Logistics warehousing system

Abstract

The invention relates to the technical field of logistics storage management, and discloses a logistics storage system, which comprises an environment sensing module used for monitoring task points, path states, obstacle positions and dynamic order changes in a storage environment in real time, and generating path weight data; the master control module is connected with the environment sensing module and is used for receiving the order demand, generating a task point set, obtaining path weight data and carrying out global scheduling on the transfer robot according to the path weight data; and the path optimization module is connected with the master control module and is used for calculating the optimal path of the transfer robot based on the multi-agent game model. According to the technical scheme, the environment sensing module is adopted to perform real-time modeling on the storage environment, and the path weight data is generated, so that the technical effect of dynamically updating the path information is achieved, and the defect of path planning lag in the complex storage environment is overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics warehousing management, and specifically to a logistics warehousing system. Background Art

[0002] With the rapid development of the logistics industry, the requirements for the efficiency and intelligent level of warehousing management are constantly increasing. Modern logistics warehousing systems widely use handling robots for sorting, handling, and storing goods. By combining information technology and automated equipment, the overall efficiency of warehousing operations can be significantly improved. In a complex and changing warehousing environment, how to achieve efficient path planning and task scheduling is the key to optimizing the logistics warehousing system.

[0003] In the prior art, some logistics warehousing systems usually adopt path planning methods based on fixed rules or static data. For example, during the task scheduling process, the warehousing system performs global scheduling and path planning for handling robots through pre-set task points and established path weights. Although these systems can complete a certain degree of automated operations, when facing dynamic environmental changes (such as blocked paths, increased or adjusted task points, etc.), their response capabilities are poor, often requiring manual intervention or re-planning the global path, resulting in a decrease in handling efficiency. In addition, during the task execution process, handling robots lack the ability to adapt to the dynamic environment in real time and may wait for a long time or be unable to complete tasks due to blocked paths.

[0004] The main problems existing in the prior art are mainly the lack of effective response capabilities to the dynamic warehousing environment. In a complex warehousing scenario, when the path becomes impassable due to obstacles or robot congestion, the existing logistics warehousing systems cannot update the path planning in a timely manner, nor can they dynamically adjust the robot task points and path selection, resulting in low task execution efficiency and affecting the overall operation effect of the system. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the present invention provides a logistics warehousing system to solve the problem that the logistics warehousing system in the prior art cannot quickly respond to dynamic environmental changes and optimize in real time.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: A logistics warehousing system, comprising: an environment perception module, configured to monitor in real time task points, path status, obstacle positions, and dynamic order changes in the warehousing environment, and generate path weight data; A general control module, connected to the environment perception module, configured to receive order requirements, generate a set of task points, and obtain path weight data, and perform global scheduling on handling robots according to the path weight data; The path optimization module, connected to the total control module, is used to calculate the optimal path of the handling robot based on the multi-agent game model; the handling robot, connected to the path optimization module and the total control module, is used to receive the optimal path instruction generated by the path optimization module and execute tasks, and at the same time feedback the execution status to the total control module; The task monitoring module, respectively connected to the total control module and the handling robot, is used to monitor the status of the handling robot and the task progress, and feedback the dynamic change information to the total control module; the path optimization module optimizes the path of the handling robot according to the total path distance, path time cost and path conflict cost, and the total control module generates the task allocation of the handling robot based on the optimization result.

[0007] Preferably, the environment perception module models the warehousing environment as a directed weighted graph, including a set of task points and a set of paths, and sends the path weight data to the total control module and the path optimization module.

[0008] Preferably, the path weight is calculated based on the path length, time cost and path conflict cost. The path conflict cost is determined by the ratio of the current occupied state of the path to the passable capacity of the path. The calculation result is updated in real time by the environment perception module and sent to the total control module and the path optimization module.

[0009] Preferably, the path optimization module performs path planning based on the multi-agent non-cooperative game model. The total control module regards the handling robot as an independent agent and sends the set of task points to the path optimization module. The path optimization module calculates the optimal path according to the utility function of the handling robot.

[0010] Preferably, the path optimization module is connected to the total control module, and is used to calculate the optimal path of the handling robot based on the path weight data and the utility function, generate the path selection result through iterative optimization, and send the optimal path to the total control module.

[0011] Preferably, the path optimization module is connected to the total control module, and the path optimization module optimizes the path selection of the handling robot by the gradient projection method, where: Receive the set of task points and path weight data sent by the total control module; Based on the set of task points and path weight data, calculate the gradient of the path utility function of each handling robot; Adjust the path selection of each handling robot, and map the adjustment result to the set of feasible paths through projection operation; After the change amount of the utility function of the path optimization module is less than the set convergence threshold, send the optimal path result of the handling robot to the total control module.

[0012] Preferably, the handling robot is connected to the environment perception module through an obstacle perception module, and can detect the blocked state of the path in real time and trigger local path replanning, which is recalculated by the path optimization module based on path weights.

[0013] Preferably, during the local path replanning process, the handling robot preferentially selects the path with the lowest weight for replacement, and feeds back the path adjustment status to the task monitoring module and the total control module through the obstacle perception module.

[0014] Preferably, the total control module is connected to the environment perception module and the path optimization module. When the environment perception module detects a major change in the environment, the total control module triggers global path replanning, which includes: the environment perception module regenerates the warehousing environment model, including updating the task point set and the path set, and recalculating the path weight data; The total control module receives the updated path weight data and generates a new task point set and task priorities; Based on the updated task point set and path weight data, re-optimize the path selection of all handling robots; Send the optimized optimal path to the total control module, and the total control module sends the new task assignment result to the corresponding handling robot.

[0015] Preferably, the task monitoring module is bidirectionally connected to the total control module. After the handling robot completes the task, it feeds back the task status through the task monitoring module, and the total control module generates a new task point set according to the task completion status and triggers the path optimization module for the next round of optimization.

[0016] The present invention provides a logistics warehousing system. It has the following beneficial effects: 1. The present invention adopts the technical solution of using an environment perception module to perform real-time modeling on the warehousing environment and generate path weight data, achieving the technical effect of dynamically updating path information. Compared with the prior art in which the acquisition and processing speed of environmental data is slow, resulting in the inability to quickly respond to environmental changes, it solves the deficiency of lagging path planning in a complex warehousing environment.

[0017] 2. The present invention calculates the optimal path of the handling robot based on a multi-agent game model through the path optimization module, achieving the technical effects of reasonable task allocation and minimizing path conflicts of the handling robot. Different from the static path planning scheme usually adopted in the prior art, the present invention can dynamically adjust the path selection of the robot, solving the deficiency in the prior art where the efficiency decreases due to robot congestion.

[0018] 3. The technical solution of the present invention combines a handling robot with an obstacle perception module for local path adjustment, achieving the technical effect that the robot can still complete the handling task even when blocked. Compared with the limitation in the prior art that the handling robot needs to wait for global planning when the path is blocked, the present invention significantly improves the flexibility of the robot to execute tasks and solves the deficiency of task interruption caused by blocked paths. Brief Description of the Drawings

[0019] Figure 1 It is a schematic diagram of the overall system flow of the present invention; Figure 2 It is a schematic diagram of the path optimization process of the present invention; Figure 3 It is a schematic diagram of the local path replanning process of the present invention; Figure 4 It is a schematic diagram of the global path replanning process of the present invention. Detailed Embodiment

[0020] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0021] Please refer to the attached Figures 1-4 , and the embodiment of the present invention provides a logistics warehousing system, including.

[0022] 1. Environment perception module The environment perception module conducts data interaction with the total control module, the path optimization module, and the handling robot, and real-time collects and processes the dynamic information in the warehousing environment. Its function is to provide real-time and accurate environment data support for the path planning and global scheduling of the system.

[0023] Generally, the changes in the warehousing environment are unpredictable. For example, the temporary adjustment of the shelf layout, the dynamic appearance of obstacles, the increase or decrease of task points, etc. Therefore, the environment perception module needs to have the ability of dynamic update and real-time data collection to ensure that the system can continuously adapt to the complex and changeable environment.

[0024] As an option, the environment perception module collects the task point location, path status, obstacle distribution, and dynamic order information through various sensors deployed in the warehouse, such as lidar, ultrasonic sensors, RFID readers, cameras, etc. These sensors can be combined and used according to the scene requirements to meet the requirements of different warehouse scales and environments.

[0025] Specifically, the environment perception module models the collected environment information as a directed weighted graph G = (V, E). In this model: V = {v1, v2,..., v n} represents the set of task points, where the task points can include shelf positions, cargo storage positions, order sorting areas, and handling endpoints.

[0026] E = {e ij} represents the set of paths, where each path e ij represents a passable path between task points v i and v j .

[0027] The weight w(e ij ) of each path e ij is used to represent the comprehensive cost of path passage. The calculation formula for the weight is: w(e ij ) = λ1·d(e ij ) + λ2·t(e ij ) + λ3·c(e ij ) Where: d(e ij ) represents the geometric distance of the path, such as the Euclidean distance between two points; t(e ij ) represents the estimated time consumption of the path, which can be calculated based on historical data, path length, and robot speed; c(e ij ) represents the path conflict cost, and its definition is: Where current load (e ij ) represents the current occupancy status of the path max capacity (e ij ) represents the maximum passable capacity of the path.

[0028] In this embodiment, the environment perception module sends the collected path weight data to the total control module and the path optimization module in real time. These path weight data can reflect the dynamic state of the current warehousing environment, for example: If the current occupancy status of a certain path is relatively high (such as an increase in robot density or the path being partially blocked by obstacles ), its conflict cost c(e ij ) will increase significantly, thereby increasing the path weight If a certain path is short and has smooth passage, its weight will be relatively low, facilitating the path optimization module to preferentially select this path.

[0029] As a possible implementation, the environment perception module can preprocess the collected information through a multi-level data processing logic. For example: performing point cloud clustering analysis on lidar data to identify dynamic obstacles on the path; performing object detection on camera image data to monitor the smoothness of the warehousing passage; performing fast matching on RFID data to locate the precise positions of goods and task points. In some other embodiments, the environment perception module can combine historical data and prediction algorithms to perform short-term prediction on the path state. For example: predicting future conflicts on the path by analyzing the historical occupancy status of the path and the current task distribution; dynamically adjusting the calculation weight coefficients λ1, λ2, λ3 of the path weight according to the order distribution and the number of handling robots to make it adapt to specific task requirements. In this embodiment, the dynamic update ability of the environment perception module is particularly important. When the warehousing environment changes, such as the addition of obstacles or a sudden increase in order demand, the environment perception module will automatically regenerate the directed weighted graph and update the path weight. The specific process is as follows: First, the sensor collects the latest environment information, such as detecting that an obstacle has appeared on a certain path; Then, the system sets the path weight w(e ij of the path where the obstacle is located ij ) to a relatively high value (close to infinity); Next, the updated directed weighted graph is sent to the total control module and the path optimization module; Finally, the total control module triggers the path optimization module to re-optimize the path of the handling robot according to the updated environment information.

[0030] To further improve the data processing efficiency, the environment perception module can adopt a distributed architecture. For example: In a large-scale warehousing scenario, the environment perception module can divide the warehousing environment into several sub-regions, and each sub-region corresponds to an independent perception node; Each perception node respectively collects the environment information of its responsible area, and after edge computing processing, uploads the result to the central server; The central server integrates the data of each region, generates a global directed weighted graph and updates it to the system.

[0031] Generally, the environment perception module needs to meet the design requirements of high real-time performance, high precision and high adaptability. For example: The calculation delay of the path weight should be controlled within 100 milliseconds to ensure the real-time performance of the data; The measurement accuracy of the path length should be controlled within ±1 centimeter to meet the requirements of high-precision path planning; The system should support processing no less than 1000 path data per second to meet the requirements of high-frequency task environments.

[0032] 2. Total Control Module The total control module receives path weight data and a set of task points from the environmental perception module, generates scheduling tasks in combination with the current order requirements, simultaneously calls the path optimization module for optimal path calculation, and allocates the calculation results to the handling robots for execution.

[0033] During the task allocation process, the total control module can dynamically adjust priorities and optimize the scheduling strategy in real-time according to the status of the handling robots. The total control module maintains information synchronization with each module through a two-way communication mechanism to ensure the efficient operation of the system in a dynamic environment.

[0034] The scheduling logic of the total control module is based on path weights, and through the set of task points and the status data of the handling robots, it reasonably allocates tasks. Through cooperation with the path optimization module, the total control module can achieve efficient allocation of multiple tasks and global path optimization.

[0035] In this embodiment, the functions of the total control module can be divided into four parts: data reception, task generation, scheduling optimization, and task allocation, which are specifically as follows: In terms of data reception, the total control module receives real-time path weight data and a set of task points from the environmental perception module. The set of task points V = {v1, v2,..., v n} represents all the target points that the current system needs to process, such as cargo storage locations, sorting endpoints, etc.; the path weight set W = {w(e ij )} contains the comprehensive weight information of all paths. The path weight w(e ij ) is dynamically calculated by the environmental perception module according to the following formula: w(e ij ) = λ1·d(e ij ) + λ2·t(e ij ) + λ3·c(e ij ) Where: d(e ij ) represents the geometric distance of the path; t(e ij ) represents the estimated time consumption; c(e ij ) represents the path conflict cost, which is defined as: λ1, λ2, and λ3 are weight coefficients used to balance the influence of different attributes of the path on the total weight.

[0036] In terms of task generation, the total control module generates a priority queue of task points according to the order requirements. Generally, the priority of a task is comprehensively evaluated based on the order deadline, the importance of the goods, and the path weight. As an option, the total control module can use the following formula to calculate the task priority p(v i ): t d is the deadline of the task, w min is the minimum path weight to reach the task point v i ; μ1 and μ2 are scheduling parameters.

[0037] In terms of scheduling optimization, the total control module globally optimally allocates tasks based on the current state of the handling robots, combined with task priorities and path weights. Specifically, the total control module calls the path optimization module, treats each handling robot as an independent agent, and models the task allocation problem as a multi-agent non-cooperative game model. In this case, the handling robots need to select the optimal path to minimize the total path distance, time cost, and path conflict cost. The optimal path result returned by the path optimization module is received by the total control module for task matching.

[0038] During the task allocation process, the total control module sends the optimized path and task information to the handling robots through a two-way communication mechanism, and at the same time receives the task execution status feedback from the handling robots. For example, when a handling robot completes the operation of a task point, the total control module updates the priorities of the remaining task points according to the task status and triggers the path optimization module to recalculate the optimal path The total control module can also optimize the task scheduling by combining historical data. For example: Predict the processing time distribution of task points based on historical order data, so as to pre-allocate high-priority tasks in advance.

[0039] Can adjust the weight coefficients λ1, λ2, λ3 according to the historical driving paths of the handling robots, and preferentially allocate paths with higher passing efficiency.

[0040] Combine the order distribution and the warehouse layout to dynamically adjust the processing order of task points to reduce the overall path length.

[0041] Adopt a distributed architecture to improve the efficiency of task allocation. For example: In a large-scale warehouse scenario, the total control module can divide the task points into several subsets by region, and each subset is responsible for scheduling by a sub-control unit; The sub-control units can process the path weights and task priorities within their respective regions through edge computing and upload the calculation results to the total control module; The total control module integrates and generates a global task allocation plan based on the results provided by each sub-control unit.

[0042] As an option, the total control module can also combine dynamic environmental changes to trigger global path replanning in real time. For example: When the environmental perception module detects a major blockage on a certain path, the total control module sets the weight w(e ij ) of this path to a relatively high value; The total control module simultaneously regenerates the set of task points and calls the path optimization module to perform global optimization calculations.

[0043] 3. Path Optimization Module The path optimization module outputs a path planning result that meets the global optimization goal of the system by constructing and solving a multi-agent game model and combining the gradient projection method. Generally, the path optimization module needs to fully consider the total path distance, time cost, and path conflict cost, dynamically balance the task conflicts between multiple handling robots, and achieve efficient cooperation of tasks.

[0044] The path optimization module can quickly generate an optimal path result that meets the current task requirements in a dynamic storage environment through a real-time iterative optimization algorithm. Through close cooperation with the total control module, the path optimization module can adapt to changes in task priorities and ensure the timely completion of high-priority tasks.

[0045] After receiving the set of task points and path weight data, the path optimization module regards the handling robots as independent agents, optimizes the path selection of each agent through a game model, and sends the optimized path planning result to the total control module.

[0046] In this embodiment, the path calculation logic of the path optimization module is based on a multi-agent non-cooperative game model. In this model: Each handling robot A i is modeled as an agent, and its strategy space is the set of all possible paths for it

[0047] Each path represents the passing plan of the robot from the task start point to the end point.

[0048] The path selections P = [P1, P2,..., P m of all agents constitute the global path plan of the system. The goal of each agent is to select an optimal path to maximize its utility function U i (Pi , P -i ) is maximized, where P -i represents the path selection of other agents. Generally, the utility function is defined as follows: U i (P i , P -i ) = -(α i · D i (P i ) + β i · T i (P i ) + γ i · C i (P i , P -i )) where: D i (P i ) is the total distance of path P i , calculated by accumulating path weights; T i (P i ) is the time cost of path P i , estimated based on the path length and the speed of the handling robot; C i (P i , P -i ) is the conflict cost of path P i , depending on the degree of overlap of the paths selected by other agents; α i , β i , γ i are the weight coefficients of the total path distance, time cost, and conflict cost, respectively.

[0049] In a possible implementation, the path optimization module optimizes the utility function through the gradient projection method. The specific process includes the following steps: First, initialize the path selection of each handling robot according to the task point set and path weight data

[0050] Then, calculate the gradient of the path utility function Update the path selection P i : where: η is the gradient descent step size, controlling the convergence speed of the optimization is the projection operator, used to map the updated path selection to the path set During the path optimization process, to avoid conflicts in path selection among agents, the path optimization module dynamically adjusts the conflict cost C of each path i (P i ,P -i ). As an option, the conflict cost can be calculated by the following formula: Where: current_load(e ij ) represents the current load of path segment e i j; max_capacity(e ij ) represents the maximum passable capacity of the path segment.

[0051] After the calculation is completed, the optimization result is sent to the total control module. After receiving the result of the path optimization module, the total control module assigns tasks based on task priorities to ensure that paths for high-priority tasks are preferentially allocated.

[0052] The path optimization module can dynamically adjust the optimization parameters in combination with historical data. For example: · Adjust the parameters λ1, λ2, λ3 in the path weight calculation formula according to the historical task distribution to adapt to different warehousing environments; Utilize the historical record of path congestion to dynamically predict future path conflicts, thereby improving the accuracy and efficiency of optimization.

[0053] It can perform differential processing on different types of handling tasks. For example: For high-priority orders that need to be processed quickly, the path optimization module can increase the weight β of time cost i , and preferentially select the path with the shortest time consumption; For general handling tasks, the path optimization module can focus on optimizing the total path distance to reduce the energy consumption of the handling robot.

[0054] The path optimization module in this embodiment can perform real-time adjustment on the optimization result in a dynamic environment. When the environment perception module detects a change in the warehousing environment (such as the addition of obstacles or path blockage), the path optimization module recalculates the weight data of the affected paths and re-executes the optimization process according to the updated environment model. Generally, this kind of real-time adjustment requires the path optimization module to have high computational efficiency and be able to complete path update within 100 milliseconds.

[0055] 4. Handling Robot The handling robot maintains data interaction with the path optimization module and the total control module, executes the handling task according to the optimal path instruction calculated by the path optimization module, and feeds back its own running state and task completion status to the total control module. Generally, the handling robot has functions such as path execution, obstacle perception, local path adjustment, and status feedback, and can flexibly adjust the path in a dynamic environment, thus ensuring the smooth completion of the task.

[0056] The handling robot interacts with other modules of the system through an embedded control system and is equipped with sensors to perceive the environment. Its core goal is to efficiently and safely complete the specified handling task, and at the same time dynamically respond through the local path adjustment function when the path is blocked.

[0057] Specifically, the handling robot receives the task points assigned by the total control module and the path planning results provided by the path optimization module. Based on this data, the handling robot can autonomously complete the handling operation from the starting point to the ending point.

[0058] In this embodiment, the handling robot has the following main functions: First, the handling robot executes the optimal path P provided by the path optimization module through the built-in path tracking algorithm i . During the execution of the path, it is necessary to compare the current position information and path points in real time to ensure that the robot moves accurately along the planned path. The path tracking algorithm of the handling robot can adopt a proportional-integral-derivative (PID) controller, and the specific formula is as follows: Where: u(t) is the control output, which is used to adjust the moving direction of the handling robot; e(t) is the current path deviation; K p ,K i ,K d are the proportional, integral, and derivative gain parameters respectively.

[0059] Secondly, the handling robot monitors the dynamic changes on the path in real time through the obstacle perception module. For example, when an obstacle appears on the path, the obstacle perception module will judge the position and size of the obstacle and trigger the local path adjustment function.

[0060] When the handling robot makes local path adjustments, it recalculates the alternative path by interacting with the path optimization module.

[0061] The local path adjustment function is based on the path weight and preferentially selects the feasible path with the lowest weight.

[0062] For example, when the weight w(e ij of a certain path e ij)When it increases significantly (e.g., the cost of conflict caused by obstacles), the handling robot will automatically avoid this path and re-plan the path to reach the target task point.

[0063] The handling robot can also improve the accuracy of obstacle perception through multi-sensor fusion technology. For example: Use the method of combining lidar and camera to achieve accurate identification of static and dynamic obstacles; Use ultrasonic sensors to detect obstacles at close range to improve the obstacle avoidance response speed.

[0064] The handling robot maintains two-way communication with the master control module and the task monitoring module through a wireless communication module. Specifically, the handling robot will send its current position, path status, and task execution progress to the task monitoring module in real time for the master control module to perform dynamic scheduling.

[0065] 5. Task monitoring module The task monitoring module is mainly responsible for real-time monitoring of the task execution status of the handling robot and feeding back the monitoring data to the master control module. The task monitoring module maintains close contact with the handling robot and ensures the consistency of the system's task scheduling and execution through full tracking of the task execution process.

[0066] Generally, the task monitoring module receives the status information fed back by the handling robot through a wireless communication interface, including the current position, task point completion status, and path execution situation. As an option, the task monitoring module can also record the task completion time in real time for the master control module to adjust the subsequent task allocation plan.

[0067] The main functions of the task monitoring module include status collection, anomaly monitoring, and task feedback.

[0068] In this embodiment, the task monitoring module first receives status data from the handling robot, and these data include the task execution status S i , the path execution progress P prog,i , and the current position x i , y i . Generally, the task execution status S i can be represented by the following coding method: S i = 0: The task has not started; S i = 1: The task is in progress; S i = 2: The task has been completed; S i = -1: The task is abnormally interrupted.

[0069] The task monitoring module generates a task execution progress report based on the status information provided by the handling robot and sends the report to the general control module. The progress report includes the completion status of each task point and the estimated completion time T est,i 。

[0070] The task monitoring module can dynamically predict the task completion time by combining historical task data and current path information. In terms of anomaly monitoring, the task monitoring module can detect anomalies in task execution in real time. For example, when the handling robot stalls on the path for a long time, the task monitoring module will send an alarm message to the general control module so that the general control module can reassign task points or trigger global path replanning.

[0071] In some embodiments, the task monitoring module can also optimize the task completion status by combining multi-task scheduling strategies. For example: When a certain handling robot causes a task delay due to a blocked path, the task monitoring module can transfer the task to other handling robots for execution; When high-priority tasks require additional resources, the task monitoring module can suspend the execution of low-priority tasks to ensure the timely completion of high-priority tasks.

[0072] In another possible implementation, the task monitoring module can store real-time monitoring data in a central database for subsequent task analysis and optimization. For example: Analyze the task completion time distribution and adjust the scheduling parameters μ1, μ2; Combine historical path data to optimize the parameters λ1, λ2, λ3 in the path weight calculation formula.

[0073] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A logistics warehousing system, characterized in that, Including: An environment perception module, which is used to monitor in real time task points, path status, obstacle positions and dynamic order changes in the warehousing environment, and generate path weight data; A total control module, connected to the environment perception module, which is used to receive order requirements, generate a set of task points and obtain path weight data, and globally schedule the handling robots according to the path weight data; A path optimization module, connected to the total control module, which is used to calculate the optimal path of the handling robots based on a multi-agent game model; Handling robots, connected to the path optimization module and the total control module, which are used to receive the optimal path instructions generated by the path optimization module and execute tasks, and at the same time feedback the execution status to the total control module; A task monitoring module, respectively connected to the total control module and the handling robots, which is used to monitor the status of the handling robots and the task progress, and feedback the dynamic change information to the total control module; the path optimization module optimizes the paths of the handling robots according to the total path distance, path time cost and path conflict cost, and the total control module generates task assignments for the handling robots based on the optimization results.

2. The logistics warehousing system according to claim 1, characterized in that, The environment perception module models the warehousing environment as a directed weighted graph, including a set of task points and a set of paths, and sends the path weight data to the total control module and the path optimization module.

3. A logistics warehousing system according to claim 1, characterized in that, The path weight is calculated based on the path length, time cost and path conflict cost. The path conflict cost is determined by the ratio of the current occupied state of the path to the passable capacity of the path. The calculation results are updated in real time by the environment perception module and sent to the total control module and the path optimization module.

4. A logistics warehousing system according to claim 1, characterized in that, The path optimization module performs path planning based on a multi-agent non-cooperative game model. The total control module regards the handling robots as independent agents and sends the set of task points to the path optimization module. The path optimization module calculates the optimal path according to the utility function of the handling robots.

5. A logistics warehousing system according to claim 4, characterized in that, The path optimization module is connected to the total control module, which is used to calculate the optimal path of the handling robots based on the path weight data and the utility function, generate a path selection result through iterative optimization, and send the optimal path to the total control module.

6. A logistics warehousing system according to claim 5, characterized in that, The path optimization module is connected to the total control module. The path optimization module optimizes the path selection of the handling robots by the gradient projection method, where: Receive the set of task points and path weight data sent by the total control module; Based on the set of task points and path weight data, calculate the gradient of the path utility function of each handling robot; Adjust the path selection of each handling robot, and map the adjustment result to the set of feasible paths through projection operation; After the change amount of the utility function is less than the set convergence threshold, the path optimization module sends the optimal path result of the handling robots to the total control module.

7. A logistics warehousing system according to claim 1, characterized in that, The handling robots are connected to the environment perception module through an obstacle perception module, and can detect the path blocked state in real time and trigger local path replanning. The local path replanning is recalculated by the path optimization module based on the path weight.

8. A logistics warehousing system according to claim 7, characterized in that, During the local path replanning process, the handling robots preferentially select the path with the lowest weight for replacement, and feedback the path adjustment status to the task monitoring module and the total control module through the obstacle perception module.

9. A logistics warehousing system according to claim 1, characterized in that, The overall control module is connected to the environment perception module and the path optimization module. When the environment perception module detects a major change in the environment, the overall control module triggers a global path replanning, which includes: The environment perception module regenerates the warehousing environment model, including updating the set of task points and the set of paths, and recalculating the path weight data; The overall control module receives the updated path weight data and generates a new set of task points and task priorities; Based on the updated set of task points and path weight data, the path selection of all handling robots is re-optimized; The optimized optimal path is sent to the overall control module, and the overall control module sends the new task assignment result to the corresponding handling robot.

10. A logistics warehousing system according to claim 1, characterized in that, The task monitoring module is bidirectionally connected to the overall control module. After the handling robot completes the task, it feeds back the task status through the task monitoring module. The overall control module generates a new set of task points according to the task completion status and triggers the path optimization module to perform the next round of optimization.

Images