Flow-based Global Scheduling Method for Multiple AGVs

Abstract

A traffic control system for global scheduling of multiple AGVs based on traffic flow, including a task allocation module, a global path planning module, and a traffic control module; among them, the task allocation module is used to allocate tasks and bind them to a certain AGV to execute this task; the global path planning module is used to use a multi-AGV collision-free path planning method based on path expected traffic flow for all AGVs with moving requirements to plan the global path for them to reach the target location; the traffic control module is used to monitor the running status of each AGV in real time, and according to the real-time traffic flow and path attributes of the remaining path, if it can continue to move forward, it will segmentally issue the remaining path sequence and the corresponding path attributes to each AGV for execution, if not, it will let the AGV wait for commands in place or re-plan the path starting from the current position. This system can avoid multi-AGV path conflicts or congestion and ensure the efficient operation of the overall system.

Description

Technical Field

[0001] The present disclosure belongs to the technical field of AGV intelligent control, and particularly relates to a multi-AGV global scheduling method based on traffic. Background Art

[0002] An automated guided vehicle (AGV) is a transportation device that can automatically complete material handling. It usually uses radio, cameras, lidar, or magnetic strips, magnetic nails, and two-dimensional codes marked on the ground for navigation. Compared with other logistics transportation devices, AGVs have the advantages of strong adaptability, high automation level, labor cost savings, and easy maintenance. With the gradual increase in labor costs and the increasingly variable production pattern, more and more enterprises adopt highly automated production systems, and AGVs are an important part of the automated production system. Therefore, the design and research of AGVs are of great significance for enterprises to improve production efficiency and reduce production costs.

[0003] Compared with a single AGV, a multi-AGV system has a stronger ability to perform handling operations. In the face of a complex and changeable factory environment, a multi-AGV system can quickly respond to changes in the external environment and complete tasks flexibly and efficiently. Therefore, the development of a multi-AGV system has become inevitable. Compared with a single AGV, a multi-AGV system needs to solve problems such as multi-robot cooperation, information interaction, and conflict resolution. Therefore, the design of a multi-AGV system is more complex. Currently, the design of a multi-AGV system with high groupability, self-organization, and self-adaptability has become a research hotspot, mainly including the following research directions: research on motion analysis and control algorithms, research on autonomous perception and networking algorithms, research on multi-robot positioning algorithms, research on multi-robot autonomous dynamic decision-making and path planning algorithms, research on multi-robot formation combination dissolution and bionic behavior simulation algorithms, etc.

[0004] The scheduling problem of multi-AGVs is also a key point of the multi-AGV system. The purpose of this problem is to plan appropriate global paths for multiple AGVs in the entire AGV system so that they can move from the starting point to the ending point without collisions. Since multiple AGVs share paths, during the walking process, in order to solve the resource competition of AGVs, multiple AGVs need to be controlled. Currently, there are two methods for scheduling and controlling AGVs: centralized control and distributed control. Centralized control means that a scheduling center uniformly schedules and plans all AGVs, while distributed control is that each AGV plans for itself separately. This application adopts the method of centralized control, which has the advantages of simple structure, convenient control, and higher scheduling efficiency.

[0005] In terms of centralized control, it can be mainly divided into two categories: the first category is to decompose the problem into single AGV global path planning, and then use specific traffic management rules or walking rules to avoid collisions between multiple AGVs when the AGV is walking; the second category will consider when and where the AGV will be, and directly plan a non-conflicting path with time constraints for each AGV, so that when the AGV walks strictly according to the path requirements, multiple AGVs will not collide. However, the above two methods have their own defects. For example, the first method cannot take the paths of other AGVs into account during planning, resulting in the planned path may cause congestion in a certain section of the road or deadlock between AGVs, and make it difficult for traffic management rules or walking rules to resolve conflicts between multiple AGVs; while the second method lacks flexibility and is only suitable for use in scenarios without random interference. In scenarios with more random interference, it may be difficult for the AGV to meet the time constraints of its path during walking. Summary of the invention

[0006] In order to solve the above technical problems, the present disclosure discloses a traffic-based multi-AGV global scheduling traffic control system, including a task allocation module, a global path planning module, and a traffic control module; wherein,

[0007] The task allocation module is used to allocate tasks and bind them to a certain AGV to perform the tasks;

[0008] The global path planning module is used to plan the global path to the target location for all AGVs with movement requirements using a multi-AGV collision-free path planning method based on the expected flow of the path;

[0009] The traffic control module is used to monitor the operating status of each AGV in real time, and according to the real-time traffic and path attributes of the remaining path, if it can continue to move forward, the remaining path sequence and corresponding path attributes are sent to each AGV in segments for execution. If not, the AGV is allowed to wait for commands on the spot or re-plan the path with the current position as the starting point.

[0010] Through the above technical solutions, the traffic-based multi-AGV global scheduling and traffic control system docks with the business logic, receives navigation tasks, selects a currently idle AGV, and assigns the tasks to this AGV. After the task assignment is completed, according to the expected traffic of each path, it is responsible for planning paths for each running AGV. Since the applicable scope of this AGV system includes wide paths that can accommodate multiple AGVs driving side by side, this solution provides a multi-AGV collision-free path planning method based on path expected traffic. On the topological map obtained by real environment modeling, based on the classic path planning algorithm A*, the A* algorithm, in the calculation process of the cost function, considers the attributes and expected traffic of the path, plans a suitable global path, and aims to minimize the AGV travel time, reducing the possibility of multi-AGV congestion. After obtaining the global path, it monitors the running status of each AGV in real time, and according to the real-time traffic and path attributes of the remaining path, if it can continue to move forward, it segments and issues the remaining path sequence and corresponding path attributes to each AGV for execution. If not, it makes the AGV wait in place or re-plan the path to avoid multi-AGV path conflicts or congestion and ensure the efficient operation of the overall system. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 is a flowchart of the traffic-based multi-AGV global scheduling solution provided in an embodiment of the present disclosure;

[0012] Figure 2 is a flowchart of the multi-AGV collision-free path planning method based on path expected traffic provided in an embodiment of the present disclosure;

[0013] Figure 3 is a flowchart of the released path selection method provided in an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0014] In order to enable those skilled in the art to understand the technical solutions disclosed in the present disclosure, the following will combine the embodiments and relevant appendices Figures 1 to 3 to describe the technical solutions of each embodiment. The described embodiments are part of the embodiments of the present disclosure, rather than all of the embodiments. The terms "first", "second", etc. used in the present disclosure are used to distinguish different objects, rather than to describe a specific order. In addition, "including" and "having" and any variations thereof are intended to cover and non-exclusively include. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally further include steps or units not listed, or may optionally further include other steps or units inherent to these processes, methods, systems, products, or devices.

[0015] References to "embodiments" in this specification mean that the particular features, structures, or characteristics described in connection with the embodiments can be included in at least one embodiment of the present disclosure. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will understand that the embodiments described herein can be combined with other embodiments.

[0016] See Figure 1 , in one embodiment, the present disclosure discloses a traffic-based multi-AGV global scheduling traffic control system, including a task assignment module, a global path planning module, and a traffic control module; wherein,

[0017] The task assignment module is used to assign tasks and bind them to a certain AGV to execute the tasks;

[0018] The global path planning module is used to use a multi-AGV collision-free path planning method based on path expected traffic for all AGVs with movement requirements to plan the global path for them to reach the target location;

[0019] The traffic control module is used to monitor the running status of each AGV in real time, and according to the real-time traffic and path attributes of the remaining path, if it can continue to move forward, segmentally issue the remaining path sequence and corresponding path attributes to each AGV for execution, if not, let the AGV wait for commands in place or re-plan the path starting from the current position.

[0020] For this embodiment, the main purpose of the system is to assign tasks and bind them to a certain AGV to execute the tasks. At the same time, it is also necessary to use a multi-AGV collision-free path planning method based on path expected traffic for all AGVs with movement requirements to plan the global path for them to reach the target location, and through the traffic control module, monitor the running status of each AGV in real time, and according to the real-time traffic and path attributes of the remaining path, select to segmentally issue the remaining path sequence and corresponding path attributes, or select to issue a wait-in-place command, or re-plan the path starting from the current position, to avoid multi-AGV path conflicts or congestion and ensure the efficient operation of the overall system.

[0021] As Figure 1 shown, the communication protocols for uploading and downloading by the system are specified. The communication protocol for the AGV to upload to the system is V2S_protocol, which mainly includes content such as online request, position initialization, status information, response instructions, etc.; the communication protocol for the system to download to the AGV is S2V_protocol, which mainly includes content such as initialization response feedback, online response feedback, path type, path number, task instructions, etc.

[0022] In another embodiment, in the task assignment module, it is necessary to continuously check whether there are new task requirements; when new task requirements are found, the new tasks need to be added to the unassigned task list; when the scenario information is updated each time, it is necessary to check whether there are idle AGVs and unassigned tasks currently; if not, continue to check whether there are new tasks; if so, assign the unassigned task with the shortest distance from the task location to the current position of the latest uploaded AGV to this AGV, and delete this assigned task from the unassigned task list.

[0023] For this embodiment, in the task assignment module, the module needs to continuously check whether new navigation task requirements are sent to the system. When new task requirements are found, the new tasks need to be added to the unassigned task list. When the scenario information is updated each time, it is necessary to check whether there are idle AGVs, that is, AGVs that have not been assigned tasks currently. If there are no idle AGVs, continue to check whether there are new tasks; if so, for the idle AGV that has uploaded the latest information to the scheduling module, select the task with the shortest distance from the task location to the current position of this AGV from the unassigned task list, assign this task to this AGV, and delete this assigned task from the unassigned task list.

[0024] In another embodiment, wherein, the multi-AGV collision-free path planning method based on path expected traffic is specifically as follows:

[0025] On the topological map obtained by real environment modeling, based on the classic path planning algorithm A* algorithm, considering the attributes of the path and the path expected traffic during the calculation of the cost function, plan a suitable global path, aiming to minimize the travel time of the AGV and reduce the possibility of multi-AGV congestion.

[0026] For this embodiment, in order to complete the global path planning for the AGV, a new multi-AGV collision-free path planning method based on path expected traffic is designed. This method is based on the classic path planning algorithm A* algorithm on the topological map obtained by real environment modeling, considering the attributes of the path and the expected traffic during the calculation of the cost function, planning a suitable global path, aiming to minimize the travel time of the AGV and reduce the possibility of multi-AGV congestion.

[0027] Different from the common A* algorithm, a new cost G value calculation method based on path expected traffic is designed. When expanding a node, if the current node is N and the node to be expanded is A, and the path from N to A is passable and not blocked, then the cost G A value of node A can be expressed as:

[0028]

[0029] Where GN is the cost G value of the current node N, d NA is the length of the path from node N to node A, r NA is the minimum radian value that needs to be deflected when moving from node N onto the path from node N to node A, e NA is the expected traffic of the path from node N to node A, m NA is the maximum traffic of the path from node N to node A, obtained from the path attributes. θ, β, and γ are constants used to adjust the preference for generating paths and can be set according to the actual situation. The cost G value of the initial node is 0.

[0030] The H of node A A value defines the straight-line distance to the target node, and the final F A value can be obtained from F A = G A + H A and calculated.

[0031] The specific process of the multi-AGV collision-free path planning method based on path expected traffic is as follows Figure 2 shown. First, the positions of the starting point and the ending point of the AGV, as well as the existing expected traffic of the current topological map, need to be given. Define the OPEN table and the CLOSE table, initialize them to be empty, and add the starting point to the OPEN table after calculating its F value. Use a loop to judge whether the OPEN table is empty each time. If it is empty, it means that the path solution fails; if it is not empty, select the node N with the smallest F value in the OPEN table at this time and put it into the CLOSE table. Judge whether node N is the ending point. If it is, it means that the solution is successful. Search backward for its previous nodes to generate the optimal path, update the expected traffic, and output the path; if it is not, obtain all adjacent node sets Adj of node N, take out each adjacent node in turn for expansion judgment until the set Adj is empty. At this time, enter the next loop to judge whether the OPEN table is empty.

[0032] When expanding the adjacent node A, first determine whether the AGV can travel from node N to node A at this time. If not, end the expansion judgment of this node; if so, calculate the G value from node N to node A according to the expected traffic. Check whether node A is in the OPEN list. If it is in the OPEN list, then determine whether the new G value is less than the old G value. If it is less, delete node A in the OPEN list, calculate the H value and F value of node A, and then add node A back to the OPEN list to complete the expansion judgment. If it is greater than or equal to the old G value, end the expansion judgment of this node; if node A is not in the OPEN list, it is necessary to determine whether it is in the CLOSE list. If it is in the CLOSE list, the same as before, determine whether the new G value is less than the old G value and perform the same subsequent operations; if not, calculate the H value and F value of node A and add it to the OPEN list to complete the expansion judgment.

[0033] In Figure 2 it should be noted that after taking out a node A from all adjacent node sets Adj of the current node N, it is necessary to determine whether this AGV can travel from node N to node A. Since there may be one-way paths in the real environment, while the paths in the topological map are all directionless, it is necessary to check whether the path from N to A can pass according to the real environment when expanding the node. In addition, it is also necessary to check whether the path from N to A is blocked. If the path planning fails, directly end the current scheduling and display a warning message.

[0034] In another embodiment, the path expected traffic will be updated in two cases: the first case is when the real-time scenario information is updated; the second case is after the optimal path is generated.

[0035] For this embodiment, the path expected traffic will be updated in two cases. The first case is when the real-time scenario information is updated. If the AGV has completed the previous path, subtract 1 from the expected traffic of the previous path, indicating that this AGV has passed this path. The second case is after the optimal path is generated, and it is necessary to add 1 to the expected traffic of each segment of the global path, indicating that this AGV will pass through these paths.

[0036] In another embodiment, in the traffic control module, first, the remaining path of each AGV is updated; then it is determined whether there is still a remaining path; if not, it is considered that this AGV has reached the task end point, the task is completed, and this AGV is set to the idle state, and this scheduling is ended; if there is, it is checked whether this AGV needs to be released; if not, this scheduling is ended and wait for the next scheduling; if so, it is checked whether the next section of the path in front of this AGV is blocked. If the path in front is blocked, it is necessary to use the end point of the path where the AGV is currently located as the starting point and the target point as the end point to eliminate the influence of its original remaining path on the expected traffic of the corresponding path, re-perform global path planning, and execute the subsequent process; if it is not blocked, a path that can be released is selected from the remaining paths; it is determined whether there is a path that can be released. If not, a waiting command is issued and this scheduling is ended; if there is, the real-time traffic of the path that can be released is updated, and at the same time, the released path sequence and the corresponding path attributes are issued, and this scheduling is ended.

[0037] For this embodiment, since each path has its maximum traffic limit, in order to avoid the traffic of AGVs on a certain section of the path exceeding its maximum traffic limit, causing danger or congestion, the system uses the traffic control module to implement the traffic management of multiple AGVs, that is, the segmented distribution and release of the remaining global paths of multiple AGVs. The specific process is as Figure 1 shown.

[0038] After each real-time scene information update, the traffic control module needs to update the real-time traffic of each section of the path in the topological map. When an AGV first reaches or leaves the end point of a certain section of the path, the real-time traffic of this path is reduced by 1. After the system determines the release path for the AGV, the real-time traffic of all the paths that can be released is increased by 1.

[0039] In addition, the system also needs to perform a scheduling for each AGV. First, the remaining path of each AGV is updated. After the remaining path is updated, it is determined whether there is still a remaining path: if not, it is considered that this AGV has reached the task end point, the task is completed, and this AGV is set to the idle state, and this scheduling is ended; if there is, it is first checked whether this AGV needs to be released. When it is determined that release is required, it is first checked whether the next section of the path in front of this AGV is blocked. If the path in front is blocked, it is necessary to use the end point of the path where the AGV is currently located as the starting point and the target point as the end point to eliminate the influence of its original remaining path on the expected traffic of the corresponding path, re-perform global path planning, and execute the subsequent process; if it is not blocked, a path that can be released is selected from the remaining paths. If there is no path that can be released, an in-place waiting command is issued to this AGV; if there is a path that can be released, the real-time traffic of the path that can be released is updated, and at the same time, the corresponding path sequence and attributes are issued, and this scheduling is ended.

[0040] In another embodiment, the specific method for selecting a passable path from the remaining paths is as follows: Initialize the passable list; Take the first path of the remaining paths as the candidate path; Determine whether the real-time traffic of the candidate path is less than the maximum traffic. If so, add the candidate path to the passable path list and delete the candidate path from the remaining paths; If not, output the passable path list and end the selection; Determine whether the total length of the paths in the passable path list is greater than the threshold. If so, output the passable path list and end the selection; If not, return to taking the first path of the remaining paths as the candidate path and continue to judge until the selection ends.

[0041] For this embodiment, the specific process method is as Figure 3 shown, and the size of the threshold can be set according to the actual situation.

[0042] In another embodiment, the maximum traffic is the number of AGVs that can pass side by side in the narrowest part of the entire path in the same driving direction; The real-time traffic means, for all passable paths, how many AGVs are currently passing or about to pass through this path.

[0043] For this embodiment, in this method, in order to meet the requirement that multiple AGVs can still theoretically pass through a path safely in the worst case, the maximum traffic of a path is defined as the maximum number of AGVs that can pass through the path side by side safely throughout the journey, that is, at the narrowest part of the path, the number of AGVs that can be accommodated side by side is the maximum traffic. Therefore, the expected traffic means, for the global path, how many AGVs are currently passing or about to pass through this path; The real-time traffic means, for all passable paths, how many AGVs are currently passing or about to pass through this path.

[0044] In another embodiment, the specific steps for updating the remaining path of each AGV are as follows: For each AGV, by locating its current position, that is, it can be obtained which path segment it is currently on. In its remaining global path, delete all paths before this path segment. If the current AGV is located at the end of this path segment, this path segment needs to be deleted together.

[0045] For this embodiment, for each AGV, by locating its current position, that is, it can be obtained which path segment it is currently on. In its remaining global path, delete all paths before this path segment. If the current AGV is located at the end of this path segment, this path segment needs to be deleted together. After the global path planning, the remaining path also needs to be updated. Specifically, it is to delete all existing remaining paths and set the new global path as the remaining path.

[0046] In another embodiment, the specific criterion for checking whether this AGV needs to be released is as follows: If no path has been issued to the AGV yet, or the distance between the AGV and the end point of the last segment of the previously issued path is less than the set threshold, it is determined that release is required; otherwise, release is not allowed. Here, the set distance threshold is the path attribute of each path segment.

[0047] For this embodiment, if no path has been issued to the AGV yet, or the distance between the AGV and the end point of the last segment of the previously issued path is less than the set threshold, it is determined that release is required, and the subsequent process is carried out; otherwise, the remaining paths do not need to be issued later, that is, release is not allowed, and this scheduling is ended, waiting for the next scheduling of this AGV. The set distance threshold is the path attribute of each path segment and needs to be given during the topological map modeling.

[0048] In another embodiment, the path being blocked specifically means that when the path is completely impassable due to some obstacles or AGVs and cannot be restored to pass through by the scheduling of the scheduling system itself, these impassable paths are the blocked paths.

[0049] Although the embodiments of the present invention have been described above in conjunction with the accompanying drawings, the present invention is not limited to the above specific embodiments and application fields. The above specific embodiments are merely illustrative and guiding, rather than restrictive. Those of ordinary skill in the art can also make many forms under the inspiration of this specification and without departing from the scope protected by the claims of the present invention, and all of these fall within the scope of protection of the present invention.

Claims

1. A multi-AGV global scheduling and traffic control system based on traffic flow, including a task allocation module, a global path planning module, and a traffic control module; wherein, The task allocation module is used to allocate tasks and bind them to a certain AGV to execute this task; The global path planning module is used to use the multi-AGV collision-free path planning method based on path expected traffic flow for all AGVs with moving requirements to plan the global path for them to reach the target location; The traffic control module is used to monitor the running status of each AGV in real time, and according to the real-time traffic flow and path attributes of the remaining path, if it can continue to move forward, it will segmentally issue the remaining path sequence and corresponding path attributes to each AGV for execution. If not, it will let the AGV wait for commands in place or re-plan the path starting from the current position; Among them, in the traffic control module, first, the remaining path of each AGV is updated; then it is judged whether there is still a remaining path; if not, it is considered that this AGV has reached the task end point, the task is completed, and this AGV is set to the idle state, and this scheduling ends; if there is, it is checked whether this AGV needs to be released; if not, this scheduling ends and waits for the next scheduling; if so, it is checked whether the next path in front of this AGV is blocked. If the path in front is blocked, it is necessary to take the end point of the path where the AGV is currently located as the starting point and the target point as the end point to eliminate the influence of its original remaining path on the expected traffic flow of the corresponding path, re-plan the global path, and execute the subsequent process; if it is not blocked, select the path that can be released from the remaining paths; judge whether there is a path that can be released. If not, issue a waiting command and this scheduling ends; if there is, update the real-time traffic flow of the path that can be released, and at the same time issue the released path sequence and corresponding path attributes to end this scheduling; The specific method for selecting the path that can be released from the remaining paths is: initialize the list of paths that can be released; take the first path of the remaining path as the candidate path; judge whether the real-time traffic flow of the candidate path is less than the maximum traffic flow. If so, add the candidate path to the list of paths that can be released and delete this candidate path from the remaining paths; if not, output the list of paths that can be released and end the selection; judge whether the total length of the paths in the list of paths that can be released is greater than the threshold. If so, output the list of paths that can be released and end the selection; if not, return to take the first path of the remaining path as the candidate path and continue to judge until the selection ends.

2. The system according to claim 1, in the task allocation module, it is necessary to continuously check whether there are new task requirements; when new task requirements are found, the new tasks need to be added to the list of unassigned tasks; when the scene information is updated each time, it is necessary to check whether there are idle AGVs and unassigned tasks currently; if not, continue to check whether there are new tasks; if there are, assign the unassigned task with the shortest distance from the task location assigned to the latest uploaded AGV to the current position of this AGV, and delete this assigned task from the list of unassigned tasks.

3. The system according to claim 1, wherein The multi-AGV collision-free path planning method based on path expected traffic flow is specifically as follows: Based on the A* algorithm, a classic path planning algorithm, on the topological map modeled from the real environment, the attributes of the path and the expected path flow are considered in the calculation of the cost function to plan a suitable global path, aiming to minimize the travel time of the AGV and reduce the possibility of multi-AGV congestion.

4. For the system according to claim 3, the expected path flow is updated in two cases: the first case is when the real-time scenario information is updated; the second case is after the optimal path is generated.

5. For the system according to claim 1, the maximum flow is the number of AGVs that can pass side by side in the narrowest part of the entire path when traveling in the same direction; the real-time flow is the number of AGVs that are currently passing or about to pass through this path for all released paths.

6. For the system according to claim 1, the specific steps for updating the remaining path of each AGV are as follows: for each AGV, by locating its current position, it can be obtained which path segment it is currently on. In its remaining global path, all paths before this path segment are deleted. If the current AGV is at the end of this path segment, this path segment needs to be deleted together.

7. For the system according to claim 1, the specific judgment criterion for checking whether this AGV needs to be released is as follows: If no path has been sent to the AGV yet, or the distance between the AGV and the end point of the last segment of the previously sent path is less than the set threshold, it is determined that release is required; otherwise, release is not allowed. Among them, The set distance threshold is the path attribute of each path segment.

8. For the system according to claim 1, a path is blocked specifically when the path is completely impassable due to some obstacles or AGVs and cannot be restored to pass through by the scheduling of the scheduling system itself. These impassable paths are the blocked paths.

Images