AGV path planning algorithm and conflict-free strategy based on Dlite algorithm

Abstract

The invention discloses an AGV path planning algorithm and a conflict-free strategy based on a Dlite algorithm. The algorithm and the strategy are suitable for multi-AGV collaborative operation in dynamic complex environments such as track equipment assembly workshops. The method comprises the following steps: (1) adopting an improved Dlite algorithm to realize dynamic path planning, and shortening the time consumed for updating a single AGV path through an incremental re-planning mechanism to adapt to the real-time change of dynamic obstacles such as personnel flow and equipment displacement; (2) constructing a space-time joint constraint model, dividing a workshop channel into limited grid units, and accurately marking the space-time occupancy state of the AGV through a time window; and (3) designing a hierarchical conflict resolution strategy, and when path crossing is detected, dynamically allocating priorities based on the task emergency degree and the AGV cargo carrying value, and realizing conflict avoidance through a speed adjustment window and path fine adjustment.

Description

Technical Field

[0001] This invention relates to the field of AGV scheduling technology, and specifically discloses an AGV path planning algorithm and a conflict-free strategy based on the Dlite algorithm. Background Technology

[0002] The bidirectional, single-lane design of AGVs means that when multiple AGVs operate simultaneously within a single area, planned paths may intersect or overlap. Factors such as changes in cargo loading / unloading cycles, personnel operation delays, or overall operational efficiency can also cause deviations in the AGV's arrival time at the task point, forcing the AGV to stop midway and obstructing the normal passage of other AGVs. One-time path planning during the task allocation phase is insufficient to handle the various uncertainties in actual operation, easily leading to deadlocks or conflict risks. In the AGV's working environment, task points and obstacles such as machines and shelves are usually fixed and can be considered static elements; while personnel flow, cargo stacking, and other factors that may affect the AGV's route exhibit dynamic characteristics and are classified as dynamic elements. Given the complexity and uncertainty of the workshop environment, simple static global path planning is insufficient to adapt to these constantly changing working conditions. Therefore, a dynamic path planning algorithm is urgently needed for path planning of a single AGV. Multi-AGV path planning requires that only one AGV can exist at the same location at any given time. Therefore, a scheduling system is needed to perform temporal and spatial scheduling to ensure that AGVs replan their paths in specific ways within a specific working scenario. This paper adopts a conflict-free strategy planning method based on time windows, introducing a time window mechanism to precisely control the passage permissions of each AGV at key nodes and road segments. This strategy, combined with a priority method, effectively detects and resolves potential conflicts in AGV paths by rationally allocating time and space resources, thus achieving the safe and efficient operation of the multi-AGV system. Summary of the Invention

[0003] (I) Purpose of the Invention

[0004] The purpose of this invention is to provide an AGV path planning algorithm and a conflict-free strategy based on the Dlite algorithm, solving the following technical problems: traditional AGV path planning is inefficient, and static path planning is costly.

[0005] There are too many resources, and the system's ability to dynamically adjust is insufficient.

[0006] (2) Technical solution

[0007] An AGV path planning algorithm and conflict-free strategy based on the Dlite algorithm are disclosed, including a path planning module, a conflict detection module, and a conflict resolution module. The detailed steps of this invention are as follows:

[0008] Step 1: Algorithm Execution Flow Planning

[0009] (1) Path planning module:

[0010] This function is responsible for calculating the optimal path from the starting point to the destination for the AGV. The optimal path typically considers multiple factors, such as path length, travel time, and energy consumption. The Dlite algorithm is used to efficiently replan the path when the environment changes, adapting to dynamic environments.

[0011] (2) Conflict detection module:

[0012] A time window algorithm is used to identify potential conflicts or collisions between AGVs.

[0013] (3) Conflict resolution module:

[0014] The conflict detection module identifies potential conflicts, and the conflict resolution module intervenes to ensure that each AGV can reach its destination safely and efficiently, while minimizing the impact on other AGVs.

[0015] Step 2: Conflict-free path planning strategy planning

[0016] (1) Detect path changes

[0017] After obtaining the task scheduling results, if the path planning module detects dynamic changes in the environment, it generates a new path based on the improved Dlite algorithm. A turning cost factor is introduced into the heuristic function calculation. When the turning cost exceeds the threshold, path replanning is triggered. When multiple AGVs are running in coordination, the time window detection module obtains the current number of path conflicts. If a path conflict occurs, the priority adjustment strategy is activated to eliminate the conflict by stopping and waiting or path reallocation.

[0018] (2) Dynamic path update

[0019] The dynamic path planning layer updates the status based on the real-time map. If the obstacle change rate Δ > 5%, it automatically calls the improved Dlite algorithm, combining turning costs to optimize and generate a smooth path. It constructs a grid map based on the start and end points and initializes the priority queue (Open). The system first creates a list of nodes and adds the destination node to a queue, calculating its heuristic cost (h value). It then extracts the node with the lowest cost from the priority queue, traverses its neighboring nodes, and updates its cost (g value) if a neighboring node is unvisited or a better path exists, reinserting it into the queue. When environmental changes are detected (e.g., new obstacles), the obstacle status of affected nodes is updated, and the relevant nodes are marked as "needs re-evaluation." The heuristic function is recalculated for the marked nodes, introducing a turning cost factor to optimize path smoothness. The system backtracks from the current AGV position to the destination, generating a temporary path. If there are conflict risk nodes in the path, a time window model is triggered to assess the conflict probability. The time window algorithm predicts the time it will take for the AGV to reach the conflict node. If the time window overlap probability exceeds a threshold, the path or speed of low-priority AGVs is adjusted according to a priority strategy. If the adjustment fails or environmental changes are frequent, the priority queue is cleared and path searching is re-executed to generate an obstacle avoidance path, ensuring path feasibility in dynamic environments. The final path is then sent to the AGV execution layer, and the deviation between the AGV position and the path is monitored in real time. If the deviation exceeds the tolerance, local replanning is triggered.

[0020] (3) Conflict avoidance

[0021] When the conflict resolution module detects that the probability of conflict at the AGV intersection point Pc is greater than 80%, it prioritizes adjusting the path of the low-priority AGV. If the adjustment fails, the emergency stop protocol is activated. After receiving the planning instruction, if the path execution layer detects that the actual number of turns exceeds the planned value by more than 2, it will automatically feed back to the scheduling system to trigger path re-optimization.

[0022] Generate a spatiotemporal trajectory for the planned path of each AGV, record the start and end times of each node it passes through, and establish a four-tuple time window data with nodes as the spatial dimension and timestamps as the time dimension; traverse the time window set of all AGVs, and filter out AGV combinations that pass through the same nodes by matching the spatial dimensions. If the time windows of multiple AGVs overlap at the same node, they are marked as potential conflict nodes. Topology analysis is performed on overlapping nodes: when the current node and the next node of two AGVs form an intersecting path, it is a node conflict; if they share the start and end points of the same one-way channel, it is an overtaking conflict; when two AGVs move towards each other in the channel, it triggers a head-on conflict. A rolling time window mechanism is adopted. After each node's conflict detection is completed, the time window parameters of subsequent nodes are dynamically adjusted according to the actual running speed of the AGVs to achieve forward compensation for conflict prediction. Priority arbitration is initiated for AGV groups that have confirmed conflicts: priority weights are calculated based on parameters such as task urgency, AGV loading status, and path completion progress. Low-priority AGVs are inserted into a waiting time window before the conflict node, while high-priority AGVs gain priority and occupy the conflict node resources through a locking mechanism.

[0023] The beneficial effects of this invention are as follows: Compared with the prior art, this invention, by employing an improved Dlite algorithm, can efficiently recalculate the optimal path when the environment changes dynamically, effectively improving path planning efficiency; it introduces a turning cost factor to optimize path smoothness, enabling AGVs to reduce unnecessary turning times during task execution, thereby improving overall operating efficiency; it utilizes a time window algorithm for potential conflict detection and adopts priority adjustment strategies or emergency stop protocols to resolve conflicts, ensuring the safe operation of each AGV while minimizing the impact on other AGVs, improving the adaptability and flexibility of path planning, and ensuring the operating efficiency of AGVs. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying 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.

[0025] Figure 1 This is a flowchart illustrating the operation of the present invention.

[0026] Figure 2 This is the flowchart of the Dlite algorithm; Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, 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.

[0028] The conflict-free strategy described in this invention is as follows:

[0029] Step 1: After the path planning module obtains the task scheduling results, if it detects dynamic changes in the environment, it generates a new path based on the improved Dlite algorithm. A turning cost factor is introduced in the heuristic function calculation. When the turning cost exceeds the threshold, path replanning is triggered.

[0030] Step 2: When multiple AGVs are running in coordination, the time window detection module obtains the current number of path conflicts. If a path conflict occurs, the priority adjustment strategy is activated to eliminate the conflict by stopping and waiting or reassigning the path.

[0031] Step 3: The dynamic path planning layer updates the status based on the real-time map. If the obstacle change rate Δ > 5%, the improved Dlite algorithm is automatically invoked, and a smooth path is generated by combining the turning cost optimization.

[0032] Step 4: When the conflict resolution module detects that the probability of conflict at the AGV intersection point Pc is greater than 80%, it prioritizes adjusting the path of the low-priority AGV. If the adjustment fails, the emergency stop protocol is activated.

[0033] Step 5: After receiving the planning instructions, if the path execution layer detects that the actual number of turns exceeds the planned value by more than 2, it will automatically feed back to the scheduling system to trigger path re-optimization.

[0034] like Figure 2 The flowchart of the Dlite algorithm is shown below. The algorithm constructs a grid map based on the start and end points, initializes a priority queue (Open List), adds the end point node to the queue, and calculates its heuristic cost (h value). The node with the minimum cost is extracted from the priority queue, and its neighboring nodes are traversed. If a neighboring node has not been visited or a better path exists, its cost value (g value) is updated and it is re-inserted into the queue.

[0035] When an environmental change is detected (such as the addition of a new obstacle), the obstacle status of the affected nodes is updated, and the relevant nodes are marked as "needs to be re-evaluated". The heuristic function is recalculated for the marked nodes, a turning cost factor is introduced, and the path smoothness is optimized.

[0036] The system traces back from the current AGV position to the destination, generating a temporary path. If there are conflict risk nodes in the path, a time window model is triggered to assess the probability of conflict. The time window algorithm is used to predict the time period for the AGV to reach the conflict node. If the probability of time window overlap exceeds the threshold, the path or speed of the low-priority AGV is adjusted according to the priority strategy.

[0037] If the adjustment fails or the environment changes frequently, clear the priority queue and re-execute the path search to generate an obstacle avoidance path to ensure the feasibility of the path in a dynamic environment; send the final path to the AGV execution layer to monitor the deviation between the AGV position and the path in real time. If the deviation exceeds the tolerance, trigger local replanning.

[0038] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. An AGV path planning algorithm and conflict-free strategy based on the Dlite algorithm, characterized in that, Includes the following modules: The path planning module is responsible for calculating the optimal path from the starting point to the destination for the AGV. This optimal path typically considers multiple factors, such as path length, travel time, and energy consumption. Employing the Dlite algorithm, it can efficiently replan paths when the environment changes, adapting to dynamic environments. Conflict detection module: Employs a time window algorithm to identify potential conflicts or collisions between AGVs; Conflict resolution module: The conflict detection module identifies potential conflicts, and the conflict resolution module intervenes to ensure that each AGV can reach its destination safely and efficiently, while minimizing the impact on other AGVs.

2. The path planning module as described in claim 1, characterized in that: (1) After the path planning module obtains the task scheduling result, if it detects dynamic changes in the environment, it generates a new path based on the improved Dlite algorithm, introduces a turning cost factor in the heuristic function calculation, and triggers path replanning when the turning cost exceeds the threshold. (2) When multiple AGVs are running in coordination, the time window detection module obtains the current number of path conflicts. If a path conflict occurs, the priority adjustment strategy is started to eliminate the conflict by stopping and waiting or reassigning the path.

3. The path planning method as described in claim 2, characterized in that, In step (1), a grid map is constructed based on the starting point and the ending point, a priority queue (Open List) is initialized, and the ending point node is added to the queue. Its heuristic cost (h value) is calculated. The node with the lowest cost is extracted from the priority queue, and its adjacent nodes are traversed. If the adjacent node has not been visited or there is a better path, its cost value (g value) is updated and it is re-inserted into the queue.

4. The path planning method as described in claim 2, characterized in that, In step (2), when an environmental change (such as the addition of a new obstacle) is detected, the obstacle status of the affected node is updated and the relevant node is marked as "needs to be re-evaluated"; the heuristic function is recalculated for the marked node, a turning cost factor is introduced, and the path smoothness is optimized.

5. The collision detection module according to claim 1, characterized in that: (1) Generate a spatiotemporal trajectory for the planned path of each AGV, record the start and end times of each node it passes through, and establish a quadruple time window data with nodes as the spatial dimension and timestamps as the time dimension. (2) Traverse the time window set of all AGVs and filter out AGV combinations that pass through the same node by matching spatial dimensions. If the time windows of multiple AGVs have an intersection in the time interval of the same node, they are marked as potential conflict nodes; (3) Perform topology analysis on overlapping nodes: When the current node and the next node of two AGVs form an intersecting path, it is a node conflict; if they share the start and end points of the same one-way channel, it is an overtaking conflict; when two AGVs move towards each other in the channel, it triggers a head-on conflict. (4) A conflict detection module according to claim 1, characterized in that, in step (1), the dynamic path planning layer updates the status according to the real-time map, and if the obstacle change rate Δ>5%, it automatically calls the improved Dlite algorithm and combines the turning cost to optimize and generate a smooth path.

6. The conflict-free detection module according to claim 5, characterized in that, In step (1), when the conflict resolution module detects that the probability of conflict at the AGV intersection point Pc is greater than 80%, it prioritizes adjusting the path of the low-priority AGV. If the adjustment fails, the emergency stop protocol is activated. After receiving the planning instruction, if the path execution layer detects that the actual number of turns exceeds the planned value by more than 2, it automatically feeds back to the scheduling system to trigger path re-optimization.

7. The conflict-free detection module according to claim 5, characterized in that, In step (2), a rolling time window mechanism is adopted. After each node's conflict detection is completed, the time window parameters of subsequent nodes are dynamically adjusted according to the actual running speed of the AGV to achieve forward compensation for conflict prediction.

8. The conflict-free detection module according to claim 5, characterized in that, In step (3), priority arbitration is initiated for the AGV group that has been confirmed to be in conflict: priority weights are calculated based on parameters such as task urgency, AGV loading status, and path completion progress. Low-priority AGVs are inserted into a waiting time window before the conflict node, while high-priority AGVs obtain priority and occupy the conflict node resources through a locking mechanism.

9. A conflict resolution module according to claim 1, characterized in that: (1) Perform topology analysis on overlapping nodes: When the current node and the next node of two AGVs form an intersecting path, it is a node conflict; if they share the start and end points of the same one-way channel, it is an overtaking conflict; when two AGVs move towards each other in the channel, it triggers a head-on conflict. (2) A rolling time window mechanism is adopted. After the collision detection of each node is completed, the time window parameters of subsequent nodes are dynamically adjusted according to the actual running speed of the AGV to realize the forward compensation of collision prediction. (3) Initiate priority arbitration for AGV groups that have confirmed conflicts: calculate priority weights based on parameters such as task urgency, AGV loading status, and path completion progress. Low-priority AGVs are inserted into a waiting time window before the conflict node, while high-priority AGVs obtain priority and occupy conflict node resources through a locking mechanism.

Images