ along the roadside are normally blocked. The edge gap d = 0.1 meters is set, the distance allowed for obstacles to approach the virtual vehicle is set to j2 = 0 meters, and the side distance of the virtual vehicle is set to h2 = w / 2 + d = 0.5 + 0.1 = 0.6 meters. Then, the path is blocked when L < h2 + j2 = 0.6 meters. That is, when the distance of the available path from the obstacle is less than or equal to 0.6 meters, the path is blocked.

[0045] like Figure 6 As shown, when obstacle 1 is detected at vehicle trajectory line 1, the distance between the middle path ③ and obstacle 1 is 0.6 meters, which satisfies L≥h2+j2=0.6 meters. When the middle path is feasible, the default optimal trajectory is the middle path ③, and the vehicle goes straight without bypassing the obstacle.​

[0046] When the vehicle continues to move forward and obstacle 2 is recognized, the distance between the middle path ③ and obstacle 2 is 0.5 meters, which satisfies L < h2 + j2 = 0.6 meters. This path is blocked. By analogy, the optional optimal trajectory is path ②, and the vehicle bypasses the obstacle to the left.

[0047] When the vehicle continues to move forward and obstacle 3 is recognized, since ①②③ all satisfy L < h2 + j2 = 0.6 meters, ①②③ are blocked, and the roadside paths ④⑤ are normally blocked. The vehicle stops and will not bypass the obstacle to the right and hit the curb.

[0048] When the vehicle is on the middle trajectory 4: h1 needs to be greater than h2. Set h1 = h2 + 0.2 = 0.8 meters. Since 0 ≤ j2 < j1, set j1 = 0.2 meters. Then, when L < h1 + j1 = 1 meter, the path is blocked. That is, when the distance between the optional path and the obstacle is less than or equal to 1 meter, the path is blocked.

[0049] As shown in the figure, when the vehicle recognizes obstacle 3 at the trajectory line 4, the distance between the middle path ③ and obstacle 3 is 1.1 meters, which satisfies L ≥ 1 meter. When the middle path is feasible, the default optimal trajectory is the middle path ③, and the vehicle goes straight without bypassing the obstacle.

[0050] When the vehicle continues to move forward and obstacle 4 is recognized, since ①②③ all satisfy L < 1 meter, ①②③ are blocked. The distance between path ④ and obstacle 4 is 1.2 meters, so the selected optimal path is path ④.

[0051] The above is only the specific implementation manner of the present invention, but the design concept of the present invention is not limited thereto. Any non-substantive modification made to the present invention using this concept shall fall within the scope of infringement of the protection scope of the present invention.

Claims

1. A method for trajectory planning and edge - following cleaning of an autonomous cleaning vehicle, which uses a system for trajectory planning and edge - following cleaning of an autonomous cleaning vehicle. The system includes: Positioning and map loading module: providing accurate positioning and basic trajectories for the vehicle; Perception module: identifying obstacles around the vehicle; Trajectory planning module: calculating the optimal path based on the basic trajectory and surrounding obstacles; Autonomous driving control module: controlling the vehicle to move along the optimal path; It is characterized in that the specific steps are as follows: S1. Generation of edge - following trajectory: Manually drive the vehicle to clean along the road edge for one circle, record the manual edge - following cleaning path as an edge - following vector map and save it. When the cleaning vehicle is driving autonomously, use the pre - recorded edge - following vector map to achieve trace - following edge - following cleaning; S2. Generation of intermediate trajectory: The edge - following vector map includes several pre - recorded trajectory lines arranged at intervals in the road width direction. The distance m between two adjacent pre - recorded trajectory lines is less than the vehicle width w. Calculate the number n of intermediate trajectory lines, which satisfies the following relationship: n=(p - 2w) / (w - k), where p is the width of the widest part of the road, k is the overlapping width, 0 < k < w, and n is rounded up; S3. Identification and processing of different trajectories: By pre - processing the data file of the pre - recorded trajectory lines during mapping, add edge - following marks to the road - edge trajectory segment. The system distinguishes the edge - following trajectory and the intermediate trajectory by identifying the marks; S4. Processing of obstacle - avoidance decision planning: Real - time judge whether the vehicle reaches the edge - following trajectory through the marks, dynamically change the obstacle - avoidance decision planning, including the virtual body side margin of the vehicle and the distance between the optimal path and the obstacle, and restrict the drivable path on one side of the road edge; The selection of the optimal path by the vehicle needs to meet the prerequisite L≥h + j, where L is the distance between the selected optimal path and the obstacle, h is the virtual body side margin, and j is the distance allowing the obstacle to approach the virtual body; Block the path with L < h + j; When L < h, block all paths, that is, once the obstacle breaks into the virtual body range, all paths are blocked and the vehicle stops; When on the intermediate trajectory, set the virtual body side margin as h1 and the distance allowing the obstacle to approach the virtual body as j1; When on the edge - following trajectory, set the virtual body side margin as h2 and the distance allowing the obstacle to approach the virtual body as j2; Satisfy w / 2 + d≤h2 < h1, 0≤j2 < j1, where w is the actual vehicle width and d is the edge - following gap.

2. The method for trajectory planning and edge cleaning of an autonomous cleaning vehicle as described in claim 1, characterized in that: The data of the edge - following vector map includes the serial number and position coordinate information of each point.

3. The method for trajectory planning and edge cleaning of an autonomous cleaning vehicle as described in claim 1, characterized in that: In step S3, the pre - processing process of the pre - recorded trajectory line data file includes designing a script, selecting the edge - following trajectory, adding road - edge marks at one key, and writing into the map file.

4. The method for trajectory planning and edge cleaning of an autonomous cleaning vehicle as described in claim 1, characterized in that: When there are no obstacles on one side of the road edge in the edge - following section and there are obstacles on the non - road - edge side, normally block the path on the road - edge side of the edge - following trajectory, and make the vehicle bypass the obstacle to the non - road - edge side.

Images