A multi-vehicle collaborative loading path planning method in open-pit mines based on roadside guidance

Abstract

The present invention relates to a method for collaborative loading path planning of multiple vehicles in an open-pit mine based on roadside guidance, which belongs to the field of multi-agent path planning and solves the problems in the prior art of low efficiency, long conflict search time and inability to resolve conflicts within the required time when multiple agents are simultaneously planning paths. The present invention establishes a search tree based on the original starting point and end point, selects transition points based on the distribution of obstacles, creates a third random search tree, and introduces roadside equipment for collision detection and re-planning the path. The method of the present invention is used to refine the multi-vehicle path in stages, which takes less time to resolve conflicts, has better real-time performance in re-planning the path, and has high path planning efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of multi-agent path planning, and relates to a multi-vehicle collaborative loading path planning method in an open-pit mine based on roadside guidance. Background Art

[0002] Path planning can be divided into global path planning and local path planning. Common path planning steps are mainly divided into environment modeling, path search and path smoothing. Environment modeling is to abstract the space we live in. Correct space modeling is an important basis for path search; path search is based on environment modeling to plan the path; path smoothing is to achieve the feasibility of the search path and provide a more reasonable solution. The global path planning method for complex multi-agent cargo loading is the focus of current research.

[0003] At present, the commonly used methods include the enhanced conflict-based search algorithm, referred to as the ECBS algorithm. The upper layer looks for conflicts, and the lower layer is responsible for replanning the path. Compared with the CBS algorithm, a focused search is added. The upper layer finds nodes with fewer collisions to expand, so that the lower layer can find suboptimal solutions faster.

[0004] However, in practice, the above method has at least the following problems:

[0005] 1. When the environment is crowded and there are many agents, path planning takes a long time or even fails.

[0006] 2. The planning process cannot be terminated for situations that cannot be planned;

[0007] 3. The planning process should be caused by the agent's own actions, which may lead to path planning errors;

[0008] 4. It is easy to fall into local search during the path planning process, which increases the path planning cost.

[0009] 5. During the expansion process, the lower bound of the optimal solution of the upper layer hardly increases, which makes it impossible to ensure that the entire planning process can be completed within a reasonable time frame.

[0010] 6. Only existing multiple agents can be routed, and new agents are not allowed to be added. Summary of the invention

[0011] In view of the above problems, the present invention provides a multi-vehicle collaborative loading path planning method for an open-pit mine based on roadside guidance, so as to solve the problems that conflicts are prone to occur when multiple intelligent agents run simultaneously in the existing methods, the planning process is prone to local jamming, and the planning cannot be completed within a reasonable time.

[0012] The present invention provides a method for planning a multi-vehicle collaborative loading path in an open-pit mine based on roadside guidance, and the specific steps are as follows:

[0013] Step 1: Divide the driving stages according to the initial path of each vehicle, and set the starting point and end point for each driving stage;

[0014] Step 2: Set the transition point for each driving phase:

[0015] The starting point and the end point of each driving stage are connected to obtain the corresponding stage connection line;

[0016] The transition point generation method is as follows: the transition point is generated based on the stage line. If there is no obstacle on the stage line, the center point of the stage line is used as the transition point. If there is an obstacle on the stage line, the obstacle is divided into left and right areas with the stage line as the boundary, and the total area of ​​obstacles in the left area S is calculated. 左障 The total area of ​​obstacles on the right side is S 右障 , if S 左障 ≥S 右障 , select a point N car body lengths away from the vertical point on the right side of the perpendicular bisector of the connecting line at this stage as the transition point. 右障 ≥S 左障 , select a point N car body lengths away from the vertical point on the left side of the perpendicular bisector of the connecting line at this stage as the transition point;

[0017] Step 3, using the starting point, transition point, and end point of each driving stage as the root nodes of three random search trees, the process path of each vehicle is generated based on the RRT random sampling method;

[0018] Step 4, setting roadside equipment for the area corresponding to each driving stage; each roadside equipment extracts the process path of each vehicle passing through its detection area;

[0019] Step 5: Each roadside device performs collision detection on the extracted process path and replans the process path:

[0020] Each roadside device calculates the timestamps of each position of the process path of each vehicle extracted through its detection area, and finds the conflict points where different vehicles overlap in time and space within the detection area of ​​the same roadside device;

[0021] The driving priority is set according to the driving stage of each vehicle at the conflict point; the vehicle with high driving priority is regarded as an obstacle to replan the process path of the vehicle with low driving priority;

[0022] The vehicle with the lowest driving priority at the conflict point is selected, and based on the obstacles formed by the vehicles with high driving priority, the transition point is regenerated according to the transition point generation method, and the process path of the vehicle is replanned; collision detection is then performed until the driving path of the vehicle is obtained without conflict points; each vehicle is traversed to obtain the driving path of each vehicle; and the driving path of each vehicle is used as the final path;

[0023] Step 6: Use the final path to replace the initial path.

[0024] Optionally, the initial path of each vehicle includes an entrance and exit of the loading area, a pre-parking position, a parking position and a loading position.

[0025] Optionally, the initial path of each vehicle is divided into a first driving stage from the entrance and exit of the loading area to the pre-parking position, a second driving stage from the pre-parking position to the parking position, a third driving stage from the parking position to the loading position, a fourth driving stage from the loading position to the parking position after loading is completed, and a fifth driving stage from the parking position out of the entrance and exit of the loading area.

[0026] Optionally, a waiting area is also provided.

[0027] Optionally, for the process path of each vehicle generated based on the RRT random sampling method, redundant nodes are pruned out by pruning.

[0028] Optionally, after the process path is generated in step 3, kinematic constraints are added to each process path, vehicle curvature constraints are set, and a quadratic programming method is used to optimize the process path;

[0029] When performing step 4, the roadside equipment extracts the optimized process path of each vehicle passing through its detection area according to its respective detection area for collision detection.

[0030] Optionally, for setting the driving priority of each vehicle at the conflict point in step 5, the specific method is as follows:

[0031] Determine the driving phase of each vehicle at the conflict point;

[0032] If the vehicles are in the first driving stage or the fifth driving stage respectively, the driving priority of the vehicle in the fifth driving stage is greater than the driving priority of the vehicle in the first driving stage;

[0033] If all vehicles are in the first driving stage, the vehicle farther from the entrance or exit of the loading area has a higher driving priority;

[0034] If all vehicles are in the fifth driving stage, the vehicle closer to the entrance and exit of the loading area has a higher driving priority;

[0035] If each vehicle is in the second driving stage, the third driving stage or the fourth driving stage, the driving priority of each vehicle is as follows:

[0036] When each vehicle is in the pre-parking position, if there is a car in the parking position, it will wait in place until there is no car in the parking position and then start to enter the parking position;

[0037] If a vehicle enters one side of the loading equipment for loading operations, the vehicle on the other side waits at the parking position.

[0038] Optionally, if the car is in the waiting area, its driving priority is the lowest.

[0039] Optionally, in step 5, after obtaining the driving path, determine whether a new vehicle enters. If a new vehicle enters, allocate a process path according to the target loading position of the new vehicle; perform collision detection on the process path of the new vehicle and the driving paths of other mining trucks; if there are conflicting points, replan the current paths of all vehicles until the final path is obtained without conflicting points.

[0040] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0041] 1. The present invention adopts a distributed planning method, with the goal of achieving full map coverage within the detection range. Multiple roadside devices are installed on the roadside. Each roadside device is responsible for collision detection and path replanning of multiple mining vehicles. When replanning the path, the starting and ending positions of the conflicting driving phases are fixed, and continuity constraints are added to facilitate subsequent unified integration, thereby improving the efficiency of the overall path planning.

[0042] 2. The present invention processes the driving priorities of multiple mining trucks in stages, effectively solving the path conflicts between the multiple mining trucks.

[0043] 3. Based on the search tree established by the original starting point and end point, the present invention selects a point relatively close to the center as a transition point according to the distribution of obstacles, creates a third search tree, and sets appropriate random probabilities for searching the starting point and end point according to the map conditions to speed up the search efficiency, and adopts pruning to remove redundant nodes and optimize the path. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The drawings are only for the purpose of illustrating particular embodiments and are not to be construed as limiting the invention.

[0045] Figure 1 A flow chart of generating paths using three random search trees and a quadratic programming method based on the RRT method of the present invention;

[0046] Figure 2 A flow chart of collision detection and re-routing for the roadside equipment of the present invention;

[0047] Figure 3 is a schematic diagram of the loading area of ​​the present invention;

[0048] Figure 4 It is a diagram of the system structure of the present invention;

[0049] Figure 5 It is a schematic diagram of redundant node pruning of the present invention. DETAILED DESCRIPTION

[0050] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. In addition, the present invention can also be implemented in other ways different from those described herein, and therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0051] A specific embodiment of the present invention, as Figure 1-5 , discloses a multi-vehicle collaborative loading path planning method in an open-pit mine based on roadside guidance, the specific steps are as follows:

[0052] Step 1: Divide the driving stages according to the initial path of each mining truck, and set the starting point and end point for each driving stage;

[0053] The initial path of the mining truck includes the entrance and exit of the loading area, the pre-parking position, the parking position and the loading position.

[0054] The pre-parking position is set between the entrance and exit of the loading area and the parking position. The mining truck pre-parks at the pre-parking position to prepare for reversing, so that the mining truck is in a reversing state and drives into the parking position, thereby making the loading process more convenient.

[0055] The parking position is set between the pre-parking position and the loading position. The parking position is adjacent to the loading position and serves as a temporary parking point to ensure safety and prevent collisions. The parking position is set at a distance of one vehicle body length from the loading position.

[0056] Optionally, one loading device corresponds to two loading positions, and the two loading positions are arranged on opposite sides for easy operation of the loading device; preferably, the loading device is an electric shovel.

[0057] The initial path of the mining truck is divided into the first driving stage from the entrance and exit of the loading area to the pre-parking position, the second driving stage from the pre-parking position to the parking position, the third driving stage from the parking position to the loading position, the fourth driving stage from the loading position to the parking position after loading is completed, and the fifth driving stage from the parking position to the entrance and exit of the loading area. The starting point and end point of the corresponding driving stage are set for each driving stage; for the second, third, fourth and fifth driving stages, the end point of the previous driving stage is the starting point of the next driving stage.

[0058] Optionally, a waiting area is also provided. When the pre-parking position, the parking position and the loading position are all fully loaded, the mining truck goes to the waiting area and waits for a vacant position in the pre-parking position, the parking position or the loading position.

[0059] Step 2, setting the transition point of each driving stage of each mining truck;

[0060] Connect the starting point and the end point of each driving stage of each mining truck to obtain the corresponding first stage connection line, second stage connection line, third stage connection line, fourth stage connection line and fifth stage connection line;

[0061] If there is no obstacle on the stage line, the center point of the stage line is used as the transition point; if there is an obstacle on the stage line, the obstacle is divided into left and right areas with the stage line as the boundary, and the total area of ​​obstacles in the left area S is calculated. 左障 The total area of ​​obstacles on the right side is S 右障 , if S 左障 ≥S 右障 , select a point 1 to 3 car body lengths away from the perpendicular point of the perpendicular bisector and the line of the stage on the right side as the transition point, otherwise select the left side; preferably, select a point 2 car body lengths away from the perpendicular bisector and the line of the stage as the transition point.

[0062] Step 3, using the starting point, transition point and end point of each driving stage as the root nodes of three random search trees, the process path is generated based on the RRT (Rapid Exploration Random Tree) random sampling method;

[0063] The three random search trees include a starting point tree, a transition tree and an end point tree; the target direction of the starting point tree search is towards the transition node in the current driving stage, the target direction of the transition tree search is towards the starting point and the end point in the current driving stage, and the target direction of the end point tree search is towards the transition point in the current driving stage.

[0064] The RRT random sampling method is used for search: each random search tree extends to the target point or to a random direction with a preset step length to obtain a scattered node each time it grows. Preferably, the probability of extending in a random direction is 50%; the probability of scattering is adjusted according to the distribution and number of obstacles; the step length is set according to the size of the area in which each mining truck is currently traveling. After setting the parameters, the three random search trees in each stage begin to search until the starting point tree and the end point tree intersect with the transition point tree to generate a process path.

[0065] Preferably, due to the randomness of RRT sampling, there are redundant nodes in the generated scattered nodes, which increases the length of the path, and the redundant nodes are cut off by pruning. Figure 5As shown, the generated path includes scattered nodes P1, P2 and P3, and a search range (i.e., a dotted circle) is generated with scattered node P2 as the center and search radius r. There are two routes to choose from from the starting point S to the end point G. One is from the starting point S through segments a and b to the end point G, and the other is from the starting point S through segment c to the end point G. Therefore, node P3 is a redundant node and node P3 is cut off.

[0066] Step 4: Add kinematic constraints to each process path, set the mine card curvature constraints, and use the quadratic programming method to optimize the process path.

[0067] Preferably, the curvature constraint of the mining card is 0.15.

[0068] Step 5: Install roadside equipment in the areas corresponding to the first driving stage, the second driving stage, the third driving stage, the fourth driving stage, and the fifth driving stage of each intelligent agent obtained in step 1, and the roadside equipment extracts the process path of each mining truck passing through its detection area according to its respective detection area;

[0069] A roadside device is set in the area of ​​each driving stage; the setting position of each roadside device in the corresponding driving stage area is set according to the network base station and the transmission signal rate, and the signal transmission of each roadside device does not overlap and the signal just covers the entire working area.

[0070] The process path of each mining truck is transmitted to all roadside devices, and each roadside device extracts the process path of each mining truck passing through its detection area according to its own detection area, stores the process path data and indexes it, where the index is the mining truck number;

[0071] Step 6: Each roadside device performs collision detection on the extracted process path and replans the process path;

[0072] Each roadside device calculates the timestamp of each position of all paths of the mining truck in its respective detection area, finds out the conflict points where different mining trucks overlap in time and space during the driving phase corresponding to the roadside device, and sets the driving priority;

[0073] The method for setting the driving priority is as follows: for the conflict point, determine the driving stage of each mining truck at the point;

[0074] If the mining trucks are in the first driving stage or the fifth driving stage respectively, the mining trucks in the first driving stage are in the light-load mode driving stage of entering the loading area without loading, and the mining trucks in the fifth driving stage are in the heavy-load mode driving stage of leaving the loading area after loading. In order to avoid excessive fuel consumption and improve safety, the heavy-loaded mining trucks are not allowed to stop, and the rule of light-loaded mining trucks giving way to heavy-loaded mining trucks is implemented. Therefore, the driving priority of the mining trucks in the fifth driving stage is greater than the driving priority of the mining trucks in the first driving stage;

[0075] If all mining trucks are in the first driving stage, the farther away from the entrance and exit of the loading area, the higher the driving priority of the mining truck;

[0076] If all mining trucks are in the fifth driving stage, the closer they are to the entrance and exit of the loading area, the higher their driving priority.

[0077] If each mining truck is in the second driving stage, the third driving stage or the fourth driving stage, the driving priority of each mining truck is as follows:

[0078] (1) When a mining truck is at the pre-parking position, if there is a vehicle at the parking position, the truck will wait until there is no mining truck at the parking position before driving into the parking position;

[0079] (2) If a mining truck enters one side of the loading equipment to carry out loading operations, the mining truck on the other side waits at the parking position; thereby, it is possible to prevent the mining truck that has not yet been loaded from interfering with the operation of the loading equipment;

[0080] (3) If there is a mining truck at each pre-parking position, the mining truck that needs to enter the pre-parking position will drive to the waiting area. When the pre-parking position is vacant, the mining truck will drive from the waiting area to the pre-parking position; if there are multiple pre-parking positions, the mining truck will drive into the pre-parking position closest to it.

[0081] If the mining truck is in the waiting area, its driving priority is the lowest.

[0082] After the driving priority is set, the index of the conflicting mining truck and the conflicting driving stage information are extracted, and the mining truck with a high driving priority is set as an obstacle as a constraint when the low-priority mining truck replans the process path; enter the re-planning path module, the roadside equipment selects the mining truck with the lowest driving priority at the conflict point in its detection area, and re-plans the process path of the mining truck with the lowest driving priority according to the method of steps 2-4 based on the obstacles formed by the vehicles with high driving priority, and enters step 6 for collision detection until the driving path of the vehicle is obtained without conflict points, that is: no collision; for the mining truck that does not need to re-plan the process path, its process path is its driving path, thereby obtaining the driving path of each mining truck, and entering step 7;

[0083] Optionally, in order to satisfy the continuity constraint, the positions and orientation angles of the starting point and the end point of the driving phase at the conflicting location are fixed and then replanned.

[0084] Step 7, determine whether there is a new mining truck entering. If there is a new mining truck entering, allocate a process path according to the target loading position of the new mining truck, and divide the process path of the new mining truck into driving stages; each roadside device performs collision detection on the process path of the new mining truck and the driving paths of other mining trucks according to the method in step 6. If there is a collision, replan the current paths of all mining trucks and then perform collision detection and path replanning, and iterate repeatedly until the final path is obtained without conflict points; if there is no new mining truck entering and the final path is obtained, proceed to step 8;

[0085] Step 8, pass the final path to the corresponding mining card to replace the initial path;

[0086] Step 9: The mining truck drives along the final route.

[0087] Another specific embodiment of the present invention discloses a mine area multi-vehicle collaborative loading path planning system based on roadside guidance, using the aforementioned path planning method, including a planning module, a roadside equipment control module and;

[0088] The planning module includes a front-end module for generating three random search trees;

[0089] The roadside equipment is used to perform collision detection on the path based on generating three random search trees and generate a new planned path to finally obtain the final path;

[0090] The control module is used to control multiple unmanned mining trucks to travel along the re-planned routes.

[0091] Optionally, the planning module further includes a back-end module for generating three random search trees and / or performing secondary optimization on the new planned path.

[0092] Optionally, the roadside equipment includes a path extraction module, a collision detection module and a re-planning module; the path extraction module is used to extract and generate three random search trees, the collision detection module is used to perform collision tests on the path, and the re-planning module is used to re-plan the path and obtain the final path.

[0093] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for multi-vehicle collaborative loading path planning in an open-pit mine based on roadside guidance, characterized in that: The specific steps are as follows: Step 1: Divide the driving stages according to the initial path of each vehicle, and set the starting point and end point for each driving stage; Among them, the initial path of each vehicle includes the entrance and exit of the loading area, the pre-parking position, the parking position and the loading position; The initial path of each vehicle is divided into a first driving stage from the entrance and exit of the loading area to the pre-parking position, a second driving stage from the pre-parking position to the parking position, a third driving stage from the parking position to the loading position, a fourth driving stage from the loading position to the parking position after loading is completed, and a fifth driving stage from the parking position to the entrance and exit of the loading area; Step 2: Set the transition point for each driving phase: The starting point and the end point of each driving stage are connected to obtain the corresponding stage connection line; The transition point generation method is as follows: the transition point is generated based on the stage line. If there is no obstacle on the stage line, the center point of the stage line is used as the transition point. If there is an obstacle on the stage line, the obstacle is divided into left and right areas with the stage line as the boundary, and the total area of ​​obstacles in the left area S is calculated. 左障 The total area of ​​obstacles on the right side is S 右障 , if S 左障 ≥S 右障 , select a point N car body lengths away from the vertical point on the right side of the perpendicular bisector of the connecting line at this stage as the transition point. 右障 ≥S 左障 , select a point N car body lengths away from the vertical point on the left side of the perpendicular bisector of the connecting line at this stage as the transition point; Step 3, using the starting point, transition point, and end point of each driving stage as the root nodes of three random search trees, the process path of each vehicle is generated based on the RRT random sampling method; Step 4, setting roadside equipment for the area corresponding to each driving stage; each roadside equipment extracts the process path of each vehicle passing through its detection area; Step 5: Each roadside device performs collision detection on the extracted process path and replans the process path: Each roadside device calculates the timestamps of each position of the process path of each vehicle extracted through its detection area, and finds the conflict points where different vehicles overlap in time and space within the detection area of ​​the same roadside device; The driving priority is set according to the driving stage of each vehicle at the conflict point; the vehicle with high driving priority is regarded as an obstacle to replan the process path of the vehicle with low driving priority; The vehicle with the lowest driving priority at the conflict point is selected, and based on the obstacles formed by the vehicles with high driving priority, the transition point is regenerated according to the transition point generation method, and the process path of the vehicle is replanned; collision detection is then performed until the driving path of the vehicle is obtained without conflict points; each vehicle is traversed to obtain the driving path of each vehicle; and the driving path of each vehicle is used as the final path; Among them, the specific method of setting the driving priority of each vehicle at the conflict point is as follows: Determine the driving phase of each vehicle at the conflict point; If the vehicles are in the first driving stage or the fifth driving stage respectively, the driving priority of the vehicle in the fifth driving stage is greater than the driving priority of the vehicle in the first driving stage; If all vehicles are in the first driving stage, the vehicle farther from the entrance or exit of the loading area has a higher driving priority; If all vehicles are in the fifth driving stage, the vehicle closer to the entrance and exit of the loading area has a higher driving priority; If each vehicle is in the second driving stage, the third driving stage or the fourth driving stage, the driving priority of each vehicle is as follows: When each vehicle is in the pre-parking position, if there is a car in the parking position, it will wait in place until there is no car in the parking position and then start to enter the parking position; If a vehicle enters one side of the loading equipment for loading operations, the vehicle on the other side waits at the parking position; Step 6: Use the final path to replace the initial path.

2. The path planning method according to claim 1, characterized in that: There is also a waiting area.

3. The path planning method according to any one of claims 1 to 2, characterized in that: For the process path of each vehicle generated based on the RRT random sampling method, redundant nodes are trimmed by pruning.

4. The path planning method according to claim 1, characterized in that: After the process path is generated in step 3, kinematic constraints are added to each process path, vehicle curvature constraints are set, and a quadratic programming method is used to optimize the process path; When performing step 4, the roadside equipment extracts the optimized process path of each vehicle passing through its detection area according to its respective detection area for collision detection.

5. The path planning method according to claim 4, characterized in that: If the car is in the waiting area, it has the lowest driving priority.

6. The path planning method according to claim 1, characterized in that: In step 5, after obtaining the driving path, determine whether a new vehicle enters. If a new vehicle enters, allocate a process path according to the target loading position of the new vehicle; perform collision detection on the process path of the new vehicle and the driving paths of other mining trucks; if there are conflicting points, replan the current paths of all vehicles until there are no conflicting points and the final path is obtained.

Images