CN-121977570-A - Multi-AGV path planning method based on regional impedance sensing and space-time dynamic cooperation

CN121977570ACN 121977570 ACN121977570 ACN 121977570ACN-121977570-A

Abstract

The invention relates to the technical field of intelligent warehouse logistics and industrial automation, in particular to a multi-AGV path planning method based on regional impedance sensing and space-time dynamic cooperation, which comprises the steps of acquiring grid map information and task information of each AGV based on a warehouse environment; the method comprises the steps of constructing a local observation window taking an AGV as a central node, analyzing nodes covered by corresponding grid map information, determining regional passage impedance indexes, constructing barrier distance penalty items, dynamically adjusting heuristic function weights in an A-algorithm by adopting the regional passage impedance indexes, building an evaluation function in combination with the barrier distance penalty items, generating a single-machine global path, generating a time window for each node in the global path based on task information of the AGV, judging path conflict of the current warehouse environment through the time window, presetting a task priority strategy, resolving the path conflict, and generating a cooperative scheduling instruction to regulate the running path of the AGV. The search efficiency is improved by sensing the degree of congestion of the environment to perform self-adaptive adjustment.

Inventors

WANG HUQING
CHEN LIUKUN
SUN ZHIXIN
XU YUHUA

Assignees

南京邮电大学

Dates

Publication Date: 20260505
Application Date: 20260209

Claims (9)

1. The multi-AGV path planning method based on the cooperation of regional impedance sensing and space-time dynamics is characterized by comprising the following steps: Acquiring grid map information and task information of each AGV based on a storage environment; Constructing a local observation window taking an AGV as a central node, analyzing nodes covered by corresponding grid map information, determining regional passing impedance indexes, and constructing barrier distance penalty items; Dynamically adjusting heuristic function weights in an A-algorithm by adopting regional pass impedance indexes, and establishing an evaluation function by combining barrier distance penalty items to generate a single-machine global path; Generating a time window for each node in the global path based on the task information of the AGV, and judging path conflict of the current storage environment through the time window; and presetting a task priority strategy, resolving the path conflict, and generating a cooperative scheduling instruction to regulate and control the running path of the AGV.
2. The multi-AGV path planning method according to claim 1 based on cooperation of regional impedance sensing and spatiotemporal dynamics, wherein acquiring the grid map information and the task information of each AGV based on the warehouse environment comprises: Establishing a two-dimensional map model for the storage environment by adopting a grid method, dividing the storage environment into a plurality of grid units, and correspondingly marking no obstacle and obstacle according to the storage environment; And receiving task information of each AGV based on the storage environment, wherein the task information comprises a task starting point, a task ending point and a task priority of each AGV.
3. The multi-AGV path planning method according to claim 2, wherein constructing a local observation window with an AGV as a central node, analyzing nodes covered by corresponding grid map information, determining a regional pass impedance index, and constructing an obstacle distance penalty term comprises: Counting the number of the obstacles in the local observation window and the distance from the central node to each obstacle, and determining the regional passing impedance index; and screening to obtain the distance from the center node to the nearest obstacle, introducing a dynamic risk gain coefficient based on exponential growth, and constructing an obstacle distance penalty term.
4. The multi-AGV path planning method based on the cooperation of regional impedance sensing and space-time dynamics according to claim 2, wherein the method for dynamically adjusting heuristic function weights in the A-algorithm by adopting regional pass impedance indexes and establishing an evaluation function by combining barrier distance penalty items to generate a single-machine global path comprises the following steps: establishing a mapping relation between the regional transit impedance indexes and weights of heuristic functions in an A-algorithm, presetting an impedance threshold value, and dynamically adjusting the weights; the task end point corresponding to the task information of the AGV is a target node, a target guide item is determined by integrating the target node and a central node, a heuristic function is corrected by combining the obstacle distance penalty item, and an evaluation function is built by combining the adjusted weight; and carrying out global path search based on the evaluation function to generate a single global path of the AGV.
5. The multi-AGV path planning method based on the cooperation of regional impedance sensing and space-time dynamics according to claim 4, wherein the preset impedance threshold value is used for dynamically adjusting the weight, specifically: Setting a region dividing mechanism, comparing an impedance threshold with a region passing impedance index, judging a sparse region, a general region or a dense region of the AGV, and correspondingly setting weights.
6. The multi-AGV path planning method according to claim 2 based on cooperation of regional impedance sensing and spatio-temporal dynamics, wherein generating a time window for each node in the global path based on the task information of the AGV, determining a path collision of the current warehouse environment through the time window, comprises: collecting the running speed and the path length of the AGV in the task information executing process, and generating a time window for each node in the global path; and traversing time windows of all AGVs, judging whether time windows overlap among the AGVs, and if so, indicating that path conflict exists in the current storage environment.
7. The multi-AGV path planning method according to claim 6 based on cooperation of regional impedance sensing and spatio-temporal dynamics, wherein the step of collecting the running speed and path length of the AGV in the task information process and generating a time window for each node in the global path comprises: determining an average cruising speed according to the running speed of the AGV in the task information executing process, obtaining the distance between adjacent nodes according to the path length, and determining physical time consumption; analyzing the steering condition of the AGV when executing the task information process, and determining the dynamic buffering time; And acquiring the entry time of each node, obtaining the exit time by combining the physical time consumption and the dynamic buffer time, and integrating the entry time and the exit time to generate a time window of the corresponding node.
8. The multi-AGV path planning method based on cooperation of regional impedance sensing and space-time dynamics according to claim 2, wherein the method for resolving path conflicts by presetting task priority strategies and generating a cooperative scheduling instruction to regulate and control an operation path of an AGV comprises the following steps: Comparing task priorities of AGVs in nodes with path conflicts, and respectively determining a high-priority AGV and a low-priority AGV; the high-priority AGV keeps an original running path and running speed, and compares a time window of the high-priority AGV with a time window of the low-priority AGV to determine waiting time of the low-priority AGV; and when the waiting time is greater than or equal to the time threshold, marking the conflict node as an obstacle, and regenerating a single global path.
9. The electronic equipment is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus, and the processor calls logic instructions in the memory to execute the multi-AGV path planning method based on the cooperation of regional impedance sensing and space-time dynamics according to any one of claims 1-8.

Description

Multi-AGV path planning method based on regional impedance sensing and space-time dynamic cooperation Technical Field The invention relates to the technical field of intelligent warehouse logistics and industrial automation, in particular to a multi-AGV path planning method based on regional impedance sensing and space-time dynamic cooperation. Background With the advancement of intelligent manufacturing strategies, intelligent logistics systems have become a central component of modern factories and warehouse centers. AGVs (Automated Guided Vehicle, automatic guided vehicles) are widely used in materials handling, electronic commerce sorting and flexible production lines as key equipment for achieving logistics automation. In practical applications, the AGV needs to operate efficiently in a complex environment full of shelves, columns and moving personnel, and therefore, the performance of the path planning algorithm directly determines the operating efficiency of the whole warehouse system. Currently, global path planning of an AGV mainly depends on a traditional a-algorithm or Dijkstra (Dijkstra) algorithm, which is used for a simple environment to perform well, but in a high-density and high-dynamic complex warehouse scene, a part of significant technical bottlenecks still exist: First, environmental suitability contradicts search efficiency. Conventional a-x algorithms typically employ fixed heuristic weights. In the sparse area of the obstacle, the searching speed is still high, but in the dense narrow channel or dead-beard area of the obstacle, the algorithm with fixed weight is easy to sink into the local optimum, so that the searching nodes are increased rapidly, a large number of redundant broken lines exist in the planned path, and the running efficiency is seriously influenced. Secondly, path planning lacks physical security constraints. Existing raster path planning generally pursues "shortest theoretical path", resulting in planned trajectories that tend to cling to obstacle edges. In the physical world, the control tracking error and the width of the vehicle body of the AGV are limited, the welted running is extremely easy to cause scraping accidents, and the scraping accidents can be relieved by increasing the expansion radius of the obstacle, but the traffic capacity of a narrow passage is sacrificed. Thirdly, the deadlock problem during multi-machine collaborative operation. Under the coexistence of multiple AGVs, dynamic conflict cannot be solved by simple static path planning, and the existing scheduling system mostly adopts simple traffic control such as mutual exclusion lock and the like, and lacks accurate pre-judgment based on time dimension. Disclosure of Invention In order to solve the technical problems of poor environmental adaptability, insufficient safety and high collision rate of a multi-AGV system of the traditional A-algorithm, the invention aims to provide a multi-AGV path planning method based on the cooperation of regional impedance sensing and space-time dynamics, and the adopted technical scheme is as follows: Acquiring grid map information and task information of each AGV based on a storage environment; Constructing a local observation window taking an AGV as a central node, analyzing nodes covered by corresponding grid map information, determining regional passing impedance indexes, and constructing barrier distance penalty items; Dynamically adjusting heuristic function weights in an A-algorithm by adopting regional pass impedance indexes, and establishing an evaluation function by combining barrier distance penalty items to generate a single-machine global path; Generating a time window for each node in the global path based on the task information of the AGV, and judging path conflict of the current storage environment through the time window; and presetting a task priority strategy, resolving the path conflict, and generating a cooperative scheduling instruction to regulate and control the running path of the AGV. Preferably, acquiring grid map information and task information of each AGV based on the warehouse environment includes: Establishing a two-dimensional map model for the storage environment by adopting a grid method, dividing the storage environment into a plurality of grid units, and correspondingly marking no obstacle and obstacle according to the storage environment; And receiving task information of each AGV based on the storage environment, wherein the task information comprises a task starting point, a task ending point and a task priority of each AGV. Preferably, constructing a local observation window with an AGV as a central node, analyzing nodes covered by corresponding grid map information, determining a regional passage impedance index, and constructing an obstacle distance penalty term, wherein the method comprises the following steps: Counting the number of the obstacles in the local observation window and the distance from the central nod