Multi-objective path collaborative optimization control method and system of warehouse handling equipment

Abstract

The multi-target path cooperative optimization control method and system of the warehouse handling equipment belong to the technical field of automatic logistics, and can perform risk prediction and hierarchical control, effectively prevent system-level congestion and deadlock, improve the operation stability under extreme working conditions, realize intelligent traffic diversion and task sequencing through local risk perception and global dynamic priority adjustment, reduce the invalid waiting and conflict of AGVs, improve the overall material turnover rate, ensure that emergency tasks such as flood prevention materials are preferentially executed on the globally optimal path, and shorten the total end-to-end completion time.

Description

Technical Field

[0001] This invention relates to the field of automated logistics technology, specifically to a multi-objective path collaborative optimization control method and system for warehousing and handling equipment. Background Technology

[0002] In modern warehouse automation systems, AGV clusters are key to improving material handling efficiency. Existing technologies mostly employ distributed control, giving individual AGVs real-time decision-making capabilities. However, when handling time-sensitive emergency allocation tasks such as flood control supplies, if sudden obstacles are frequent and concentrated at critical path nodes, existing distributed decision-making mechanisms, due to incomplete information, are prone to system-level path oscillations and deadlocks, severely impacting material allocation efficiency and reliability. This is a pressing technical problem that needs to be solved in this field.

[0003] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] The purpose of this invention is to provide a multi-objective path collaborative optimization control method and system for warehousing and handling equipment, so as to solve the problems mentioned in the background art.

[0005] The technical solution of the present invention includes the following steps:

[0006] Step 1: Obtain real-time status data of each automated guided vehicle in the warehouse environment, and based on the real-time status data and the preset historical data of the warehouse management system, evaluate the preset critical path nodes and construct the spatiotemporal instability index of the critical path nodes.

[0007] Step 2: Preset the coordination activation threshold, and generate control mode instructions based on the comparison result between the spatiotemporal instability index and the coordination activation threshold;

[0008] Step 3: In response to the control mode command, when the control mode command is a global coordination optimization mode, perform global task and path coordination optimization; when the control mode command is a local autonomous decision-making mode, perform local predictive risk avoidance to output the coordinated control path of the automated guided vehicle.

[0009] Preferably, step one includes:

[0010] S11. The automated guided vehicle (AGV) uses its own sensors to acquire the AGV's position data, velocity vector data, and planned path point data, which together form real-time status data.

[0011] S12. Collect the real-time status data and count the number of new obstacles that appear within the influence radius of the critical path node within the preset observation time window to obtain the obstacle burst frequency.

[0012] S13. Based on the real-time status data, calculate the ratio of the number of automated guided vehicles (AGVs) entering the influence radius of the critical path node to the preset capacity of the critical path node, and obtain the dynamic AGV density.

[0013] S14. Weighted fusion of the frequency of sudden obstacles, the density of dynamic automated guided vehicles, and historical congestion factors mined from historical data of the warehouse management system is performed to construct the spatiotemporal instability index of the critical path node.

[0014] Preferably, step two includes:

[0015] S21. Compare the spatiotemporal instability index of each critical path node with the coordination activation threshold in real time.

[0016] S22. When the spatiotemporal instability index of any of the critical path nodes reaches or exceeds the coordination activation threshold, a global coordination optimization mode instruction is generated.

[0017] S23. When the spatiotemporal instability index of all critical path nodes is lower than the coordination activation threshold, a local autonomous decision-making mode instruction is generated.

[0018] Preferably, when the control mode instruction is a global coordination optimization mode, the steps for performing global task and path coordination optimization include:

[0019] S31. Identify the risk areas defined by critical path nodes whose spatiotemporal instability index exceeds the limit, and determine the automated guided vehicles (AGVs) whose path planning passes through the risk areas as associated AGVs.

[0020] S32. Calculate the effective priority of the tasks performed by the associated automated guided vehicle;

[0021] S33. Based on the calculated effective priority, replan the route corridor and passage order for the associated automated guided vehicles, and issue route instructions.

[0022] Preferably, the steps for calculating the effective priority include:

[0023] S321. Based on the spatiotemporal instability index of the critical path nodes traversed by the associated automated guided vehicle, calculate the risk-weighted length of the planned path of the associated automated guided vehicle.

[0024] S322. Calculate the path executability score based on the ratio of the geometric length of the planned path to the risk-weighted length;

[0025] S323. Using the logistic function model, combined with the path executability score and the preset basic priority of the task, the effective priority is generated.

[0026] Preferably, when the control mode instruction is a local autonomous decision-making mode, the steps for performing local predictive risk avoidance include:

[0027] S34. Preset a risk warning threshold and compare the spatiotemporal instability index of the critical path node with the risk warning threshold;

[0028] S35. Based on the comparison results of S34, calculate the node risk repulsion force exerted by the critical path node on the automated guided vehicle; wherein, when the spatiotemporal instability index exceeds the risk warning threshold, the node risk repulsion force is the calculated value, and when the spatiotemporal instability index does not exceed the risk warning threshold, the node risk repulsion force is zero.

[0029] S36. The node risk repulsion force, target attraction force, obstacle repulsion force, and automatic guidance workshop repulsion force are vector summed to form the total navigation force;

[0030] S37. Based on the total navigation force, control the real-time path of the automated guided vehicle.

[0031] A multi-objective path collaborative optimization control system for warehouse handling equipment includes:

[0032] The system status assessment module is used for:

[0033] Acquire real-time status data of each automated guided vehicle in the warehouse environment, and acquire historical data from the preset warehouse management system;

[0034] Based on the real-time status data and the historical data, the preset critical path nodes are evaluated to construct the spatiotemporal instability index of the critical path nodes.

[0035] The control mode decision module is used for:

[0036] Store preset coordination activation thresholds;

[0037] Based on the comparison between the spatiotemporal instability index and the coordination activation threshold, a global coordination optimization mode instruction or a local autonomous decision-making mode instruction is generated.

[0038] The global collaborative optimization module is used for:

[0039] In response to the global coordination optimization mode instruction;

[0040] Identify associated automated guided vehicles within the risk area and calculate the effective priority of their tasks;

[0041] Based on the effective priority, the associated automated guided vehicle replans its route corridor and passage order to generate route instructions.

[0042] The local risk avoidance module is used for:

[0043] In response to the local autonomous decision-making mode instruction;

[0044] Based on the aforementioned spatiotemporal instability index, the autonomous guided vehicle's own navigation power is adjusted to perform local predictive risk avoidance.

[0045] Upon receiving a path instruction from the global collaborative optimization module, navigation is performed based on the path instruction.

[0046] This invention improves the multi-objective path collaborative optimization control method and system for warehousing and handling equipment, and compared with the prior art, it has the following improvements and advantages:

[0047] 1. This invention can perform risk prediction and hierarchical control, effectively prevent system-level congestion and deadlock, and improve operational stability under extreme conditions;

[0048] 2. This invention achieves intelligent flow management and task sequencing through local risk perception and global dynamic priority adjustment, reducing AGV's ineffective waiting and conflicts, and improving the overall material turnover rate; Attached Figure Description

[0049] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0050] Figure 1 This is a flowchart of the system of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0052] Example 1:

[0053] Please see Figure 1 The present invention provides a multi-objective path collaborative optimization control method and system technical solution for warehousing and handling equipment, including: Step 1: Obtain the real-time status data of each automated guided vehicle in the warehousing environment, and evaluate the preset critical path nodes based on the real-time status data and the preset historical data of the warehousing management system, and construct the spatiotemporal instability index of the critical path nodes.

[0054] Step 2: Preset the coordination activation threshold, and generate control mode instructions based on the comparison results between the spatiotemporal instability index and the coordination activation threshold;

[0055] Step 3: Responding to control mode commands, when the control mode command is the global coordination and optimization mode, global task and path coordination optimization is performed; when the control mode command is the local autonomous decision-making mode, local predictive risk avoidance is performed to output the coordinated control path of the automated guided vehicle.

[0056] The preset critical path nodes are not simply arbitrary path points, but rather nodes that, determined according to the following systematic method, have a decisive impact on the overall efficiency of the warehousing and handling system; the determination method includes the following steps:

[0057] Step A: Topology analysis; Based on the grid map or topology map of the warehouse environment, automatically identify all path intersections, merging points, diverging points, warehouse entrances and exits, charging pile docking points, and bottleneck sections where the width changes significantly, forming a candidate node set.

[0058] Step B: Historical data-driven node criticality scoring. From the historical data of the warehouse management system, for each candidate node, extract the following data over a long period: Historical average traffic flow: the average number of AGVs passing through the node per unit time; Historical congestion event rate: the frequency of events near the node where the average speed of AGVs falls below a preset threshold due to traffic conflicts or obstacle avoidance; High-priority task relevance: the frequency with which planned paths of high-priority tasks have historically passed through the node. Step C: Weighted scoring and filtering. The node criticality score of each candidate node is calculated using the following formula. :

[0059] ;

[0060] in, , , These are the normalized values ​​of the three historical data points mentioned above. The three weighting coefficients are preset, and their sum is 1. Nodes with scores higher than the preset criticality threshold are selected as critical path nodes for continuous evaluation in this invention. This method ensures the representativeness of the selected nodes and their sensitivity to system risks. Their specific values ​​can be determined by combining the analytic hierarchy process with expert scoring, or by optimizing the results through regression analysis of historical data to maximize the consistency between the scores and the actual bottleneck nodes.

[0061] The criticality threshold is determined based on the distribution of node criticality scores in historical data, such as taking the top 5% of nodes, or by the system administrator based on the actual management needs and operational goals of the warehouse.

[0062] Preset coordination activation threshold The determination of the instability index is based on statistical analysis of the system's historical operating states, aiming to scientifically define the boundary between stability and critical instability. The method is as follows: collect the spatiotemporal instability index of all critical path nodes under normal operating conditions. Historical data is used to form a dataset; the probability distribution of this dataset is calculated, and values ​​at the 95th or 99th percentile are set as the coordinated activation threshold. This statistically based setting method allows the threshold to adapt to the inherent fluctuations of the specific warehousing environment, avoiding the problem of excessively frequent or sluggish system mode switching due to improper experience settings.

[0063] The multi-objective path collaborative optimization control method for warehousing and handling equipment provided in this embodiment aims to overcome the system-level path oscillation and deadlock problem caused by high frequency and concentration of sudden obstacles at critical path nodes in high-time-sensitivity emergency dispatch tasks. The core technology of this method lies in establishing a hierarchical control architecture, which dynamically switches between local autonomous decision-making and global coordinated optimization modes by constructing and evaluating spatiotemporal instability indices that characterize system risks in real time. This not only endows individual automated guided vehicles (AGVs) with efficient local obstacle avoidance capabilities, but also activates a global coordination mechanism before potential risks escalate into systemic congestion, thereby achieving macro-level guidance and traffic diversion for the entire AGV cluster. This ensures the absolute reliability and timeliness of high-priority tasks such as flood control materials, while significantly improving the overall operating efficiency and robustness of the AGV cluster in complex dynamic environments. The excellent computational economy brought about by its hierarchical design provides solid technical support for the deployment of large-scale AGV clusters.

[0064] Example 2:

[0065] Step one includes:

[0066] S11. The automated guided vehicle (AGV) uses its own sensors to acquire the AGV's position data, velocity vector data, and planned path point data, which together form real-time status data.

[0067] S12. Collect real-time status data and count the number of new obstacles that appear within the influence radius of the critical path node within the preset observation time window to obtain the obstacle burst frequency.

[0068] S13. Based on real-time status data, calculate the ratio of the number of automated guided vehicles (AGVs) entering the influence radius of the critical path node to the preset capacity of the critical path node, and obtain the dynamic AGV density.

[0069] S14. Weighted fusion of obstacle occurrence frequency, dynamic automated guided vehicle density, and historical congestion factors mined from historical data of the warehouse management system is performed to construct a spatiotemporal instability index for critical path nodes.

[0070] Step two includes:

[0071] S21. Compare the spatiotemporal instability index of each critical path node with the coordinated activation threshold in real time.

[0072] S22. When the spatiotemporal instability index of any critical path node reaches or exceeds the coordination activation threshold, a global coordination optimization mode instruction is generated.

[0073] S23. When the spatiotemporal instability index of all critical path nodes is lower than the coordination activation threshold, generate a local autonomous decision-making mode instruction.

[0074] In this embodiment, the construction of the spatiotemporal instability index of critical path nodes is the technological cornerstone for realizing the shift from passive response to predictive risk management. The construction of this index aims to address the technical pain point of existing distributed decision-making mechanisms being unable to cope with high-frequency, concentrated obstacles due to incomplete information. The motivation is to create a comprehensive indicator that can accurately quantify and predict the congestion risk of critical nodes, serving as a direct and reliable basis for dynamic decision-making in subsequent hierarchical control modes. The construction logic of this index is not a linear superposition, but rather a deep integration of risk engineering theory, mathematically modeling the three dimensions of recent trends, current state, and historical experience required for risk prediction.

[0075] Spacetime Instability Index The calculation formula is presented as follows:

[0076] ;

[0077] In this formula, all parameters are either normalized or are dimensionless ratios, ensuring that... It is a dimensionless comprehensive risk score; where:

[0078] Represents the critical path node At any moment The instability index;

[0079] This indicates the frequency of obstacle occurrences; the data comes from the AGV sensors over the past observation window. Within, for example, 60 seconds, at the node The number of new obstacles detected within the influence radius is used to quantify the real-time dynamic changes in the environment;

[0080] The maximum tolerable frequency is an engineering parameter set based on actual warehouse environment stress testing. It is used to normalize the frequency of obstacle occurrences, ensuring that this term is dimensionless. This stress test refers to maintaining a constant AGV task load in a simulation environment and gradually increasing the frequency of randomly generated temporary obstacles near critical path nodes. When the observed decrease in total system throughput exceeds the preset performance tolerance, the obstacle generation frequency at this point is recorded as the maximum tolerable frequency. ;

[0081] The dynamic AGV density is defined as follows: This is used to reflect the real-time congestion level of the nodes; here... The total number of AGVs, indicated by the function. Used to determine if ID is AGV position Is it at the node? Location radius of influence Internally, and node capacity It is a priori parameter based on road network geometry information, therefore the density is a dimensionless ratio.

[0082] The historical congestion factor is a normalized historical congestion rate of nodes statistically mined from historical data of the Warehouse Management System (WMS), reflecting the model's self-learning and experience-based learning capabilities.

[0083] These represent dimensionless weighted coefficients, with the sum of the three being 1. These coefficients are key adjustable parameters of the model; their values ​​are not arbitrarily set but rather determined based on a specific warehouse layout. They are obtained through multiple regression analysis in a simulation environment or by training and optimizing using reinforcement learning algorithms. The goal is to optimize the constructed model. The index has the highest predictive correlation with actual future traffic congestion events;

[0084] During system operation, the central coordination server periodically or event-triggeredly calculates this index; after the calculation is completed, the system immediately transmits the results to each key node. With preset coordination activation threshold Perform real-time comparison; if all nodes All below If the system is stable, it is determined to be stable, and instructions for a local autonomous decision-making mode are generated; otherwise, if any node... Reaching or exceeding If the system is about to enter an unstable state, it will immediately generate a global coordination and optimization mode instruction. This decision-making mechanism based on quantitative risk prediction enables the system to accurately identify potential sources of congestion and preemptively switch to a more macroscopic control mode, thereby effectively avoiding deadlocks and path oscillations that are prone to occur under traditional distributed control, and greatly enhancing the operational stability of the entire warehousing system.

[0085] Example 3:

[0086] When the control mode command is in global coordination and optimization mode, the steps for performing global task and path coordination optimization include:

[0087] S31. Identify the risk areas defined by critical path nodes whose spatiotemporal instability index exceeds the limit, and determine the Automated Guided Vehicles (AGVs) whose path planning passes through the risk areas as associated AGVs.

[0088] S32. Calculate the effective priority of the tasks performed by the associated automated guided vehicle;

[0089] S33. Based on the calculated effective priority, replan the route corridor and passage order for the associated automated guided vehicles and issue route instructions.

[0090] The steps for calculating effective priority include:

[0091] S321. Based on the spatiotemporal instability index of the critical path nodes traversed by the associated automated guided vehicle (AGV), calculate the risk-weighted length of the planned path of the associated AGV.

[0092] S322. Calculate the path executability score based on the ratio of the geometric length to the risk-weighted length of the planned path;

[0093] S323. Using the logistic function model, combined with the path executability score and the preset basic priority of the task, an effective priority is generated.

[0094] A path corridor is a macroscopic path planning instruction, not a rigid trajectory accurate to the centimeter level, but a sequence of ordered critical path nodes. And the suggested passage time window associated with each node. A shared data structure; after receiving a path corridor instruction, the AGV will use the node sequence as its navigation target. At the same time, while moving from one node to the next, it will retain its local, sensor-based artificial potential field obstacle avoidance capability to deal with unforeseen obstacles inside the corridor. This flexible corridor definition achieves a balance between the guidance of global planning and the flexibility of local decision-making.

[0095] The global task and path coordination optimization in this embodiment is the core strategy for the system to cope with high-risk conditions. The essence of this strategy is that it does not handle tasks rigidly, but introduces the concept of opportunity cost from operations research. It readjusts the effective priority of tasks by dynamically evaluating the actual executability of the path. When the system triggers the global coordination optimization mode, in order to avoid delays of high-priority tasks due to path congestion or low-priority tasks occupying critical channels, this invention proposes a dynamic task priority function. The motivation is to enable task priority to reflect the risk level of its execution path in real time, ensuring that truly urgent tasks with smooth paths receive the highest priority.

[0096] This process begins with accurately calculating the risk-weighted length of the path and the path executability score for AGVs. The planned path The path is defined as a path consisting of... An ordered sequence of key path nodes The path contains A path segment connecting consecutive nodes. Each segment (from node) point to The geometric length of ) is denoted as .

[0097] Calculation path Risk-weighted length It is defined as the sum of the products of the geometric length of each segment of the path and the risk index of its target node:

[0098] ;

[0099] In the formula, It is the first Target node of segment path At any moment The spatiotemporal instability index. The physical meaning of this formula is: the cost of path segmentation depends not only on its length, but also amplified by the congestion risk of its endpoint;

[0100] Based on the total geometric length of the path Its risk-weighted length Execution score of the computation path The formula is:

[0101] ;

[0102] In the formula, here It is a dimensionless ratio with a range of . This score intuitively reflects the "cost-effectiveness" of the path; the closer the score is to 1, the lower the risk and the higher the feasibility of the path.

[0103] Based on this score, a dynamic task priority function is proposed. It was constructed by drawing on the logistic function model in statistics to achieve smooth and bounded dynamic adjustment of priorities: In this formula, and All are dimensionless priority scores, and the output range of the logistic function is... The dimensionless adjustment factor ensures that the dimensions of both sides of the equation are consistent; where:

[0104] Indicates task At any moment The effective priority is the direct basis for scheduling decisions;

[0105] This indicates the static base priority of the task specified by the business layer;

[0106] It is the path executability score calculated in the previous step, which is the core link between path status and task priority, and reflects the direct impact of path status on task priority;

[0107] and These are the adjustment parameters of the logistic function, and they need to be jointly calibrated. To strike a balance for feasibility, the source can be set to the system under normal operating conditions. The statistical average; The steepness of the penalty can be controlled, and its source can be set according to operational needs; for example, for flood control materials that absolutely cannot be delayed, a smaller penalty can be set. This makes its priority decay more gradual; for ordinary tasks, a larger priority setting can be used. This causes its priority to drop rapidly when the path deteriorates, prompting it to give way proactively;

[0108] In global coordination mode, after the central coordination server identifies the associated AGVs affected by the risk area, it calculates the dynamic effective priority of the tasks to be performed by these AGVs using the formula mentioned above, only for these AGVs. Based on the adjusted new priority sequence, the system re-plans the macro-path corridors and passage order through risk areas. The effect is that if a task with a high basic priority experiences a surge in risk and a sharp drop in feasibility score on its planned path, its effective priority will automatically decrease, making way for other tasks with better paths. This intelligent traffic diversion and task reordering not only resolves congestion at its source but also ensures that truly urgent tasks such as flood control materials are always assigned to the globally optimal path for priority execution, significantly improving overall operational efficiency and the reliability of high-priority tasks.

[0109] The replanning process can be implemented using an improved A* algorithm, where the cost function for path search no longer considers only the geometric length of the path, but also the effective priority of the task. As a key adjustment factor, for example, when selecting the next expansion node, priority should be given to paths that can serve higher-priority tasks. For planning the passage order, different passage time windows can be allocated to associated automated guided vehicles (AGVs) passing through the same risk area based on their effective priority, ensuring that high-priority tasks proceed first.

[0110] Example 4:

[0111] When the control mode instruction is a local autonomous decision-making mode, the steps for performing local predictive risk avoidance include:

[0112] S34. Preset risk warning thresholds and compare the spatiotemporal instability index of critical path nodes with the risk warning thresholds;

[0113] S35. Based on the comparison results of S34, calculate the node risk repulsion force exerted by the critical path node on the automated guided vehicle; where, when the spatiotemporal instability index exceeds the risk warning threshold, the node risk repulsion force is the calculated value, and when the spatiotemporal instability index does not exceed the risk warning threshold, the node risk repulsion force is zero.

[0114] S36. The repulsive forces of node risks, the attractive forces of targets, the repulsive forces of obstacles, and the repulsive forces of the automated guidance workshop are vector summed to form the total navigation force;

[0115] S37. Based on total navigation power, control the real-time path of the automated guided vehicle.

[0116] Risk warning threshold The settings are intended to create an effective early warning buffer; the value is consistent with the coordinated activation threshold. The association is confirmed and set as follows:

[0117] ;

[0118] in, It is a buffer coefficient, and its value usually ranges from 100 to 1000. Between, for example, can be set This setting ensures that when the node's risk index exceeds... When the system reaches 70% capacity, the AGV begins preventative local avoidance, rather than waiting until the system approaches the critical point of global congestion. The timely response demonstrates the foresight in risk management;

[0119] The local predictive risk avoidance in this embodiment is a significant technological innovation of the classic artificial potential field method. The core technology lies in transforming macroscopic system-level risk warning into microscopic driving forces that each AGV can perceive. The motivation for this is to make each AGV not only an independent decision-maker, but also a perceiver and responder of the entire system's state. By preventing potential risks in the distance, we can jointly maintain the smooth operation of the system and reduce the frequency of triggering global coordination.

[0120] This avoidance capability is achieved through a newly added node risk repulsion force. To achieve this; AGV Total navigation force It consists of four vector parts: The first three terms are known in the field as attraction and repulsion terms, while the innovation lies in the fourth term, nodal risk repulsion. The calculation formula is: To ensure dimensional consistency, Force, dimensionless ; right side of the formula and Both are dimensionless; position vector difference The dimension of is length its cube The dimensions are Therefore, the dimensions of the fractional term are: To make the equation hold, the gain coefficient... The physical dimension of the force must be the force multiplied by the square of the length, that is... ;in:

[0121] Indicates key nodes For AGV The applied virtual repulsive force;

[0122] This represents the risk repulsion gain coefficient, whose physical dimension is force multiplied by the square of length, such as... This is an adjustable engineering parameter, the value of which is calibrated through simulation experiments. The goal is to achieve a balance between effectively managing traffic flow and avoiding unnecessary detours.

[0123] It is a real-time instability index of each key node, broadcast by the central coordination server.

[0124] This represents the risk warning threshold; the logic behind setting this threshold is to provide a buffer zone between normal and dangerous conditions for the system, and its value must be lower than the coordination activation threshold. For example, if , can be set This allows the AGV to begin preventative avoidance within this buffer zone, rather than waiting until the system becomes critically unstable before reacting.

[0125] They represent AGVs respectively and nodes The position vector;

[0126] The function ensures that the repulsion only takes effect when the node risk exceeds the warning threshold;

[0127] Node capacity These are prior parameters based on road network geometry information, influencing the radius. The size is based on the physical path width at the node's location. Determined, it can usually be set to This is to ensure effective coverage of the area near the node where interactions may occur;

[0128] In the local autonomous decision-making mode, in addition to performing routine pathfinding and obstacle avoidance, each AGV will continuously receive key node broadcasts from the central server. Index; once any node of The index exceeded The AGV will then calculate the repulsive force pointing away from the node based on the above formula. It is then superimposed with other navigation force vectors to form the total navigation force. This allows for fine-tuning of the AGV's real-time path; this mechanism creates a self-organizing traffic management effect: when the risk of congestion in a certain area is just emerging but not yet serious enough to require global intervention, the passing AGVs have already begun to smoothly detour; this predictive local risk avoidance enables intelligent micro-adjustment of traffic flow, effectively reducing the ineffective waiting of AGVs and local conflicts, thereby improving the overall material turnover rate and reducing the probability of the system entering a high-cost global coordination mode.

[0129] Example 5:

[0130] A multi-objective path collaborative optimization control system for warehouse handling equipment includes: a system status assessment module, used for:

[0131] Acquire real-time status data of each automated guided vehicle in the warehouse environment, and acquire historical data from the preset warehouse management system;

[0132] Based on real-time status data and historical data, the preset critical path nodes are evaluated to construct a spatiotemporal instability index for critical path nodes.

[0133] The control mode decision module is used for:

[0134] Store preset coordination activation thresholds;

[0135] Based on the comparison results between the spatiotemporal instability index and the coordination activation threshold, global coordination optimization mode instructions or local autonomous decision-making mode instructions are generated.

[0136] The global collaborative optimization module is used for:

[0137] Responding to global coordination optimization mode instructions;

[0138] Identify associated automated guided vehicles within the risk area and calculate the effective priority of their tasks;

[0139] Based on effective priorities, the associated automated guided vehicles (AGVs) are replanned for route corridors and passage order to generate route instructions.

[0140] The local risk avoidance module is used for:

[0141] Responding to instructions from a local autonomous decision-making mode;

[0142] Based on the spatiotemporal instability index, the autonomous vehicle's navigation capabilities are adjusted to perform local predictive risk avoidance.

[0143] Upon receiving a path instruction from the global collaborative optimization module, navigation is performed based on the path instruction.

[0144] The multi-objective path collaborative optimization control system for warehousing and handling equipment provided in this embodiment is physically implemented by typically including an AGV vehicle cluster, a central coordination server, and control software deployed on it. The functional division of the control software closely corresponds to the aforementioned method steps, and its core consists of a local perception and rapid response module deployed on the AGV end, and a global coordination and optimization module deployed on the central coordination server end.

[0145] The system status assessment module and control mode decision module are implemented on the central coordination server; the former is responsible for collecting real-time status data such as position and speed from all AGVs, and combining it with historical data retrieved from the warehouse management system (WMS), based on the aforementioned... The formula continuously calculates the spatiotemporal instability index for each key node; the latter then calculates the... Coordinated activation threshold of exponent and storage The comparison is performed, and based on the comparison results, control mode instructions that the entire system should follow are generated and broadcast.

[0146] The global collaborative optimization module is also located on the central coordination server; this module is activated upon receiving a global collaborative optimization mode command; based on the exceeded limits... The index identifies risk areas and associated AGVs, calls the dynamic task priority function model to recalculate the effective priorities for the tasks of these AGVs, plans new path corridors and passage sequences based on the new priorities, and issues these macro-path instructions to specific associated AGVs.

[0147] The local risk avoidance module is mainly implemented on each AGV body; after receiving the local autonomous decision-making mode instruction, this module will continuously receive broadcasts from the server. The index, based on the node risk repulsion formula, adjusts its total navigation force in real time to achieve predictive risk avoidance at the micro level; when an AGV that was originally in local mode is identified as an associated AGV by the global module and receives a new path instruction, the local risk avoidance module will follow the macro instruction as the new navigation target, and continue to perform fine real-time obstacle avoidance within the path corridor, thus achieving seamless connection between global instructions and local execution.

[0148] This system architecture achieves optimal allocation of computing resources by placing high-frequency local environmental awareness and rapid response at the AGV end, while placing risk assessment, pattern decision-making, and macro-planning, which require global information and have a larger computational load, at the central coordination server. This hierarchical collaborative working mode ensures that the system can make both rapid local responses and far-sighted global planning when dealing with complex scenarios such as the emergency allocation of flood control materials, thereby fundamentally improving the efficiency, stability, and reliability of the automated warehousing system.

[0149] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A multi-objective path collaborative optimization control method for warehousing and handling equipment, characterized by the following steps: include: Step 1: Obtain real-time status data of each automated guided vehicle in the warehouse environment, and based on the real-time status data and the preset historical data of the warehouse management system, evaluate the preset critical path nodes and construct the spatiotemporal instability index of the critical path nodes. Step 2: Preset the coordination activation threshold, and generate control mode instructions based on the comparison result between the spatiotemporal instability index and the coordination activation threshold; Step 3: In response to the control mode command, when the control mode command is a global coordination optimization mode, perform global task and path coordination optimization; when the control mode command is a local autonomous decision-making mode, perform local predictive risk avoidance to output the coordinated control path of the automated guided vehicle. When the control mode command is the global coordination optimization mode, the steps for performing global task and path coordination optimization include: S31. Identify the risk areas defined by critical path nodes whose spatiotemporal instability index exceeds the limit, and determine the automated guided vehicles (AGVs) whose path planning passes through the risk areas as associated AGVs. S32. Calculate the effective priority of the tasks performed by the associated automated guided vehicle; S33. Based on the calculated effective priority, replan the route corridor and passage order for the associated automated guided vehicles, and issue route instructions; The steps for calculating effective priority include: Based on the spatiotemporal instability index of the critical path nodes traversed by the associated automated guided vehicle (AGV), the risk-weighted length of the planned path of the associated AGV is calculated. Based on the total geometric length of the path Its risk-weighted length Execution score of the computation path The formula is: ； In the formula, The feasibility score of a path is a dimensionless ratio. The total geometric length of the path; Risk-weighted length of the path; This represents the total number of path segments that connect consecutive nodes within the path. : The first of the paths Each segment; : No. The geometric length of each segment; : No. Target node of segment path At any moment The spatiotemporal instability index; The effective priority is generated by using the logistic function model and combining the path executability score with the preset basic priority of the task.

2. The multi-objective path collaborative optimization control method for warehousing and handling equipment according to claim 1, characterized in that, Step one includes: S11. The automated guided vehicle (AGV) uses its own sensors to acquire the AGV's position data, velocity vector data, and planned path point data, which together form real-time status data. S12. Collect the real-time status data and count the number of new obstacles that appear within the influence radius of the critical path node within the preset observation time window to obtain the obstacle burst frequency. S13. Based on the real-time status data, calculate the ratio of the number of automated guided vehicles (AGVs) entering the influence radius of the critical path node to the preset capacity of the critical path node, and obtain the dynamic AGV density. S14. Weighted fusion of obstacle occurrence frequency, dynamic automated guided vehicle density, and historical congestion factors mined from historical data of the warehouse management system is performed to construct the spatiotemporal instability index of the critical path node.

3. The multi-objective path collaborative optimization control method for warehousing and handling equipment according to claim 1, characterized in that, Step two includes: S21. Compare the spatiotemporal instability index of each critical path node with the coordination activation threshold in real time. S22. When the spatiotemporal instability index of any of the critical path nodes reaches or exceeds the coordination activation threshold, a global coordination optimization mode instruction is generated. S23. When the spatiotemporal instability index of all critical path nodes is lower than the coordination activation threshold, a local autonomous decision-making mode instruction is generated.

4. The multi-objective path collaborative optimization control method for warehousing and handling equipment according to claim 1, characterized in that, When the control mode instruction is a local autonomous decision-making mode, the steps for performing local predictive risk avoidance include: S34. Preset a risk warning threshold and compare the spatiotemporal instability index of the critical path node with the risk warning threshold; S35. Based on the comparison results of S34, calculate the node risk repulsion force exerted by the critical path node on the automated guided vehicle; wherein, when the spatiotemporal instability index exceeds the risk warning threshold, the node risk repulsion force is the calculated value, and when the spatiotemporal instability index does not exceed the risk warning threshold, the node risk repulsion force is zero. S36. The node risk repulsion force, target attraction force, obstacle repulsion force, and automatic guidance workshop repulsion force are vector summed to form the total navigation force; S37. Based on the total navigation force, control the real-time path of the automated guided vehicle.

5. A multi-objective path collaborative optimization control system for warehousing and handling equipment, used to implement the multi-objective path collaborative optimization control method for warehousing and handling equipment as described in claim 1, characterized in that, include: The system status assessment module is used for: Acquire real-time status data of each automated guided vehicle in the warehouse environment, and acquire historical data from the preset warehouse management system; Based on the real-time status data and the historical data, the preset critical path nodes are evaluated to construct the spatiotemporal instability index of the critical path nodes. The control mode decision module is used for: Store preset coordination activation thresholds; Based on the comparison between the spatiotemporal instability index and the coordination activation threshold, a global coordination optimization mode instruction or a local autonomous decision-making mode instruction is generated. The global collaborative optimization module is used for: In response to the global coordination optimization mode instruction; Identify associated automated guided vehicles within the risk area and calculate the effective priority of their tasks; Based on the effective priority, the associated automated guided vehicle replans the route corridor and passage order to generate route instructions; The local risk avoidance module is used for: In response to the local autonomous decision-making mode instruction; Based on the aforementioned spatiotemporal instability index, the autonomous guided vehicle's own navigation power is adjusted to perform local predictive risk avoidance. Upon receiving a path instruction from the global collaborative optimization module, navigation is performed based on the path instruction.

Images