AGV (Automatic Guided Vehicle) taking path optimization method for intelligent warehousing

Abstract

The invention relates to the technical field of AGV trolley material taking path optimization, in particular to an AGV trolley material taking path optimization method for intelligent warehousing. According to the method, an AGV which executes a new task is used as a target vehicle, an AGV which does not update a path is used as a reference vehicle, and an initial driving path of the target vehicle is determined based on a running path of the reference vehicle; when the target vehicle runs to a termination node of the initial driving road section according to the initial driving path, the analysis data of the subsequent candidate road sections are predicted again; obtaining a second driving road section of the target vehicle according to the difference between the initial prediction analysis data and the actual analysis data of the initial driving road section, the difference between the second prediction analysis data and the initial prediction analysis data of the subsequent candidate road sections and the number of alternative paths of the subsequent candidate road sections; and the driving road sections are sequentially acquired until the target vehicle arrives at the task end point. The running state of the target vehicle is analyzed based on the reference vehicle, and the accuracy and stability of target vehicle path setting are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of AGV (Automated Guided Vehicle) material handling path optimization technology, and specifically to an AGV material handling path optimization method for intelligent warehousing. Background Technology

[0002] Smart warehousing, as a core component of modern logistics and manufacturing upgrades, integrates cutting-edge technologies such as automation, the Internet of Things, big data analytics, and artificial intelligence. It has achieved unmanned or semi-unmanned operations in core processes such as goods storage, order picking, material handling, and sorting. Among these, AGVs (Automated Guided Vehicles), as the mobile execution units of smart warehousing, undertake the crucial tasks of precise material transfer and retrieval. Based on order demand and inventory information, the system can automatically allocate retrieval tasks to AGVs and guide them from the starting point to the destination through path planning, significantly improving the efficiency and accuracy of warehousing operations.

[0003] The operating environment of intelligent warehousing is not static. AGVs may experience sudden malfunctions, encounter temporary obstacles, or have tasks inserted. These dynamic disturbances require AGVs to have the ability to adjust their material picking paths in real time. However, existing technologies only make local adjustments through simple obstacle avoidance algorithms and cannot reconstruct the path from a global task perspective, which may reduce the overall material picking efficiency. In addition, existing AGV path optimization algorithms mostly focus on the efficiency of a single AGV, only paying attention to the shortest path length or the least travel time for a single AGV, ignoring the need for coordinated operation of multiple AGVs in the warehousing system. Since AGVs share the warehousing road network, when an AGV adjusts its path due to dynamic disturbances, if the operating status of other AGVs is not considered, it is very easy to cause path conflicts, which may even cause the system to lock up in severe cases, increasing the total energy consumption and total time of the system and exacerbating task delays. Summary of the Invention

[0004] To address the technical problem that existing AGV path optimization algorithms fail to consider the operating status of other AGVs, leading to path conflicts and exacerbating task delays, the present invention aims to provide an AGV trolley material retrieval path optimization method for intelligent warehousing. The specific technical solution adopted is as follows: This invention provides a method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing, the method comprising the following steps: The AGVs that need to perform new tasks are taken as target vehicles, and the candidate road segments and the initial predicted arrival times of the corresponding nodes of each target vehicle are obtained; the AGVs that do not need to update their paths are taken as reference vehicles, and the running paths of each reference vehicle are obtained. Based on the reference vehicle's running path, the initial predicted reference vehicle flow for each candidate road segment is predicted, and the initial driving path for each target vehicle is determined. When the target vehicle travels to the end node of the initial travel segment according to the initial travel path, the second predicted reference traffic flow of the subsequent candidate road segments and the second predicted arrival time of its nodes are re-predicted; the arrival time and reference traffic flow are used as analysis data; based on the difference between the initial predicted analysis data and the actual analysis data corresponding to the initial travel segment, the difference between the second predicted analysis data and the initial predicted analysis data corresponding to the subsequent candidate road segments, and the number of alternative paths for the subsequent candidate road segments, the second travel segment of the target vehicle is obtained; The target vehicle travels along the second driving segment, and then obtains subsequent driving segments in sequence until the target vehicle reaches the mission endpoint.

[0005] Furthermore, the method for obtaining the initial driving path is as follows: For any target vehicle, the initial driving path of the target vehicle is obtained by using the dynamic A* algorithm based on the initial predicted reference traffic flow of each candidate road segment and the initial predicted arrival time of the corresponding node of the candidate road segment.

[0006] Furthermore, the method for obtaining the second travel segment is as follows: For any target vehicle, the termination node of the initial driving segment of the target vehicle is taken as the first analysis node. Based on the difference between the initial predicted analysis data and the actual analysis data corresponding to the initial driving segment of the target vehicle, and the difference between the second predicted analysis data and the initial predicted analysis data corresponding to the subsequent candidate segments of the first analysis node, the first predicted influence degree of each candidate segment of the target vehicle after the first analysis node is obtained. Starting from the first analysis node, alternative paths to the target vehicle's mission endpoint are obtained, and all of them are used as the first analysis path; For any candidate road segment of the target vehicle following the first analysis node, the first analysis path containing the candidate road segment will be used as the designated path for that candidate road segment. For any given path, the sum of the first predicted impact of the candidate road segment and its subsequent candidate road segments on the given path is taken as the overall impact of the candidate road segment on the given path. Based on the overall impact of the candidate road segment on each specified path, the replacement cost of the candidate road segment is obtained; The correction cost of the candidate road segment is obtained based on the replacement cost of the candidate road segment and the number of first analysis paths. The product of the correction cost and the first predicted impact level of the candidate road segment is taken as the comprehensive impact level of the candidate road segment; Based on the dual objectives of minimizing the overall impact and minimizing the path length, the second driving path of the target vehicle is obtained through the dynamic A* algorithm, and the second driving segment of the target vehicle is determined.

[0007] Furthermore, the method for obtaining the first predicted degree of impact is as follows: The difference between the initial predicted reference traffic flow and the actual reference traffic flow of the initial travel segment of the target vehicle is taken as the first difference; The difference between the initial predicted arrival time and the actual arrival time of the target vehicle at the initial end point of its initial travel segment is taken as the second difference; For any candidate road segment of the target vehicle after the first analysis node, the difference between the initial predicted reference traffic flow and the second predicted reference traffic flow of the candidate road segment is taken as the third difference; The difference between the initial predicted arrival time and the second predicted arrival time of the candidate road segment termination node is taken as the fourth difference; The product of the first preset weight and the first difference is used as the first feature value; the product of the second preset weight and the second difference is used as the second feature value; the product of the third preset weight and the third difference is used as the third feature value; the product of the fourth preset weight and the fourth difference is used as the fourth feature value; wherein, the sum of the first preset weight, the second preset weight, the third preset weight and the fourth preset weight is 1; The sum of the first, second, third, and fourth eigenvalues, followed by normalization, is used as the first predicted influence level of the candidate road segment.

[0008] Furthermore, the method for obtaining the replacement cost is as follows: When there are at least two specified paths for a candidate road segment, the overall impact of the candidate road segment is arranged in ascending order to obtain a sequence of overall impact. If the first overall impact degree in the overall impact degree sequence is not equal to the second overall impact degree, then the difference between the second overall impact degree and the first overall impact degree shall be taken as the first value; The ratio of the first value to the second overall impact level is used as the replacement cost of the candidate road segment; If the first overall impact level in the overall impact level sequence is equal to the second overall impact level, then the default value is 0 as the replacement cost of the candidate road segment; When there is only one specified path for the candidate road segment, the default value is 1, which is the replacement cost of the candidate road segment.

[0009] Furthermore, the method for obtaining the correction cost is as follows: The ratio of the replacement cost of the candidate road segment to the number of paths analyzed in the first analysis is used as the correction cost of the candidate road segment.

[0010] Furthermore, the alternative path refers to all feasible paths that start from the end node of the initial driving segment, reach the task endpoint of the corresponding target vehicle, and can cover the corresponding candidate segment.

[0011] Furthermore, the method for obtaining the predicted arrival time is as follows: For any target vehicle, the shortest path from the current node to the termination node of each subsequent candidate road segment is obtained through the dynamic A* algorithm, and used as the reference path for the corresponding node. The product of the length of the reference path and the average speed of the target vehicle is used as the predicted arrival time of the corresponding node.

[0012] Furthermore, the method for obtaining the initial predicted reference traffic flow is as follows: For any candidate road segment of any target vehicle, the time period consisting of the initial predicted arrival times of the start and end nodes of the candidate road segment is taken as the initial predicted travel time period of the candidate road segment. The number of all reference vehicles located on the candidate road segment is counted and predicted within the initial predicted driving time period, and this number is used as the initial predicted reference traffic flow for the target vehicle on that candidate road segment.

[0013] Furthermore, the optimization of the material retrieval path for the AGV in the intelligent warehousing establishes a data interaction link with the enterprise resource planning system, warehouse management system, and warehouse control system through the AGV scheduling system. This allows the system to obtain the task information of the target vehicle, the warehouse road network information, and the real-time operating status of the reference vehicle. Based on the task information and the warehouse road network information, candidate road segments for the target vehicle are determined. The operating path of the reference vehicle is obtained through historical planning data and real-time positioning data from the warehouse control system.

[0014] The present invention has the following beneficial effects: This invention first clearly divides AGVs into target vehicles that need to perform new tasks and reference vehicles that do not require path updates. Candidate route selection and arrival time prediction calculations are performed only for the target vehicles, eliminating the need for full path reconstruction for all AGVs. This reduces system data processing volume and computational pressure, avoiding computational overload caused by simultaneous planning of a large number of AGVs. Simultaneously, the target vehicle can quickly focus on new task requirements, accurately acquire candidate routes and initial predicted arrival times, while the reference vehicles maintain stable travel on their original routes. This significantly shortens the response time for path adjustments after temporary task allocation, ensuring that warehousing operations are not disrupted by large-scale path changes. Furthermore, based on the actual operating paths of the reference vehicles, the initial predicted reference vehicle traffic flow for each candidate route of the target vehicle is predicted. This incorporates the spatiotemporal interaction risk of multiple AGVs into the core consideration of initial path planning, prioritizing the selection of reference vehicle traffic flow. By selecting candidate road segments with low traffic volume, the probability of the target vehicle and the reference vehicle meeting on the same road segment and at the same time is reduced from the planning source. This avoids the problem of frequent conflicts among multiple AGVs due to the single AGV path being optimal in traditional algorithms. It effectively reduces AGV congestion, waiting at intersections, and other situations, improving the traffic efficiency and AGV operation stability of the entire warehouse road network. When the target vehicle reaches the end node of the initial driving segment, the second predicted reference vehicle traffic and arrival time of the subsequent candidate road segments are re-predicted. The initial prediction is compared with the actual data. The differences between the initial prediction and the second prediction are captured in real time to capture dynamic changes in road conditions. The path planning strategy is adjusted based on the difference data, so that the second driving segment of the target vehicle can accurately adapt to the current actual scenario. This avoids path failure caused by the disconnect between the initial planning and real-time road conditions, and significantly improves the path adaptability and fault tolerance of AGVs in dynamic interference environments. When acquiring the second driving segment, the number of alternative routes for subsequent candidate segments is used as a key decision-making factor. Segments with sufficient alternative routes are prioritized, allowing the target vehicle to quickly switch to an alternative route without prolonged stagnation or the need to replan a long-distance alternative route, significantly reducing the risk of route interruption. Simultaneously, the introduction of multiple alternative routes provides more options for route optimization, ensuring that while maintaining low interference and high efficiency, the reliability and resilience to unforeseen circumstances of route planning are further enhanced. After the target vehicle travels along the second driving segment, it continuously repeats the process of reaching the node, re-predicting, analyzing differences, and selecting a new segment until it reaches the task endpoint, forming a closed loop of dynamic path optimization throughout the entire task cycle. This allows for real-time correction of path deviations and avoids continuous travel to congested sections. At the same time, the step-by-step planning can flexibly adapt to new changes based on the progress of the task, ensuring that the target vehicle always advances the task along the optimal path, guaranteeing that the material collection task is completed efficiently and on time. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A schematic flowchart illustrating an AGV (Automated Guided Vehicle) material retrieval path optimization method for intelligent warehousing, provided as an embodiment of the present invention; Figure 2 This is a structural diagram of an AGV (Automated Guided Vehicle) material handling path optimization system for intelligent warehousing, provided in one embodiment of the present invention. Figure 3 This is a schematic diagram of a computer device provided according to an embodiment of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the AGV (Automated Guided Vehicle) material handling path optimization method for intelligent warehousing proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0019] The following describes in detail, with reference to the accompanying drawings, the specific scheme of the AGV (Automated Guided Vehicle) material picking path optimization method for intelligent warehousing provided by the present invention.

[0020] Example 1: This invention proposes a method for optimizing the material handling path of AGVs used in intelligent warehousing. Please refer to [link / reference]. Figure 1 The diagram illustrates a schematic flowchart of an AGV (Automated Guided Vehicle) material retrieval path optimization method for intelligent warehousing, according to an embodiment of the present invention. The method includes the following steps: Step S1: Take the AGV that needs to perform the new task as the target vehicle, and obtain the candidate road segment and the initial predicted arrival time of the corresponding node of each target vehicle; take the AGV that does not need to update its path as the reference vehicle, and obtain the running path of each reference vehicle.

[0021] Specifically, in the AGV (Automated Guided Vehicle) material handling path optimization system used for intelligent warehousing, the AGV scheduling system interacts with the enterprise resource planning system, warehouse management system, and warehouse control system to build a complete information scheduling platform. This platform acquires comprehensive data, including inventory information, order data, material information, AGV running paths, and real-time load status, and makes AGV task distribution decisions based on this information.

[0022] When a new task is assigned to an AGV, the AGV needs to temporarily take on this new task in addition to performing its existing tasks. The system first determines the start and end nodes of the new task and adds them to the AGV's path planning as the path nodes that the AGV must traverse. This process is achieved by updating the path graph. The path graph is represented using a graph theory model. The nodes in the path graph correspond to nodes in the warehouse environment, such as shelf points, intersections, and workstations. The edges in the path graph represent road segments that the AGV can travel on. It should be noted that in the path graph, each road segment corresponds to two nodes: a start node and an end node. Adjacent road segments are directly connected by nodes; that is, the end node of one road segment is the start node of its next adjacent road segment. After updating the path graph, the system uses the dynamic A* algorithm to calculate the AGV's material retrieval path (a complete path from the start and end nodes of the task, composed of continuous road segments). However, the obtained material retrieval path has limitations; it only optimizes the path from the local perspective of the AGV, ignoring the overall collaborative work of multiple AGVs in the system. Because all AGVs share limited path resources, a change in the path of one AGV can trigger a chain reaction, affecting the path decisions of other AGVs. This can lead to local or even global path congestion, and in severe cases, even system lock-up, rendering all AGVs unable to operate normally. The dynamic A* algorithm is well-known and will not be elaborated upon further.

[0023] Existing methods employ a reactive obstacle avoidance strategy, which temporarily activates the obstacle avoidance program to replan the path when the AGV detects an obstacle in its path. While this method can resolve instantaneous conflicts, it has the following drawbacks: First, the system needs to continuously monitor the real-time status of all vehicles and perform frequent path replanning, resulting in enormous computational resource pressure. Second, due to the lack of forward-looking planning, the vehicle's material picking path is in a constantly adjusting and unstable state, frequently deviating from the original route to avoid conflicts, which may reduce the overall material picking efficiency and increase the uncertainty of the system.

[0024] To better optimize the material handling paths of AGVs, this embodiment divides AGVs into two types when temporary task assignments occur: AGVs requiring new tasks are designated as target vehicles, while those not requiring path updates are designated as reference vehicles. Target vehicles are used only for handling urgent orders or unexpected tasks and constitute a minority in the system; reference vehicles maintain their material handling paths and constitute a large proportion of the system. By maintaining the stability of the reference vehicle paths, a predictable operating environment can be maintained. The motion status of the reference vehicles changes in real time and can be acquired and directly invoked, providing a foundation for accurately predicting the overall traffic flow of the system. If frequent path replanning is performed on reference vehicles, it will not only significantly increase the real-time computing load of the system but also require the acquisition and processing of a large amount of dynamic data, easily leading to system resource shortages and affecting overall operating efficiency. Therefore, keeping the original operating paths of the reference vehicles unchanged and only dynamically optimizing the paths of a few target vehicles ensures that urgent tasks can be handled in a timely manner while maintaining the overall stability and predictability of the system, providing a reliable data foundation for subsequent path conflict prediction and optimization.

[0025] To accurately plan the target vehicle's travel path, this embodiment obtains candidate road segments and their corresponding nodes' initial predicted arrival times using the target vehicle's task information, warehouse road network information, and the real-time operating status of a reference vehicle. Simultaneously, it obtains the reference vehicle's travel path using historical planning data and real-time positioning data from the warehouse control system. Specifically, the predicted arrival time is obtained as follows: for any target vehicle, the shortest path from the current node to the termination nodes of each subsequent candidate road segment is obtained using a dynamic A* algorithm, serving as the reference path for that node. Then, the product of the length of each reference path and the target vehicle's average speed is used as the predicted arrival time for that node. It should be noted that candidate road segments are the road segments included in the path that the target vehicle can travel from the task start point to the task end point.

[0026] Step S2: Based on the reference vehicle's running path, predict the initial predicted reference vehicle flow for each candidate road segment, and determine the initial driving path for each target vehicle.

[0027] Specifically, in the AGV scheduling system, the reference vehicles maintain a fixed operating path, forming a highly predictable motion system. Based on their preset material picking paths, driving speeds, and standard travel times between nodes, the system can accurately calculate the path location information of each reference vehicle at any future time. Therefore, for any given future time period, by analyzing the preset positions of all reference vehicles within that time period, the system can predict the reference vehicle traffic flow of each road segment during that time period. This provides crucial prior knowledge for the entire path optimization system, enabling the system to proactively identify potential congestion areas and provide a reliable decision-making basis for the dynamic path planning of target vehicles. Thus, in this embodiment, the initial predicted reference vehicle traffic flow of each candidate road segment is predicted based on the operating path of the reference vehicles, thereby determining the initial driving path of each target vehicle.

[0028] Preferably, in one feasible embodiment of this method, the initial predicted reference traffic flow is obtained as follows: for any candidate road segment of any target vehicle, the time period consisting of the initial predicted arrival times of the start and end nodes of the candidate road segment is taken as the initial predicted travel time period of the candidate road segment; within the initial predicted travel time period, the number of all reference vehicles predicted to be located on the candidate road segment is counted, and this number is taken as the initial predicted reference traffic flow of the target vehicle on that candidate road segment. It should be noted that the initial predicted arrival time of the start node of the candidate road segment corresponding to the target vehicle's task starting point is known and can be used directly.

[0029] Preferably, in one feasible implementation of this embodiment, for any target vehicle, based on the initial predicted reference traffic flow of each candidate road segment and the initial predicted arrival time of the corresponding node of the candidate road segment, the initial travel path of the target vehicle is obtained through a dynamic A* algorithm. It should be noted that the traditional dynamic A* algorithm uses road segment length or travel time as a single cost. This embodiment optimizes the cost function of the dynamic A* algorithm to a weighted form of the initial predicted reference traffic flow cost and the path length cost, specifically as follows: ;in, The total cost of the i-th candidate road segment; The normalized result of the initial predicted reference traffic flow for the i-th candidate road segment; This is the normalized result of the path length of the i-th candidate road segment. In this embodiment, the initial predicted reference traffic flow and the path length are normalized using a linear normalization method. The linear normalization method is a well-known technique and will not be described in detail here. The smaller the value, the better the i-th candidate road segment; this embodiment sets... It is 0.6. The value is 0.4, and implementers can set it according to the actual situation. and The size is not limited here, but + It equals 1, set here. and In order to determine and In acquiring The reference weight at the time. Taking the task start node of the target vehicle as the starting point and the task end node as the ending point, based on the above cost function, the dynamic A* algorithm is used to traverse and search all candidate road segments. The specific process is as follows: (1) Starting from the starting point, calculate the total cost of all adjacent candidate road segments; (2) Prioritize the road segment with the smallest total cost to enter the next node, and record the traversed path at the same time; Repeat steps (1) and (2) until the end node is searched to form a complete path sequence; If there are multiple paths with the same total cost, the shortest total path length is used as an auxiliary screening condition to determine the unique initial driving path of the target vehicle.

[0030] At this point, the initial driving path of each target vehicle is obtained.

[0031] Step S3: When the target vehicle travels to the end node of the initial travel segment according to the initial travel path, re-predict the second predicted reference traffic flow of the subsequent candidate road segments and the second predicted arrival time of its nodes; use the arrival time and reference traffic flow as analysis data; based on the difference between the initial predicted analysis data and the actual analysis data corresponding to the initial travel segment, the difference between the second predicted analysis data and the initial predicted analysis data corresponding to the subsequent candidate road segments, and the number of alternative paths for the subsequent candidate road segments, obtain the second travel segment of the target vehicle.

[0032] Specifically, as the target vehicle travels along the initial segment of its initial route, it provides real-time feedback on its location. When the system detects that the target vehicle has reached the end node of the initial segment, it automatically initiates the planning process for the second segment. This is because the initial segment is based solely on predicted data from the initial location. During the target vehicle's journey, minor changes may occur in the storage environment. For example, a path deviation caused by mechanical errors in the reference vehicle, or the presence of small, temporarily scattered obstacles that, while not triggering emergency obstacle avoidance, still affect traffic efficiency. Furthermore, the target vehicle's actual driving status (such as speed fluctuations or time differences due to load variations) may deviate from the initial prediction. Therefore, replanning is necessary at the end node of the initial segment to prevent accumulated deviations from causing subsequent path conflicts or reduced efficiency. Subsequently, after the target vehicle reaches the end node of its initial segment, based on the current operating status of the reference vehicle, the system re-predicts the second predicted reference vehicle traffic flow and the second predicted arrival time of its nodes for subsequent candidate segments.

[0033] To more accurately plan the second travel segment of the target vehicle, this embodiment first uses arrival time and reference traffic flow as analysis data. Then, it analyzes the difference between the initial predicted analysis data and the actual analysis data corresponding to the initial travel segment, reflecting the delay of the target vehicle and the error in the initial prediction. Further analysis examines the difference between the second predicted analysis data and the initial predicted analysis data corresponding to subsequent candidate travel segments, reflecting the impact of changes in the target vehicle's travel progress on subsequent time planning and the impact of changes in the reference vehicle's operating status on the congestion risk of subsequent road segments. Finally, this embodiment combines the differences between the initial predicted analysis data and the actual analysis data corresponding to the initial travel segment, and the differences between the second predicted analysis data and the initial predicted analysis data corresponding to subsequent candidate travel segments, to quantify the interference intensity of the target vehicle's selection of each candidate road segment on the normal operation of the reference vehicle. The lower the interference intensity, the more likely the corresponding candidate road segment is to be used as the target vehicle's travel segment. The initial predicted analysis data includes the initial predicted reference traffic flow and the initial predicted arrival time; the actual analysis data includes the actual reference traffic flow and the actual arrival time; the second predicted analysis data includes the second predicted reference traffic flow and the second predicted arrival time.

[0034] Furthermore, considering that a larger number of alternative routes for a certain road segment indicates that even if congestion or obstacles occur on that segment, the target vehicle can quickly switch to other routes, resulting in lower route adjustment costs; that is, the more alternative routes there are, the stronger the fault tolerance of that road segment, indirectly indicating that the correction cost of that road segment is smaller, and the more suitable that road segment is as the target vehicle's driving segment. It should be noted that the number of alternative routes corresponding to a candidate road segment after the initial driving segment's termination node refers to the total number of all feasible paths that start from the initial driving segment's termination node, reach the corresponding target vehicle's mission endpoint, and cover that candidate road segment.

[0035] Therefore, in this embodiment, the second driving segment of the target vehicle is obtained based on the difference between the initial prediction analysis data and the actual analysis data corresponding to the initial driving segment, the difference between the second prediction analysis data and the initial prediction analysis data corresponding to the subsequent candidate segments, and the number of alternative paths for the subsequent candidate segments.

[0036] Preferably, in one feasible embodiment, the method for obtaining the second driving segment is as follows: For any target vehicle, the termination node of the initial driving segment of the target vehicle is taken as the first analysis node. Based on the difference between the initial predicted analysis data and the actual analysis data corresponding to the initial driving segment of the target vehicle, and the difference between the second predicted analysis data and the initial predicted analysis data corresponding to the subsequent candidate segments of the first analysis node, the first predicted influence degree of each candidate segment of the target vehicle after the first analysis node is obtained; the greater the first predicted influence degree, the more the reference vehicle operation of the corresponding candidate segment is affected by the target vehicle. The method for obtaining the first predicted impact level is as follows: the absolute value of the difference between the initial predicted reference traffic flow and the actual reference traffic flow of the target vehicle's initial travel segment is taken as the first difference; the absolute value of the difference between the initial predicted arrival time and the actual arrival time of the target vehicle's initial travel segment termination node is taken as the second difference; the larger the first and second differences are, the greater the impact of the target vehicle's actual travel on the reference vehicle, and the less accurate the current prediction is; further, for any candidate road segment of the target vehicle after the first analysis node, the absolute value of the difference between the initial predicted reference traffic flow and the second predicted reference traffic flow of the candidate road segment is taken as the third difference; the larger the third difference is, the greater the impact of the change in the reference vehicle's operating status on the subsequent road segment's congestion risk may be; the absolute value of the difference between the initial predicted arrival time and the second predicted arrival time of the candidate road segment termination node is taken as the fourth difference; the larger the fourth difference is, the greater the impact of the target vehicle's travel progress change on the subsequent time planning may be. To accurately analyze the interference of the target vehicle on the normal operation of the reference vehicle when driving on the candidate road segment, the product of the first preset weight and the first difference is used as the first feature value; the product of the second preset weight and the second difference is used as the second feature value; the product of the third preset weight and the third difference is used as the third feature value; and the product of the fourth preset weight and the fourth difference is used as the fourth feature value. The sum of the first, second, third, and fourth preset weights is 1. In this embodiment, the first preset weight is set to 0.2, the second preset weight to 0.2, the third preset weight to 0.3, and the fourth preset weight to 0.3. The implementer can set the values ​​of the first, second, third, and fourth preset weights according to the actual situation; this is not limited here. The first, second, third, and fourth preset weights are set here to determine the participation weights of the first, second, third, and fourth differences in sequence. Finally, the sum of the first, second, third, and fourth eigenvalues ​​and the result of linear normalization are used as the first predictive influence of the candidate road segment. The greater the first predictive influence, the more unstable the prediction result of the candidate road segment, which indirectly indicates that the target vehicle's driving on the candidate road segment has a greater interference with the reference vehicle. Starting from the first analysis node, alternative paths to the target vehicle's task endpoint are obtained and used as the first analysis path. For any candidate road segment of the target vehicle after the first analysis node, the first analysis path containing the candidate road segment is used as the designated path for that candidate road segment. For any designated path, the sum of the first predicted influence levels of the candidate road segment and its subsequent candidate road segments on that designated path is used as the overall influence level of the candidate road segment on that designated path. The smaller the overall influence level, the lower the interference risk of the candidate road segment to the reference vehicle, and the more suitable it is as a candidate road segment to ensure the overall stability of the AGV's operation. To accurately analyze the target vehicle's driving situation on the candidate road segment, the replacement cost of the candidate road segment is obtained based on the overall influence level of the candidate road segment on each designated path. The larger the replacement cost, the smaller the degree of adjustment the target vehicle can make when driving the candidate road segment, i.e., the worse the fault tolerance, and the less suitable the candidate road segment is for the target vehicle to drive. The replacement cost is obtained as follows: When there are at least two specified paths for a candidate road segment, the overall impact of the candidate road segment is arranged in ascending order to obtain an overall impact sequence. If the first overall impact in the overall impact sequence is not equal to the second overall impact, the difference between the second and first overall impact is taken as the first value. The larger the first value, the smaller the first overall impact, the better the driving effect of the specified path corresponding to the first overall impact, and the greater the cost of subsequent replacement. The ratio of the first value to the second overall impact is then taken as the replacement cost of the candidate road segment. If the first overall impact in the overall impact sequence is equal to the second overall impact, the default value is 0 for the replacement cost of the candidate road segment. When there is only one specified path for the candidate road segment, the default value is 1 for the replacement cost of the candidate road segment. The more first analysis paths a candidate road segment has, the stronger its fault tolerance. To more accurately analyze the fault tolerance of a candidate road segment, the correction cost of the candidate road segment is obtained based on its replacement cost and the number of first analysis paths. The smaller the correction cost, the better the candidate road segment and the stronger its fault tolerance. The correction cost is obtained by taking the ratio of the replacement cost of the candidate road segment to the number of first analysis paths as the correction cost of the candidate road segment. The smaller the predicted impact, the less direct interference the target vehicle will cause to the reference vehicle when traveling on the candidate road segment; the smaller the correction cost, the stronger the fault tolerance of the candidate road segment. The product of the correction cost and the predicted impact of the candidate road segment is then used as the comprehensive impact of the candidate road segment. This leads to the acquisition of the comprehensive impact of each candidate road segment for the target vehicle after the first analysis node. Based on the dual objectives of minimizing the comprehensive impact and shortest path length, the second driving path of the target vehicle is obtained using the dynamic A* algorithm, thus determining the second driving segment of the target vehicle. The priority rule for the dual objectives is as follows: first, all paths with a comprehensive impact greater than a preset impact threshold are excluded, forming a candidate path set; if the candidate path set is empty, the preset impact threshold needs to be increased (e.g., adjusted from 0.5 to 0.6) and recalculated to ensure at least one feasible path; the preset impact threshold can be set according to actual conditions and is not limited here; then, in the candidate path set, the path with the shortest total path length is selected. If two or more paths have the same length, the path with the smallest comprehensive impact is further selected to ensure a unique optimal solution. It should be noted that the second driving path of the target vehicle is obtained based on the dual objectives of minimizing the overall impact and the shortest path length, using a dynamic A* algorithm, and is independent of the cost function in step S2.

[0037] Step S4: The target vehicle travels along the second driving segment, and then obtains the subsequent driving segments in sequence until the target vehicle reaches the mission endpoint.

[0038] Specifically, when the system detects that the target vehicle has reached the end node of the second driving segment, it automatically starts the planning process for the third driving segment according to the method in step S3. The specific process is as follows: For any target vehicle, after it reaches the termination node of its initial second travel segment, based on the current operating status of the reference vehicle, the third predicted reference traffic flow and the third predicted arrival time of its node for subsequent candidate road segments are re-predicted. Simultaneously, the termination node of the target vehicle's second travel segment is designated as the second analysis node; the absolute value of the difference between the second predicted reference traffic flow and the actual reference traffic flow of the target vehicle's second travel segment is designated as the fifth difference; the absolute value of the difference between the second predicted arrival time and the actual arrival time of the target vehicle's second travel segment termination node is designated as the sixth difference. For any candidate road segment of the target vehicle following the second analysis node, the absolute value of the difference between the second and third predicted reference traffic flows of the candidate road segment is designated as the seventh difference; the absolute value of the difference between the second and third predicted arrival times of the candidate road segment termination node is designated as the seventh difference. The absolute value of the time difference is taken as the eighth difference. To accurately analyze the interference of the target vehicle on the normal operation of the reference vehicle when the candidate road segment is in motion, the product of the fifth preset weight and the fifth difference is taken as the fifth feature value; the product of the sixth preset weight and the sixth difference is taken as the sixth feature value; the product of the seventh preset weight and the seventh difference is taken as the seventh feature value; and the product of the eighth preset weight and the eighth difference is taken as the eighth feature value. The sum of the fifth, sixth, seventh, and eighth preset weights is 1. In this embodiment, the fifth preset weight is set to 0.2, the sixth preset weight to 0.2, the seventh preset weight to 0.3, and the eighth preset weight to 0.3. The implementer can set the values ​​of the fifth, sixth, seventh, and eighth preset weights according to the actual situation, which is not limited here. Finally, the sum of the fifth, sixth, seventh, and eighth feature values ​​and the result of linear normalization are taken as the second predicted influence degree of the candidate road segment.

[0039] Starting from the second analysis node, alternative paths to the target vehicle's task endpoint are obtained and all are considered as second analysis paths. For any candidate road segment of the target vehicle after the second analysis node, the second analysis path containing the candidate road segment is considered as the second designated path for that candidate road segment. For any second designated path, the sum of the second predicted influence levels of the candidate road segment and its subsequent candidate road segments on the second designated path is taken as the second overall influence level of the candidate road segment on the second designated path. To accurately analyze the target vehicle's driving choice for the candidate road segment, the second replacement cost of the candidate road segment is further obtained based on the overall influence level of the candidate road segment on each second designated path. The method for obtaining the second replacement cost is as follows: When there are at least two second designated paths for the candidate road segment, the second overall influence degree corresponding to the candidate road segment is arranged in ascending order to obtain a second overall influence degree sequence; if the first second overall influence degree in the second overall influence degree sequence is not equal to the second second overall influence degree, the difference between the second and first second overall influence degrees is taken as the second value; the ratio of the second value to the second second overall influence degree is taken as the second replacement cost of the candidate road segment; if the first second overall influence degree in the second overall influence degree sequence is equal to the second second overall influence degree, the default value is 0 as the second replacement cost of the candidate road segment; when there is only one second designated path for the candidate road segment, the default value is 1 as the second replacement cost of the candidate road segment. The more second analysis paths a candidate road segment has, the stronger its fault tolerance. To more accurately analyze the fault tolerance of a candidate road segment, the ratio of the second replacement cost of the candidate road segment to the number of second analysis paths is used as the second correction cost of the candidate road segment. The smaller the second correction cost, the better the candidate road segment. The product of the second correction cost and the second predicted impact of the candidate road segment is taken as the second comprehensive impact of the candidate road segment. Then, the second comprehensive impact of each candidate road segment for the target vehicle after the second analysis node is obtained. Based on the dual objectives of minimizing the second comprehensive impact and minimizing the path length, the third driving path of the target vehicle is obtained through the dynamic A* algorithm, thereby determining the third driving segment of the target vehicle. It should be noted that the third driving path of the target vehicle is obtained solely based on the dual objectives of minimizing the second comprehensive impact and minimizing the path length, using the dynamic A* algorithm.

[0040] This process continues, obtaining the subsequent travel segments for each target vehicle until each target vehicle reaches the mission endpoint.

[0041] In summary, this embodiment uses the AGV performing the new task as the target vehicle and the AGV that does not update its path as the reference vehicle. The initial travel path of the target vehicle is determined based on the reference vehicle's path. When the target vehicle reaches the termination node of the initial travel segment, the analysis data for subsequent candidate road segments is re-predicted. Based on the difference between the initial predicted analysis data and the actual analysis data of the initial travel segment, the difference between the second predicted analysis data and the initial predicted analysis data of the subsequent candidate road segments, and the number of alternative paths for the subsequent candidate road segments, the second travel segment of the target vehicle is obtained. This process continues until the target vehicle reaches the task endpoint. This invention analyzes the target vehicle's operating status based on the reference vehicle, effectively improving the accuracy and stability of the target vehicle's path setting.

[0042] Example 2: This invention also proposes an AGV (Automated Guided Vehicle) material handling path optimization system for intelligent warehousing; please refer to [link to relevant documentation]. Figure 2 The diagram illustrates a structural diagram of an AGV (Automated Guided Vehicle) material retrieval path optimization system for intelligent warehousing, provided by an embodiment of the present invention. The system includes: a data acquisition module 10, an initial travel path acquisition module 20, a second travel segment acquisition module 30, and a data processing module 40.

[0043] The data acquisition module 10 is used to take the AGV that needs to perform a new task as the target vehicle, and obtain the candidate road segment and the initial predicted arrival time of the corresponding node of each target vehicle; and to take the AGV that does not need to update its path as the reference vehicle, and obtain the running path of each reference vehicle.

[0044] The initial driving path acquisition module 20 is used to predict the initial predicted reference traffic flow of each candidate road segment based on the running path of the reference vehicle, and to determine the initial driving path of each target vehicle.

[0045] The second driving segment acquisition module 30 is used to re-predict the second predicted reference traffic flow and the second predicted arrival time of the node of the subsequent candidate road segment when the target vehicle travels to the end node of the initial driving segment according to the initial driving path; the arrival time and reference traffic flow are used as analysis data; and the second driving segment of the target vehicle is obtained based on the difference between the initial predicted analysis data and the actual analysis data corresponding to the initial driving segment, the difference between the second predicted analysis data and the initial predicted analysis data corresponding to the subsequent candidate road segment, and the number of alternative paths of the subsequent candidate road segment.

[0046] The data processing module 40 is used to obtain subsequent driving segments as the target vehicle travels along the second driving segment until the target vehicle reaches the mission endpoint.

[0047] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the AGV cart material picking path optimization system for intelligent warehousing and the AGV cart material picking path optimization method embodiment for intelligent warehousing provided in the above embodiments belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0048] Example 3: This invention also proposes an AGV (Automated Guided Vehicle) material handling path optimization device for intelligent warehousing. The device includes a memory and a processor. The memory stores executable program code, and the processor calls and executes this executable program code to perform the AGV material handling path optimization method for intelligent warehousing provided in the embodiments of this application. Specifically, the device may be a chip, component, or module. The chip may include a connected processor and memory; the memory stores instructions, and when the processor calls and executes the instructions, the chip can perform the AGV material handling path optimization method for intelligent warehousing provided in the above embodiments.

[0049] Furthermore, this application also protects a computer device; please refer to [link to relevant documentation]. Figure 3 The computer device includes a memory 401, a processor 402, and a computer program 403 stored in the memory 401 and running on the processor 402. When the processor 402 executes the computer program 403, the computer device can execute any of the aforementioned AGV material handling path optimization methods for intelligent warehousing.

[0050] Example 4: This embodiment also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer, the computer executes the above-described method steps to implement the AGV material picking path optimization method for intelligent warehousing provided in the above embodiment.

[0051] Example 5: This embodiment also provides a computer program product. When the computer program product is run on a computer, it causes the computer to perform the above-mentioned related steps to realize the AGV trolley material picking path optimization method for intelligent warehousing provided in the above embodiment.

[0052] In this embodiment, the device, computer-readable storage medium, computer program product, or chip are all used to execute the corresponding methods provided above. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0053] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0054] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing, characterized in that, The method includes the following steps: For each AGV that needs to perform a new task, the candidate road segment and the initial predicted arrival time of the corresponding node are obtained. For each AGV that does not need to update its path, the running path of each AGV is obtained. Based on the reference vehicle's running path, the initial predicted reference vehicle flow for each candidate road segment is predicted, and the initial driving path for each target vehicle is determined. When the target vehicle travels to the end node of the initial travel segment according to the initial travel path, the second predicted reference traffic flow of the subsequent candidate road segments and the second predicted arrival time of its nodes are re-predicted; the arrival time and reference traffic flow are used as analysis data; based on the difference between the initial predicted analysis data and the actual analysis data corresponding to the initial travel segment, the difference between the second predicted analysis data and the initial predicted analysis data corresponding to the subsequent candidate road segments, and the number of alternative paths for the subsequent candidate road segments, the second travel segment of the target vehicle is obtained; The target vehicle travels along the second driving segment, and then obtains subsequent driving segments in sequence until the target vehicle reaches the mission endpoint.

2. The method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing as described in claim 1, characterized in that, The method for obtaining the initial driving path is as follows: For any target vehicle, the initial driving path of the target vehicle is obtained by using the dynamic A* algorithm based on the initial predicted reference traffic flow of each candidate road segment and the initial predicted arrival time of the corresponding node of the candidate road segment.

3. The method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing as described in claim 1, characterized in that, The method for obtaining the second driving segment is as follows: For any target vehicle, the termination node of the initial driving segment of the target vehicle is taken as the first analysis node. Based on the difference between the initial predicted analysis data and the actual analysis data corresponding to the initial driving segment of the target vehicle, and the difference between the second predicted analysis data and the initial predicted analysis data corresponding to the subsequent candidate segments of the first analysis node, the first predicted influence degree of each candidate segment of the target vehicle after the first analysis node is obtained. Starting from the first analysis node, alternative paths to the target vehicle's mission endpoint are obtained, and all of them are used as the first analysis path; For any candidate road segment of the target vehicle following the first analysis node, the first analysis path containing the candidate road segment will be used as the designated path for that candidate road segment. For any given path, the sum of the first predicted impact of the candidate road segment and its subsequent candidate road segments on the given path is taken as the overall impact of the candidate road segment on the given path. Based on the overall impact of the candidate road segment on each specified path, the replacement cost of the candidate road segment is obtained; The correction cost of the candidate road segment is obtained based on the replacement cost of the candidate road segment and the number of first analysis paths. The product of the correction cost and the first predicted impact level of the candidate road segment is taken as the comprehensive impact level of the candidate road segment; Based on the dual objectives of minimizing the overall impact and minimizing the path length, the second driving path of the target vehicle is obtained through the dynamic A* algorithm, and the second driving segment of the target vehicle is determined.

4. The method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing as described in claim 3, characterized in that, The method for obtaining the first predicted impact level is as follows: The difference between the initial predicted reference traffic flow and the actual reference traffic flow of the initial travel segment of the target vehicle is taken as the first difference; The difference between the initial predicted arrival time and the actual arrival time of the target vehicle at the initial end point of its initial travel segment is taken as the second difference; For any candidate road segment of the target vehicle after the first analysis node, the difference between the initial predicted reference traffic flow and the second predicted reference traffic flow of the candidate road segment is taken as the third difference; The difference between the initial predicted arrival time and the second predicted arrival time of the candidate road segment termination node is taken as the fourth difference; The product of the first preset weight and the first difference is used as the first feature value; The product of the second preset weight and the second difference is used as the second feature value; The product of the third preset weight and the third difference is used as the third feature value; The product of the fourth preset weight and the fourth difference is taken as the fourth feature value; wherein, the sum of the first preset weight, the second preset weight, the third preset weight and the fourth preset weight is 1; The sum of the first, second, third, and fourth eigenvalues, followed by normalization, is used as the first predicted influence level of the candidate road segment.

5. The method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing as described in claim 3, characterized in that, The method for obtaining the replacement cost is as follows: When there are at least two specified paths for a candidate road segment, the overall impact of the candidate road segment is arranged in ascending order to obtain a sequence of overall impact. If the first overall impact degree in the overall impact degree sequence is not equal to the second overall impact degree, then the difference between the second overall impact degree and the first overall impact degree shall be taken as the first value; The ratio of the first value to the second overall impact level is used as the replacement cost of the candidate road segment; If the first overall impact level in the overall impact level sequence is equal to the second overall impact level, then the default value is 0 as the replacement cost of the candidate road segment; When there is only one specified path for the candidate road segment, the default value is 1, which is the replacement cost of the candidate road segment.

6. The method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing as described in claim 3, characterized in that, The method for obtaining the correction cost is as follows: The ratio of the replacement cost of the candidate road segment to the number of paths analyzed in the first analysis is used as the correction cost of the candidate road segment.

7. The method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing as described in claim 1, characterized in that, The alternative path refers to all feasible paths that start from the end node of the initial driving segment, reach the mission endpoint of the corresponding target vehicle, and can cover the corresponding candidate segment.

8. The method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing as described in claim 1, characterized in that, The method for obtaining the predicted arrival time is as follows: For any target vehicle, the shortest path from the current node to the termination node of each subsequent candidate road segment is obtained through the dynamic A* algorithm, and used as the reference path for the corresponding node. The product of the length of the reference path and the average speed of the target vehicle is used as the predicted arrival time of the corresponding node.

9. The method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing as described in claim 1, characterized in that, The method for obtaining the initial predicted reference traffic flow is as follows: For any candidate road segment of any target vehicle, the time period consisting of the initial predicted arrival times of the start and end nodes of the candidate road segment is taken as the initial predicted travel time period of the candidate road segment. The number of all reference vehicles located on the candidate road segment is counted and predicted within the initial predicted driving time period, and this number is used as the initial predicted reference traffic flow for the target vehicle on the candidate road segment.

10. The method for optimizing the material handling path of an AGV (Automated Guided Vehicle) for intelligent warehousing as described in claim 1, characterized in that, The optimization of the material retrieval path for the AGV in the intelligent warehousing establishes a data interaction link with the AGV scheduling system, enterprise resource planning system, warehouse management system, and warehouse control system to obtain the task information of the target vehicle, warehouse road network information, and the real-time operating status of the reference vehicle. Based on the task information and warehouse road network information, candidate road segments for the target vehicle are determined. The operating path of the reference vehicle is obtained through historical planning data and real-time positioning data from the warehouse control system.

Images