An AGV obstacle avoidance path planning method and system for intelligent warehousing

Abstract

This invention provides an AGV obstacle avoidance path planning method and system for intelligent warehousing, belonging to the field of path planning technology. It includes: collecting warehousing data; constructing a map reflecting the topological structure of the warehousing space using a topology map; obtaining an initial path using a pruned and optimized Dijkstra's algorithm based on a set start and end point; optimizing the initial path by setting repulsion factors and safety buffers to obtain a secondary optimized path; determining whether the environment has changed, and if so, recalculating a new obstacle avoidance path based on the secondary optimized path. This invention improves the overall efficiency and safety of warehousing cargo handling by constructing the initial path using a pruned and optimized Dijkstra's algorithm and optimizing the initial path by setting repulsion factors and safety buffers, and enhances the adaptability of the warehousing system to complex environmental changes.

Description

Technical Field

[0001] This invention relates to the field of path planning technology, and in particular to an AGV obstacle avoidance path planning method and an AGV obstacle avoidance path planning system for intelligent warehousing. Background Technology

[0002] With the booming development of e-commerce, manufacturing, and other industries, the volume of warehousing and logistics has increased dramatically, placing higher demands on the efficiency, accuracy, and intelligence of warehousing operations. Traditional warehousing models relying on manual handling and simple mechanical assistance are no longer sufficient to meet the needs of large-scale, high-frequency cargo processing. Intelligent warehousing has emerged to address this need, aiming to achieve efficient and precise control of warehousing processes through automated equipment and information technology. AGVs (Automated Guided Vehicles), as key automated handling equipment in intelligent warehousing, need to operate stably in complex and dynamically changing warehousing environments. Effectively avoiding obstacles and rationally planning paths have become core technical points for ensuring their efficient operation, driving the continuous development of AGV obstacle avoidance path planning technology.

[0003] Warehouses contain different types of rack layouts, such as traditional row racks and high-rise racks in automated storage and retrieval systems. The racks form aisles of varying widths, and there are often temporarily stacked goods and forklifts used for replenishment moving in these aisles. For AGVs to navigate in such a complex spatial structure, reliable obstacle avoidance path planning is essential to avoid collisions and ensure smooth passage.

[0004] Warehousing operations are dynamic, with goods constantly being put into storage, sorted, and shipped out. During this process, the stacking location and space occupied by the goods are constantly changing, and new obstacles may appear at any time. The situation on the AGV's running path is uncertain, so corresponding path planning technology is needed to respond to these changes in real time and ensure that it can flexibly adjust its route to avoid obstacles. Summary of the Invention

[0005] This invention provides an AGV obstacle avoidance path planning method and system for intelligent warehousing, which solves the shortcomings of existing technologies in minimizing losses and the ease with which items can be lost.

[0006] On one hand, the present invention provides an AGV obstacle avoidance path planning method for intelligent warehousing, comprising:

[0007] S1: Scan the warehouse environment to obtain warehouse data.

[0008] S2: Use a topology map to divide the warehouse data into different grids and construct a warehouse space topology map.

[0009] S3: Using the pruning-optimized Dijkstra algorithm, an initial path is generated on the warehouse space topology map based on the set start and end points.

[0010] S4: Set a repulsion factor so that the initial path automatically avoids the area where the obstacle is located based on the repulsion factor and obtains an optimized path.

[0011] S5: Set a safety buffer so that the primary optimized path can be adjusted to deal with minor deviations during the AGV's movement based on the safety buffer, resulting in a secondary optimized path.

[0012] S6: Real-time monitoring of AGV operating status to determine if the environment has changed. If so, based on the secondary optimized path, the current AGV position is used as the new starting point, the target point remains unchanged, and the updated environmental map is used to recalculate the new obstacle avoidance path.

[0013] According to the present invention, an AGV obstacle avoidance path planning method for intelligent warehousing is provided. In step S1, the warehousing data includes obstacle location data, passable channel data, shelf location data, AGV starting point and target point data, and AGV real-time location data.

[0014] According to the AGV obstacle avoidance path planning method for intelligent warehousing provided by the present invention, the specific steps for constructing the warehouse space topology map in step S2 are as follows:

[0015] S21: Analyze the intersections of aisles between various shelves and the connection points between different functional areas based on warehouse data.

[0016] S22: Determine whether there is a passable path between each node based on the intersection node and the connection position. If so, construct an edge between the two nodes and assign edge attribute characteristics to each edge.

[0017] S23: Integrate the nodes and edges to obtain the map framework of the spatial topology.

[0018] S24: Establish a dynamic update mechanism to update the map framework of the initial spatial topology in real time.

[0019] According to the AGV obstacle avoidance path planning method for intelligent warehousing provided by the present invention, the specific steps of finding the path using the pruned and optimized Dijkstra algorithm in step S3 are as follows:

[0020] S31: Add the starting point to the set of visited nodes, initialize a distance array, set the distance from the starting point to itself as the leader, initialize the distance from the starting point to all other nodes as infinity, and record the predecessor node of each node as empty.

[0021] S32: Traverse all adjacent nodes of the starting point, update the distance values ​​of the corresponding adjacent nodes in the distance array according to the weight of the edge from the starting point to the adjacent node, and set the predecessor node of these adjacent nodes as the starting point.

[0022] S33: Select the node with the smallest distance among the unvisited nodes and add it to the visited set.

[0023] S34: Repeat the step of adding nodes to the visited set until the endpoint is added to the visited node set.

[0024] S35: When the endpoint is added to the set of visited nodes, start from the endpoint, backtrack from the predecessor node, and obtain each node from the starting point to the endpoint in sequence. Combine these nodes in order to obtain the initial path.

[0025] According to the AGV obstacle avoidance path planning method for intelligent warehousing provided by the present invention, the specific steps of selecting nodes to add to the set in step S33 include:

[0026] S331: Select the node with the smallest current distance value from the unvisited nodes and add it to the set of visited nodes.

[0027] S332: Re-examine and update all adjacent nodes of the node with the smallest current distance value. Determine whether the distance from all adjacent nodes to the starting point is smaller than the distance from the node with the smallest current distance value to the starting point. If so, update the distance value and add the predecessor node of the adjacent node to the visited set of nodes.

[0028] According to the obstacle avoidance path planning method for AGVs used in intelligent warehousing provided by the present invention, the specific steps of path planning adjustment in step S4 are as follows:

[0029] S41: Define the corresponding repulsion function for different types of obstacles.

[0030] S42: Calculate the repulsive force vector generated by each obstacle based on the repulsive force function, superimpose the repulsive force vectors of all obstacles to obtain the total repulsive force vector, and synthesize the total repulsive force vector with the attractive force vector to obtain the final resultant force direction of the AGV.

[0031] S43: As the AGV moves and the position of obstacles changes, update the distance parameters between each obstacle and the AGV according to the repulsion vector.

[0032] S44: Recalculate the repulsion function value based on the distance parameter, and recalculate the resultant force direction based on the repulsion function value to obtain an optimized path.

[0033] According to the obstacle avoidance path planning method for AGVs used in intelligent warehousing provided by the present invention, in step S42, the repulsion function formula is expressed as:

[0034]

[0035] In the formula, where k rep ρ is the repulsion coefficient, d is the distance between the AGV and obstacle i, and ρ is the radius of the obstacle's influence range. Let be the vector pointing from the obstacle to the AGV. The obstacle's coordinates are (x obs ,y obs ), AGV coordinates are (x agv ,y agv ).

[0036] According to the AGV obstacle avoidance path planning method for intelligent warehousing provided by the present invention, the specific steps for dealing with minor deviations during the AGV's movement in step S5 are as follows:

[0037] S51: Determine the scope of the safety buffer zone based on the storage data.

[0038] S52: Compare the planned position of the primary optimized path with the range of the safety buffer zone to determine whether the AGV's driving has deviated. If so, adjust the primary optimized path according to the deviation and the range of the safety buffer zone to obtain the secondary optimized path.

[0039] The obstacle avoidance path planning method for AGVs in intelligent warehousing provided by the present invention further includes multi-AGV collaborative operation, dynamically adjusting the path of each AGV through a conflict-based search algorithm, planning a preliminary path for each AGV independently, and detecting and judging whether there is a conflict between the paths, and coordinating and replanning if so.

[0040] On the other hand, the present invention also provides an AGV obstacle avoidance path planning system for intelligent warehousing, comprising:

[0041] The data acquisition module is used to scan warehouse data to obtain warehouse data.

[0042] The data partitioning module is used to divide warehouse data into different grids using a topology map, thereby constructing a map that reflects the topological structure of the warehouse space.

[0043] The path building module is used to generate an initial path on a pre-built map using the Dijkstra algorithm with pruning optimization, based on the set start and end points.

[0044] The path optimization module is used to set repulsion factors, so that the initial path can automatically avoid the area where the obstacle is located and obtain an optimized path based on the repulsion factors.

[0045] The secondary optimization module is used to set a safety buffer, so that the primary optimized path can be adjusted according to the safety buffer to deal with minor deviations during the AGV's movement, and a secondary optimized path can be obtained.

[0046] The real-time update module is used to monitor the AGV's operating status in real time and determine whether the environment has changed. If so, it recalculates a new obstacle avoidance path based on the secondary optimized path, taking the current AGV's position as the new starting point and keeping the target point unchanged, combined with the updated environmental map.

[0047] This invention provides an AGV obstacle avoidance path planning system and method for intelligent warehousing. It constructs an initial path using a pruned and optimized Dijkstra algorithm, and optimizes the initial path by setting repulsion factors and safety buffers. The beneficial effects achieved are as follows:

[0048] This study utilizes a topological mapping method to highlight various locations and their connections within the environment. By rationally identifying nodes, constructing edges, and assigning attributes, combined with a dynamic update mechanism, the resulting map accurately reflects the real-time state of the warehouse environment. This provides practical and timely foundational data for path planning, facilitating more precise path design. During path planning, an improved version of the pruned Dijkstra algorithm is employed. This algorithm considers obstacle avoidance and efficiency within the warehouse context to guide the search direction and quickly derive an initial path. By introducing repulsive forces and referencing the artificial potential field method, the path automatically avoids obstacle areas based on defined repulsive functions and resultant force calculations, resulting in a first-order optimized path. A safety buffer zone is established, and by monitoring minor deviations and fine-tuning the path, a second-order optimized path is obtained. Through this series of optimizations, the planned path becomes safer, more efficient, and adaptable to the complexities of the warehouse environment. The improved version of the pruned Dijkstra algorithm, combined with the warehouse context, considers obstacle avoidance and efficiency to guide the search direction and quickly derive an initial path. By introducing repulsive forces and referencing the artificial potential field method, and based on the defined repulsive force function and resultant force calculation, the path can automatically avoid areas containing obstacles, resulting in a primary optimized path. A safety buffer zone is then set up, and through monitoring minor deviations and fine-tuning the path, a secondary optimized path is obtained. Through this series of optimizations, the planned path becomes safer, more efficient, and adaptable to the complexities of the warehouse environment.

[0049] The system monitors the AGV's operating status and environmental changes in real time. Once an environmental change is detected, it can quickly use the current AGV position as a new starting point, determine the target point as needed, recalculate the obstacle avoidance path based on the updated map, and verify and correct its rationality. The accurate path information is then transmitted to the control system to ensure that the AGV can travel along the new path in a timely manner. At the same time, the system continuously and cyclically executes the environmental monitoring and path adjustment process, enabling the AGV to operate continuously, safely, and efficiently in the warehousing environment. This effectively avoids safety accidents such as collisions, greatly improves the overall efficiency and safety of warehouse cargo handling, and enhances the warehousing system's adaptability to complex environmental changes. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0051] Figure 1 This is a flowchart illustrating an AGV obstacle avoidance path planning method for intelligent warehousing provided by an embodiment of the present invention;

[0052] Figure 2 This is a schematic diagram of a module of an AGV obstacle avoidance path planning system for intelligent warehousing provided in an embodiment of the present invention. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0054] The following combination Figure 1-Figure 2 This invention describes an AGV obstacle avoidance path planning method and system for intelligent warehousing.

[0055] Figure 1 This is a flowchart illustrating an AGV obstacle avoidance path planning method for intelligent warehousing provided by an embodiment of the present invention.

[0056] like Figure 1 As shown in the figure, an AGV obstacle avoidance path planning method for intelligent warehousing is provided by an embodiment of the present invention. The method includes:

[0057] S1: Warehouse data includes obstacle location data, passable aisle data, shelf location data, AGV starting and target point data, AGV real-time location data, other AGV location information, aisle feasibility status data, and environmental map update data. Ultrasonic sensors can serve as a supplementary short-range detection method, especially advantageous for detecting obstacles in corners and low-lying areas. Visual sensors identify different types of obstacles, acquiring richer image feature information. These sensors are strategically installed at various locations on the AGV (Automated Guided Vehicle) to ensure comprehensive, blind-spot-free coverage of the area surrounding the AGV, enabling real-time perception of the surrounding environment.

[0058] Obstacle location data includes: the coordinates of the obstacles, static obstacles including shelves and columns, and dynamic obstacles including temporarily stacked goods, forklifts, and other AGVs. The obstacle location data is detected in real time using LiDAR, ultrasonic sensors, and vision sensors.

[0059] Obstacle types include fixed obstacles and moving obstacles, used to define different repulsion functions. The radius of the obstacle's influence range determines the setting of the safety buffer zone.

[0060] The passable passage data includes: coordinates of passage intersection nodes (as nodes on the topology map), passage width, slope, permitted AGV types (as edge attributes), and real-time passability status of the passage. This data is obtained through sensor scanning combined with warehouse layout analysis.

[0061] Shelf location data includes: shelf coordinates, including shelf corners and the junctions between storage and sorting areas; shelf functional areas, including storage, sorting, and shipping areas, used for cross-area path planning. Warehouse layout information is obtained by combining sensor positioning.

[0062] The AGV starting point and target point data include: starting point coordinates, AGV current position, target point coordinates, task instructions or allocation by the warehouse management system.

[0063] AGV real-time location data includes: AGV's current coordinates and orientation, AGV's speed, acceleration, and other motion status.

[0064] Other AGV location information includes: real-time coordinates, direction of travel, and speed of other AGVs, as well as path planning information for each AGV. This information is obtained using the communication module in the multi-AGV collaborative system.

[0065] The passability status data includes: whether the passage is blocked, including temporarily stacked goods and malfunctioning AGVs. Passage difficulty includes width, slope, and the number of temporary obstacles. Real-time monitoring using sensors combined with a dynamic update mechanism is employed.

[0066] Environmental map update data includes: the location and type of new obstacles, removal markers for disappeared obstacles, and the real-time location and task status of other AGVs. Sensor fusion and dynamic map update algorithms are used.

[0067] The collected warehouse data undergoes preprocessing, such as removing noise data caused by sensor malfunctions, unifying coordinates and synchronizing time for data from different sensors to ensure that each data point accurately reflects the actual environmental conditions at the same moment. The processed data is then integrated into a unified format for easier subsequent analysis and judgment. Data preprocessing includes:

[0068] Remove noise data caused by the sensor itself.

[0069] Coordinate unification and time synchronization are performed on data from different sensors.

[0070] The processed data is integrated into a unified format.

[0071] S2: Utilizing information collected by sensors, a topology map is used to mark the corresponding grids as impassable areas based on the locations of obstacles detected by the sensors. Simultaneously, key information such as passable aisles, shelf locations, AGV starting points, and target points are marked, constructing a map that reflects the real-time state of the warehouse environment and providing fundamental data support for subsequent path planning. Furthermore, the map content is continuously updated as the AGV moves and the environment changes, ensuring the map's accuracy and timeliness. Emphasis is placed on describing various locations in the environment and the connections between them. Key location points perceived by sensors, such as aisle intersections and shelf corners, are used as nodes. Passable paths between nodes are abstracted as edges, with each edge carrying attribute information such as distance and passage difficulty.

[0072] The following are the specific steps for constructing a map reflecting the topological structure of a warehouse space using the topological mapping method:

[0073] S21: Sensors installed on the AGV scan the warehouse environment to acquire environmental information and analyze the intersections of aisles between shelves and the connections between different functional areas. These intersections are often key decision points for path selection and are identified as nodes in the topology map. For example, in a warehouse with multiple rows of shelves and crisscrossing aisles, locations such as crossroads and T-junctions are marked as nodes. The connections between different functional areas are important nodes when the AGV changes paths during different task phases. Entrances from the storage area to the sorting area, or transition points from the sorting area to the shipping area, are defined as nodes in the topology map to facilitate subsequent planning of cross-area transport routes.

[0074] The surrounding locations such as pillars, elevator entrances, and fire-fighting facilities in the warehouse can affect the passage of AGVs or are places that AGVs need to bypass or stop at. These can also be used as nodes so that the topology map can more comprehensively cover the key locations in the warehousing environment.

[0075] S22: Based on the information acquired by the sensors and the actual layout of the warehouse, determine whether there are passable passageways between the nodes. If there is a direct connection between two nodes that allows AGVs to travel normally, construct an edge between the two nodes to indicate a direct path connection. Assign relevant attribute features to each edge, including distance, passage difficulty, and permitted AGV types. These attributes help in making more reasonable choices and decisions during subsequent path planning. If a node at a certain passage intersection is detected that leads directly to another area's junction, establish an edge between these two nodes. Distance is determined by actual measurement or by estimating the length of the passageway between the two nodes based on the map scale. Passage difficulty includes factors such as passageway width, slope, and the likelihood of temporary obstacles. For example, narrower passageways are relatively more difficult to traverse, so a higher passage difficulty coefficient can be set. Permitted AGV types include AGVs with different load capacities and sizes; indicate which AGVs can travel on the path based on the passageway conditions.

[0076] S23: Integrate the identified nodes and the edges constructed between them to form a preliminary map framework reflecting the topology of the warehouse space. Use a graph data structure to store this topology map to facilitate subsequent computer queries, calculations, and path planning operations.

[0077] S24: Due to the dynamic nature of the warehousing environment, activities such as stacking and handling of goods may change the passability of passageways. Therefore, a dynamic update mechanism is established. Sensors continuously monitor environmental changes. When a new obstacle is detected blocking a passageway corresponding to a certain edge, or when a previously impassable area becomes passable, the state of the edges in the topology map is updated promptly. This ensures that the map always matches the actual environment, providing a reliable basis for AGV obstacle avoidance path planning.

[0078] S3: Using the Dijkstra algorithm with pruning optimization, an initial path is obtained on the constructed map based on the set start and end points. However, given the complexity of the warehousing environment and the real-time requirements, these algorithms need to be improved and optimized. By improving the heuristic function, it can be made more suitable for the comprehensive consideration of obstacle avoidance and efficiency in the warehousing scenario, guiding the search direction towards a better and safer path.

[0079] S31: Add the starting point to the set of visited nodes, initialize a distance array, set the distance from the starting point to itself as the leader, initialize the distance from the starting point to all other nodes as infinity, and record the predecessor node of each node as empty.

[0080] S32: Traverse all adjacent nodes of the starting point, update the distance values ​​of the corresponding adjacent nodes in the distance array according to the weight of the edge from the starting point to the adjacent node, and set the predecessor node of these adjacent nodes as the starting point.

[0081] S33: Select the node with the smallest distance among the unvisited nodes and add it to the visited set, and update its neighbor distance and predecessor node.

[0082] S331: Select the node with the smallest distance value from the unvisited nodes and add it to the set of visited nodes.

[0083] S332: Recheck and update the distances to the starting point of all adjacent nodes of the node that has just been added to the visited set. Determine whether the distance to its adjacent nodes through the newly added node to the visited set is smaller than the previously recorded distance. If so, update the distance value and update the predecessor node of the corresponding adjacent node to the newly added node to the visited set.

[0084] S34: Repeat the operation of selecting the node with the smallest distance among the unvisited nodes and adding it to the visited set, as well as updating the distances of its adjacent nodes and predecessor nodes, until the endpoint is added to the visited node set.

[0085] S35: Once the destination enters the set of visited nodes, start from the destination and backtrack along the predecessor nodes to obtain each node from the starting point to the destination in sequence. Combine these nodes in order to obtain the initial path.

[0086] S4: Set a repulsion factor so that the initial path automatically avoids the area where the obstacle is located based on the repulsion factor and obtains an optimized path.

[0087] S41: Referring to the idea of ​​the artificial potential field method, define corresponding repulsive force functions for different types of obstacles.

[0088] The formula for the repulsive force function is expressed as:

[0089]

[0090] In the formula, where k rep ρ is the repulsion coefficient, used to adjust the magnitude of the repulsion force; d is the distance between the AGV and obstacle i; and ρ is the radius of the obstacle's influence range. Let be the vector pointing from the obstacle to the AGV. The obstacle's coordinates are (x obs,y obs ), AGV coordinates are (x agv ,y agv ).

[0091] S42: Based on the direction of the line connecting the target and the AGV as the direction of gravity, calculate the resultant direction of the repulsive forces generated by each obstacle. Superimpose the repulsive force vectors calculated from the repulsive force function of each obstacle to obtain the total repulsive force vector. Then, combine this vector with the gravity vector to determine the final resultant force direction acting on the AGV. This guides the AGV's path planning, allowing it to automatically avoid areas containing obstacles. For example, if the target is directly in front of the AGV, with the gravity direction forward, but there is an obstacle on the right generating a rightward repulsive force, after vector synthesis, the AGV's path will adjust to a direction that both approaches the target and avoids the obstacle on the right.

[0092] The formula for calculating the repulsive force vector of each obstacle i is expressed as follows:

[0093]

[0094] in, =d is the distance between the AGV and obstacle i. and These are the components of the repulsive force vector along the x-axis and y-axis, respectively.

[0095]

[0096]

[0097] In the formula, Let x be the repulsive force function, and let the obstacle coordinates be (x, y). obs ,y obs ), AGV coordinates are (x agv ,y agv ).

[0098] The formula for the total repulsive force vector is expressed as:

[0099]

[0100] In the formula, n is the total number of obstacles.

[0101] By combining the total repulsive force vector and the attractive force vector, we obtain the final resultant force vector Ftotal acting on the AGV:

[0102]

[0103] In the formula, The gravitational vector, .

[0104] S43: As the AGV moves and the position of obstacles changes, the distance and relative direction between each obstacle and the AGV are continuously updated. Then, the repulsion function value and resultant force direction are dynamically recalculated to ensure that the path planning can adapt to environmental changes in real time and always guide the AGV to avoid detected obstacles.

[0105] S5: Set a safety buffer so that the primary optimized path can be adjusted to deal with minor deviations during the AGV's movement based on the safety buffer, resulting in a secondary optimized path.

[0106] S51: Determine the scope of the safety buffer based on the warehouse data, and set it around the primary optimization path so that the safety buffer is closely related to the primary optimization path, and the primary optimization path becomes the reference basis for setting the safety buffer.

[0107] S52: When the AGV travels along the optimized path, its position changes are monitored in real time. The actual travel position is compared with the optimized path and the set safety buffer boundary to determine whether there is a slight deviation. The optimized path and safety buffer are the key references for judging the deviation.

[0108] S53: Once a minor deviation is detected and it is within the safety buffer, the path is fine-tuned based on the specific direction and distance of the deviation, the range of the safety buffer, and the original direction of the primary optimization path. By continuously making dynamic adjustments based on the deviation, a secondary optimization path that can handle minor deviations is finally obtained. The primary optimization path, safety buffer, and detected deviations mentioned in the previous steps are all important inputs for obtaining the secondary optimization path.

[0109] S6: Real-time monitoring of AGV operating status to determine if the environment has changed. If so, based on the secondary optimized path, the current AGV position is used as the new starting point, the target point remains unchanged, and the updated environmental map is used to recalculate the new obstacle avoidance path.

[0110] The AGV's surrounding environment is continuously monitored by sensors at a high frequency. Once an environmental change is detected—such as the appearance of new obstacles, other AGVs changing routes, or previously passable passages being blocked—a path adjustment process is immediately triggered. Upon triggering the path adjustment request, an optimized path planning algorithm quickly recalculates a feasible obstacle avoidance path, using the current AGV position as the new starting point and the target point unchanged, combined with an updated environmental map. This recalculation process must be completed within a very short time to ensure the AGV can travel along the new path promptly, avoiding collisions and other safety accidents. The environmental map is immediately updated based on the latest sensor data. If a grid map is used, newly appearing obstacles are marked as impassable, and the impassable markers for previously obscured obstacles are removed. Simultaneously, the position information of other AGVs and the passability status of passages are updated to ensure the environmental map reflects the current warehouse conditions in real time.

[0111] The current actual position of the AGV is clearly defined as the starting point for the new path planning, and its coordinate position in the environmental map is accurately obtained through the AGV's own positioning system.

[0112] If the AGV's mission objective remains unchanged due to environmental changes, the original objective point will remain unchanged. If the mission objective changes due to special circumstances, a new objective point will be determined according to the new mission instructions, and the objective point marker in the environmental map will be updated simultaneously.

[0113] The generated path undergoes feasibility verification, checking for issues such as traversing impassable areas or conflicting with other AGVs' planned paths. If any issues are found, algorithm parameters are adjusted promptly, or alternative path optimization strategies are employed to correct the problem until a feasible obstacle avoidance path that meets the requirements is obtained. The recalculated obstacle avoidance path information is then transmitted to the AGV's control system. This path information includes detailed instructions such as the coordinates of each path node, turning angle, and desired speed, enabling the AGV's control system to accurately understand and execute the new path planning. Based on the path information, the control system adjusts the AGV's drive wheel speed and steering mechanism in real time to ensure the AGV can quickly travel along the new path. Simultaneously, it continues to monitor the environment, preparing for potential environmental changes. This environmental monitoring and path adjustment process is repeated cyclically to ensure the continuous, safe, and efficient operation of the AGV in the warehouse environment.

[0114] In summary, this embodiment provides an AGV obstacle avoidance path planning method for intelligent warehousing. It constructs an initial path using a pruned and optimized Dijkstra algorithm, and optimizes the initial path by setting repulsion factors and safety buffers. The beneficial effects achieved are as follows:

[0115] This study utilizes a topological mapping method to highlight various locations and their connections within the environment. By rationally identifying nodes, constructing edges, and assigning attributes, combined with a dynamic update mechanism, the resulting map accurately reflects the real-time state of the warehouse environment. This provides practical and timely foundational data for path planning, facilitating more precise path design. During path planning, an improved version of the pruned Dijkstra algorithm is employed. This algorithm considers obstacle avoidance and efficiency within the warehouse context to guide the search direction and quickly derive an initial path. By introducing repulsive forces and referencing the artificial potential field method, the path automatically avoids obstacle areas based on defined repulsive functions and resultant force calculations, resulting in a first-order optimized path. A safety buffer zone is established, and by monitoring minor deviations and fine-tuning the path, a second-order optimized path is obtained. Through this series of optimizations, the planned path becomes safer, more efficient, and adaptable to the complexities of the warehouse environment. The improved version of the pruned Dijkstra algorithm, combined with the warehouse context, considers obstacle avoidance and efficiency to guide the search direction and quickly derive an initial path. By introducing repulsive forces and referencing the artificial potential field method, and based on the defined repulsive force function and resultant force calculation, the path can automatically avoid areas containing obstacles, resulting in a primary optimized path. A safety buffer zone is then set up, and through monitoring minor deviations and fine-tuning the path, a secondary optimized path is obtained. Through this series of optimizations, the planned path becomes safer, more efficient, and adaptable to the complexities of the warehouse environment.

[0116] Based on the same general inventive concept, this invention also protects an AGV obstacle avoidance path planning system for intelligent warehousing. The following describes an AGV obstacle avoidance path planning system for intelligent warehousing provided by this invention. The AGV obstacle avoidance path planning system for intelligent warehousing described below can be referred to in correspondence with the AGV obstacle avoidance path planning method for intelligent warehousing described above.

[0117] Figure 2 This is a schematic diagram of a module of an AGV obstacle avoidance path planning system for intelligent warehousing provided in an embodiment of the present invention.

[0118] like Figure 2 As shown, an AGV obstacle avoidance path planning system for intelligent warehousing includes:

[0119] The data acquisition module is used to scan warehouse data to obtain warehouse data.

[0120] The data partitioning module is used to divide warehouse data into different grids using a topology map, thereby constructing a map that reflects the topological structure of the warehouse space.

[0121] The path building module is used to generate an initial path on a pre-built map using the Dijkstra algorithm with pruning optimization, based on the set start and end points.

[0122] The path optimization module is used to set repulsion factors, so that the initial path can automatically avoid the area where the obstacle is located and obtain an optimized path based on the repulsion factors.

[0123] The secondary optimization module is used to set a safety buffer, so that the primary optimized path can be adjusted according to the safety buffer to deal with minor deviations during the AGV's movement, and a secondary optimized path can be obtained.

[0124] The real-time update module is used to monitor the AGV's operating status in real time and determine whether the environment has changed. If so, it recalculates a new obstacle avoidance path based on the secondary optimized path, taking the current AGV's position as the new starting point and keeping the target point unchanged, combined with the updated environmental map.

[0125] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, are embodied in the form of a software product. This computer software product is stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. These modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An obstacle avoidance path planning method for AGVs used in intelligent warehousing, characterized in that, include: S1: Scan the warehouse environment to obtain warehouse data; S2: Use a topology map to divide the warehouse data into different grids and construct a warehouse space topology map; S21: Analyze the intersections of aisles between various shelves and the connection points between different functional areas based on the warehouse data; S22: Determine whether there is a passable passage between each node based on the intersection node and the connection position. If so, construct an edge between the two nodes and assign edge attribute characteristics to each edge. S23: Integrate the nodes and edges to obtain a map framework with a spatial topology; S24: Establish a dynamic update mechanism to update the map framework of the initial spatial topology in real time; S3: Using the pruning-optimized Dijkstra algorithm, an initial path is generated on the warehouse space topology map based on the set start and end points; S4: Set a repulsion factor so that the initial path automatically avoids the area where the obstacle is located based on the repulsion factor and obtains an optimized path; S5: Set a safety buffer zone so that the primary optimized path can be adjusted to accommodate minor deviations during AGV travel, resulting in a secondary optimized path. S51: Determine the range of the safety buffer zone based on the storage data; S52: Compare the planned position of the primary optimized path with the range of the safety buffer zone to determine whether the AGV's driving has deviated. If so, adjust the primary optimized path according to the deviation and the range of the safety buffer zone to obtain the secondary optimized path. S53: Once a minor deviation is detected and the deviation is within the safety buffer, the path is fine-tuned based on the specific direction and distance information of the deviation, combined with the range limitation of the safety buffer and the original direction of the primary optimization path. By continuously adjusting dynamically according to the deviation, a secondary optimization path that can cope with minor deviations is finally obtained. S6: Monitor the AGV's operating status in real time to determine if the environment has changed. If so, based on the secondary optimization path, take the current AGV's position as the new starting point, keep the target point unchanged, and recalculate the new obstacle avoidance path in conjunction with the updated environment map.

2. The AGV obstacle avoidance path planning method for intelligent warehousing according to claim 1, characterized in that, In step S1, the warehouse data includes obstacle location data, passable passage data, shelf location data, AGV starting point and target point data, and AGV real-time location data.

3. The AGV obstacle avoidance path planning method for intelligent warehousing according to claim 1, characterized in that, In step S3, the specific steps for finding the path using the pruning-optimized Dijkstra's algorithm are as follows: S31: Add the starting point to the set of visited nodes, initialize a distance array, set the distance from the starting point to itself as the leader, initialize the distance from the starting point to all other nodes as infinity, and record the predecessor node of each node as empty; S32: Traverse all adjacent nodes of the starting point, update the distance values ​​of the corresponding adjacent nodes in the distance array according to the weight of the edge from the starting point to the adjacent nodes, and set the predecessor node of these adjacent nodes as the starting point. S33: Select the node with the smallest distance among the unvisited nodes and add it to the visited set; S34: Repeat the step of adding nodes to the visited set until the endpoint is added to the visited node set; S35: When the endpoint is added to the set of visited nodes, start from the endpoint, backtrack from the predecessor node, and obtain each node from the starting point to the endpoint in sequence. Combine these nodes in order to obtain the initial path.

4. The AGV obstacle avoidance path planning method for intelligent warehousing according to claim 3, characterized in that, In step S33, the specific steps for selecting nodes to add to the set include: S331: Select the node with the smallest current distance value from the unvisited nodes and add it to the set of visited nodes; S332: Re-examine and update all adjacent nodes of the node with the smallest current distance value, and determine whether the distance from all adjacent nodes to the starting point is smaller than the distance from the node with the smallest current distance value to the starting point. If so, update the distance value and add the predecessor node of the adjacent node to the visited set of nodes.

5. The AGV obstacle avoidance path planning method for intelligent warehousing according to claim 1, characterized in that, The specific steps for path planning adjustment in step S4 are as follows: S41: Define the corresponding repulsion function for different types of obstacles; S42: Calculate the repulsive force vector generated by each obstacle according to the repulsive force function, superimpose the repulsive force vectors of all obstacles to obtain the total repulsive force vector, and combine the total repulsive force vector with the attraction force vector to obtain the final resultant force direction of the AGV; S43: As the AGV moves and the position of the obstacles changes, update the distance parameters between each obstacle and the AGV according to the repulsion vector; S44: Recalculate the repulsion function value based on the distance parameter, and recalculate the resultant force direction based on the repulsion function value to obtain an optimized path.

6. The AGV obstacle avoidance path planning method for intelligent warehousing according to claim 5, characterized in that, In step S42, the repulsive force function formula is expressed as: In the formula, k rep ρ is the repulsion coefficient, d is the distance between the AGV and obstacle i, and ρ is the radius of the obstacle's influence range. Let be the vector pointing from the obstacle to the AGV. The obstacle's coordinates are (x obs y obs ), AGV coordinates are (x agv y agv ).

7. The AGV obstacle avoidance path planning method for intelligent warehousing according to claim 1, characterized in that, It also includes multi-AGV collaborative operation, dynamically adjusting the paths of each AGV through a conflict-based search algorithm, planning an initial path for each AGV independently, and detecting whether there are conflicts between the paths. If so, coordination and replanning are carried out.

8. An AGV obstacle avoidance path planning system for intelligent warehousing, which is applied to the AGV obstacle avoidance path planning method for intelligent warehousing as described in any one of claims 1 to 7, characterized in that, The path planning system includes: The data acquisition module is used to scan warehouse data to obtain warehouse data; The data partitioning module is used to divide the warehouse data into different grids using a topology map, and to construct a map that reflects the topological structure of the warehouse space. The path building module is used to obtain the initial path on the constructed map based on the set start and end points using the Dijkstra algorithm with pruning optimization. The path optimization module is used to set repulsion factors, so that the initial path can automatically avoid the area where the obstacle is located and obtain an optimized path according to the repulsion factors. The secondary optimization module is used to set a safety buffer, so that the primary optimized path can be adjusted according to the safety buffer to deal with minor deviations during the AGV's movement, and a secondary optimized path can be obtained. The real-time update module is used to monitor the AGV's operating status in real time and determine whether the environment has changed. If so, based on the secondary optimization path, the current position of the AGV is used as the new starting point, the target point remains unchanged, and the updated environmental map is used to recalculate the new obstacle avoidance path.

Images