Agv scheduling and station cooperative control method for battery module assembly line

Abstract

This application relates to the technical field of automated logistics scheduling, and in particular to a method for AGV scheduling and workstation collaborative control in battery module assembly lines. The method includes the following steps: configuring workstation type labels for each workstation on the assembly line; calculating the dynamic priority of each workstation based on the workstation type labels; predefining AGVs into multiple resource types, with each resource type having a predefined matching degree relationship with the workstation type labels; and performing hierarchical scheduling decisions based on dynamic priority: for workstations with a dynamic priority higher than a first set threshold, selecting AGVs that fully match the workstation type labels, and further selecting AGVs in a schedulable state for allocation; for workstations with a dynamic priority lower than a second set threshold, selecting AGVs that partially match the workstation type labels for allocation. This application can achieve precise dynamic scheduling of material handling in mixed-line battery assembly by matching workstation production characteristics with AGV capability types.

Description

Technical Field

[0001] This application relates to the technical field of automated logistics scheduling, and in particular to a method for AGV scheduling and workstation collaborative control in battery module assembly lines. Background Technology

[0002] As a key component in vehicles, consumer electronics, and other fields, the assembly efficiency and quality of battery modules directly impact the overall production pace and reliability of products. With the diversification of market demands, battery module assembly lines are evolving towards flexibility and "mixed-line production," where multiple types of products need to be assembled alternately or in parallel on the same production line. For example, simultaneously processing heavy, slow-assembly battery modules for electric vehicles and lightweight, fast-assembly battery modules for portable electronic devices. In such complex scenarios, Automated Guided Vehicles (AGVs) are responsible for material delivery between warehouses and assembly stations; their intelligent and collaborative scheduling directly determines the overall throughput and stability of the production line.

[0003] Currently, numerous general methods have been proposed and applied for scheduling multi-AGV systems. The core logic of these methods is typically based on task queue management. They generate task request priorities by sensing the material demand status of each workstation and allocate tasks according to rules such as nearest-distance priority and first-to-idle priority, based on information such as the real-time location and idle status of the AGVs. Meanwhile, existing research largely focuses on global path optimization, deadlock avoidance, and multi-vehicle cooperative obstacle avoidance, aiming to reduce the AGV's empty-running rate and shorten the average task response time at the system level.

[0004] However, in mixed-line assembly scenarios with significantly differentiated production needs, existing scheduling systems fail to align their workstation task priority assessments with the inherent production rhythm characteristics and the physical capabilities of the AGVs. This leads to a disconnect between priority ranking and actual production logic, resulting in dynamic allocation misalignment and resource competition conflicts. Specifically, in the typical scenario where electric vehicle batteries and electronic device batteries co-produce on the same line, existing methods only prioritize workstations based on emergency signals. If an electronic device battery workstation is marked as high priority due to order backlog, the system might dispatch a heavy-duty AGV with strong load capacity but a design speed of only 0.5 m / s to replenish the battery, based solely on this label and the nearest-distance principle. Although this AGV can complete the transport operation, its slow speed is completely incompatible with the fast-paced inherent needs of the electronic device workstation. As a result, the workstation is forced to stagnate while waiting for the slow-speed AGV, with waiting time increasing dramatically from approximately 1 minute to over 5 minutes, creating the paradox that "the more urgent the system determines the situation, the longer the actual production wait becomes," severely hindering the productivity of fast-paced workstations. Conversely, if a light-load, fast-moving AGV is mistakenly assigned to a heavy-load workstation for battery carts, it may require multiple round trips due to insufficient single-trip carrying capacity, or pose stability risks when handling heavy objects, thus reducing actual efficiency. Furthermore, because the scheduling logic does not deeply bind workstation type and AGV type, the system cannot effectively predict and handle the effects caused by dynamic changes in priority. For example, when an electronic device battery workstation suddenly receives an urgent order, its priority spikes instantly. Multiple light-load AGVs may simultaneously respond to this single high-intensity demand signal, converging on the same target point, easily causing congestion at path intersections. AGVs slow down and stop to avoid each other, not only increasing additional traffic waiting time, but also, in extreme cases, the sudden stops and starts at close range may endanger the safety of the battery modules being transported, leading to an increase in product defect rates. This problem, caused by the mismatch between scheduling decisions and the deep characteristics of the physical production scenario, stems from the fact that the existing technology's assessment of "priority" remains at the superficial task queue level, failing to delve into and integrate the production rhythm of specific workstations, the physical characteristics of materials, and the core capability requirements of the executing equipment. This results in a disconnect between decision-making and execution, becoming the core bottleneck restricting the overall efficiency improvement of mixed-line assembly lines. Summary of the Invention

[0005] In order to achieve precise dynamic scheduling of material handling in mixed-line battery assembly by matching the production characteristics of workstations with the capabilities of AGVs, this application provides a method for AGV scheduling and workstation collaborative control in battery module assembly lines.

[0006] The AGV scheduling and workstation collaborative control method for a battery module assembly line provided in this application adopts the following technical solution: A method for AGV scheduling and workstation collaborative control for a battery module assembly line includes the following steps: Configure workstation type labels for each workstation on the assembly line, wherein the workstation type labels are defined according to the inherent production attributes of the battery modules assembled at the workstation; Based on the workstation type label, calculate the dynamic priority of each workstation; AGVs are predefined into multiple resource types, and each resource type has a predefined matching degree relationship with the workstation type label; Perform hierarchical scheduling decisions based on dynamic priorities: For workstations with dynamic priority higher than the first set threshold, AGVs that perfectly match the workstation type label are selected based on the matching degree relationship, and AGVs in a schedulable state are further selected from them for allocation. For workstations with dynamic priority lower than the second set threshold, AGVs that partially match the workstation type label are selected and allocated based on the matching degree relationship, and adaptive adjustments to the AGV operating parameters are triggered based on this partial matching relationship. For workstations with dynamic priorities between the first and second set thresholds, a standard scheduling decision is executed. The first set threshold is greater than the second set threshold.

[0007] Optionally, the calculation of the dynamic priority specifically includes: Based on the workstation type label, obtain the process urgency factor that represents the urgency of the current task; Obtain the workstation type fit factor, which characterizes the degree of fit between the requirements defined by the workstation type label and the AGV resource type; Obtain the historical waiting time factor based on the historical material replenishment performance of this workstation; By assigning a higher calculation weight to the workstation type suitability factor than to the process urgency factor, the process urgency factor, workstation type suitability factor, and historical waiting time factor are weighted and fused to generate the dynamic priority.

[0008] Optionally, the multiple resource types include a first resource type and a second resource type; The first resource type has a load capacity of not less than the first load threshold and its operating speed does not exceed the first speed threshold, and is associated with the workstation type label with heavy-load production requirements; The second resource type has a load capacity not exceeding the second load threshold and its operating speed is not lower than the second speed threshold, and is associated with a workstation type tag that has light-load, fast-paced production requirements.

[0009] Optionally, the predefined matching degree relationship is specifically as follows: The first resource type and the workstation type tag with the heavy production demand are at a perfect match level; The second resource type and the workstation type tag with the light-load, fast-paced production requirements are a perfect match. Among them, the first resource type and the workstation type label with the light-load fast-paced production needs, and the second resource type and the workstation type label with the heavy-load production needs, are both partially matched.

[0010] Optionally, the allocation of workstations with dynamic priorities higher than a first set threshold is achieved through the following steps: Perform matching locking, and based on the predefined matching degree relationship, filter out the set of AGVs whose resource type completely matches the current workstation type label from all AGVs; The matching and locked AGV set is filtered for status. Based on the real-time acquired task status information, only AGVs that are in an immediate responsive state or whose current task is expected to be completed in a time shorter than a first time threshold are retained. For AGVs that pass the status filter, the best dispatch is performed. Based on their spatial location information, the nearest AGV is selected for allocation. When the distance is the same, the dispatch decision is further made based on the logical rules of their energy status and the reliability of their task history.

[0011] Optionally, when it is detected that the dynamic priority of a workstation increases and exceeds the priority mutation threshold within a short period of time, a conflict resolution strategy is executed, the conflict resolution strategy including: Based on the predefined matching degree relationship, multiple candidate AGVs with a matching degree not lower than the pre-screening threshold and an expected available time shorter than the second time threshold are pre-screened from all AGVs; The candidate AGVs are initially sorted based on their spatial distance from the workstation. For candidate AGVs with the same spatial distance, a secondary sort is performed based on their matching degree with the current workstation type label; Based on the results of the preliminary and secondary sorting, a unique AGV is finally dispatched.

[0012] Optionally, the adaptive adjustment of the AGV operating parameters triggered based on this partial matching relationship specifically includes: When assigning a partially matched AGV to a workstation type tag with the aforementioned heavy-duty production requirements, its operating speed is reduced to a speed value suitable for heavy-duty production requirements, and the maximum amount of materials it can transport in a single trip is simultaneously limited.

[0013] Optionally, during the execution of the hierarchical scheduling decision based on dynamic priority, if it is necessary to allocate an AGV to a workstation with a dynamic priority higher than the first set threshold, thus occupying an AGV that is currently executing a task, then a task transfer process is triggered: Based on the real-time location information of the occupied AGV, it is determined and guided to the nearest idle material transfer station to unload and temporarily store the materials it is currently carrying; The task interrupted by the occupancy event will be reinserted into the task scheduling sequence based on the dynamic priority assigned to it before the interruption.

[0014] Optionally, it may also include system self-optimization steps, which include: Periodically collect and update several key performance indicators that reflect the system's scheduling efficiency; When any of the key performance indicators exceeds its preset improvement trigger threshold, the weight parameters associated with the dynamic priority calculation or the decision threshold associated with the hierarchical scheduling decision are dynamically adjusted according to the preset adjustment mapping relationship. Within each verification cycle after adjustment, the improvement effect of the key performance indicators is evaluated, and a decision is made based on the evaluation results as to whether to solidify the adjusted parameters or rules.

[0015] In summary, this application includes the following beneficial technical effects: 1. This application establishes a matching relationship between workstations and AGVs by configuring type tags that are strongly related to production attributes and predefining corresponding resource types for AGVs. Combined with dynamic priority calculation based on workstation adaptability as the core weight, it ultimately executes hierarchical scheduling decisions based on this priority. This fundamentally solves the resource mismatch and efficiency paradox problem in existing technologies, which is caused by the deep disconnect between scheduling decisions and the inherent production rhythm, material characteristics, and required core capabilities of AGVs at workstations. The goal is to achieve accurate dynamic matching between material handling needs and AGV execution capabilities in battery mixed-line assembly scenarios.

[0016] 2. This application designs a progressive allocation logic that includes matching locking, state filtering, and optimal dispatch. It also introduces a candidate AGV pre-screening and dual sorting strategy based on matching degree and expected available time for priority change scenarios. This effectively avoids path intersection congestion and potential safety risks caused by multiple AGVs converging disorderly in response to the same high emergency signal, and significantly improves the scheduling stability and traffic flow of the system in a high-dynamic, high-load production environment.

[0017] 3. This application distinguishes between fully matched and partially matched scenarios, and when allocating partially matched AGVs to low-priority workstations, it synchronously triggers adaptive adjustments to parameters such as AGV running speed and single transport volume. This allows idle or suboptimal AGV resources to be fully utilized for replenishment without affecting high-priority tasks, thereby optimizing the overall resource utilization rate of the system and reducing equipment idle rate while ensuring operational safety.

[0018] 4. This application establishes a system self-optimization process. By periodically collecting key performance indicators such as workstation response efficiency, AGV matching and adaptability, and conflict and interruption rates, and dynamically adjusting priority calculation weights, matching degree relationships, or decision thresholds when indicators are abnormal, the entire scheduling system can continuously perceive changes in production status and perform self-calibration. This ensures that the scheduling strategy and real-time production needs are optimally matched in the long term, effectively overcoming the problem of performance degradation of fixed rule systems as production conditions change. Attached Figure Description

[0019] Figure 1 This is the overall scheduling logic diagram; Figure 2 It is a logic diagram of the conflict resolution and task transfer process; Figure 3 It is a logic diagram of the system's self-optimization closed loop. Detailed Implementation

[0020] The following is in conjunction with the appendix Figure 1-3 This application will be described in further detail.

[0021] This application discloses an AGV scheduling and workstation collaborative control method for battery module assembly lines. For example... Figure 1 As shown, the specific implementation method of AGV scheduling and workstation collaborative control in the battery module assembly line includes the following steps: S1 workstation type label configuration and dynamic priority calculation S1 establishes a core reference system for accurate matching of AGV resources by standardizing the definition of workstation attributes and quantifying multi-dimensional priorities.

[0022] S11 workstation type label configuration The system first assigns workstation type labels to each workstation on the assembly line. These labels are defined entirely based on the inherent production attributes of the battery modules being assembled at each workstation. For the typical application scenario of mixed-line assembly of battery modules, the system sets two types of workstation type labels. The specific attributes and setting basis of these two types of labels are as follows: The first category is labeled "Heavy-duty, slow-paced workstation." This label corresponds to workstations that assemble electric vehicle battery modules. Their inherent production attributes are that a single battery pack weighs no less than 30kg, and only 10 battery packs can be assembled per hour. These parameters are set based on the actual physical characteristics and assembly process requirements of the electric vehicle battery modules. 30kg is the lower limit of the standard single-pack weight for electric vehicle batteries, and 10 packs per hour is the standard assembly efficiency for this type of battery module. The second category is the "Lightweight, Fast-Paced Workstation" label. This label corresponds to workstations that assemble battery modules for electronic devices. Their inherent production attributes are that a single battery pack weighs no more than 5kg, and they can complete the assembly of 30 battery packs per hour. These parameters are set based on the lightweight characteristics of portable electronic device batteries and the need for rapid turnaround assembly. 5kg is the upper limit for the standard weight of a single battery pack, and 30 packs per hour is their standard assembly efficiency.

[0023] The system completes the workstation type label configuration in this way, which can establish a direct link between workstation production attributes and subsequent scheduling logic. This effectively avoids the AGV allocation misalignment problem caused by the lack of standardized definitions of workstation attributes in existing technologies, and at the same time provides a stable and unified basic reference for the accurate calculation of subsequent dynamic priorities.

[0024] S12 Dynamic Priority Three-Dimensional Weighted Calculation The system calculates the dynamic priority of each workstation based on the configured workstation type labels. This process requires the collection of three types of quantitative factors first, and then the final priority is generated through weighted fusion.

[0025] S121 Collection of Three Types of Quantitative Factors The system collects process urgency factor, workstation type suitability factor, and historical waiting time factor for each workstation. All three factors are uniformly quantified within a range of 0 to 10. The collection rules and reasons for each factor are as follows: Process urgency factor collection: The system quantifies factors based on workstation type tags and the urgency of the current production task at each workstation. Critical processes are assigned 10 points, non-critical processes with a backlog of at least 5 batteries are assigned 8 points, and non-critical processes with a backlog of less than 5 batteries are assigned 5 points. This rule is based on the fact that critical processes directly determine the overall production line's flow rhythm and therefore require the highest urgency weight. The backlog of batteries directly reflects the urgency of material replenishment at each workstation, allowing for accurate characterization of the actual urgency of the task.

[0026] Workstation type compatibility factor collection: The system quantifies the factor based on the preset matching relationship between workstation type tags and AGV resource types. A complete match between workstation type and AGV type is assigned 10 points, a partial match 6 points (corresponding to a 60%-70% matching range), and a complete mismatch 0 points. This rule is set because existing scheduling technologies generally ignore the compatibility between workstations and AGVs. This factor can proactively avoid the misallocation problem of heavy-load AGVs being matched with fast-paced workstations, providing a core reference for priority calculation based on the compatibility dimension.

[0027] Historical waiting time factor collection: The system retrieves material replenishment waiting data for the past 7 days from each workstation to complete factor quantification. An average waiting time of at least 10 minutes is assigned 10 points, 5 to 10 minutes is assigned 8 points, and less than 5 minutes is assigned 5 points. This rule is set because historical waiting time objectively reflects past scheduling shortcomings at each workstation. Using this as a basis, a compensation dimension can be provided for priority calculation, avoiding production line scheduling imbalances caused by the continuous tilting of scheduling resources towards a single dimension.

[0028] S122 Factor Weighted Fusion and Priority Hierarchy Classification The system performs weighted fusion of the three types of factors collected to generate a dynamic priority total score for the workstation, and then determines the priority level based on the total score. The specific operation is as follows: Weighting and Factor Fusion: The system assigns differentiated weights to three types of factors: workstation type suitability factor (1.25), process urgency factor (1.0), and historical waiting time factor (1.3). The core principle of this weighting is that existing technologies prioritize task urgency, while this method prioritizes workstation and AGV suitability while compensating workstations with long historical waiting times with scheduling resources, thus avoiding a disconnect between scheduling decisions and actual production needs. The dynamic priority score is calculated using a linear weighting method. Specifically, the dynamic priority score equals the process urgency score multiplied by 1.0, plus the workstation type suitability score multiplied by 1.25, plus the historical waiting time score multiplied by 1.3. Through this weighted calculation, the theoretical range of the dynamic priority score is 0 to 35.5. Based on verification results from a large amount of battery module assembly line scheduling data, a first threshold of 28 points and a second threshold of 20 points are set. This threshold range can accurately distinguish the actual scheduling priority of workstations, avoiding resource mismatch caused by ambiguous hierarchical division.

[0029] Priority Hierarchy: The system defines three priority levels based on a dynamic total priority score. Workstations with a total score of 28 or higher are classified as Core Emergency Level; workstations with a total score of 20 or higher but less than 28 are classified as Important Regular Level; and workstations with a total score less than 20 are classified as General Pending Level. The Core Emergency Level corresponds to high-priority supply needs, the Important Regular Level corresponds to regular supply needs, and the General Pending Level corresponds to low-priority supply needs. This hierarchical classification enables precise targeted allocation of scheduling resources, providing a clear and explicit priority basis for subsequent hierarchical scheduling decisions.

[0030] S2 resource type definition, matching relationship establishment and status acquisition Based on the two workstation type labels of heavy-load slow pace and light-load fast pace already configured in S1, S2 completes the targeted classification of AGV resources, solidifies matching rules, and realizes real-time status perception, providing a standardized resource basis for the subsequent accurate matching of AGVs and workstations.

[0031] S21 Resource Type Classification Definition Based on the differences in production needs between the two types of workstations in S1, the system predefines AGVs into two resource types. The core indicators for classification are clearly defined as rated load capacity and operating speed. The parameter settings and basis for the two resource types are as follows: The system defines the first resource type as a heavy-duty AGV. The system sets the rated load capacity of this type of AGV to be no less than 50kg, which is set as the first load threshold. This threshold is based on the fact that the weight of a single battery pack in the heavy-duty, slow-paced workstation S1 is no less than 30kg. A load capacity of 50kg not only ensures the transport of at least one battery module per trip but also allows for a 20kg safety margin, effectively preventing mechanical damage caused by the load approaching the equipment's upper limit during heavy-duty transport. The system also sets the operating speed of this type of AGV to be no more than 0.6m / s, which is set as the first speed threshold. This threshold is based on the weight characteristics of the battery modules. Low-speed operation significantly reduces inertial impact during AGV start-up and shutdown. Actual measurements show that material swaying can be controlled within 2mm, effectively ensuring the structural integrity of the battery modules.

[0032] The system defines the second resource type as a light-load AGV. The system sets the rated load capacity of this type of AGV to 30kg, which is designated as the second load threshold. This threshold is based on the fact that the weight of a single battery pack in the light-load, fast-paced workstation S1 does not exceed 5kg. A load capacity of 30kg can support the transport of multiple electronic device battery modules in a single trip, which is highly compatible with the assembly efficiency of 30 packs per hour for this type of workstation, effectively reducing the number of AGV round trips to improve replenishment efficiency. The system also sets the operating speed of this type of AGV to be no less than 0.8m / s, which is designated as the second speed threshold. This threshold is based on the high-frequency replenishment requirements of the light-load, fast-paced workstation. Production line testing shows that an operating speed of 0.8m / s can stably control the single replenishment response time to within 1 minute, perfectly adapting to the fast-paced production characteristics of this type of workstation.

[0033] This classification method establishes a clear demand correspondence between AGV resource types and S1 workstation types, effectively solving the matching chaos problem caused by the ambiguity of AGV resource types in existing technologies, and providing a clear resource benchmark for subsequent scheduling decisions based on workstation priority.

[0034] S22 Predefined matching degree relationship solidification The system establishes and solidifies the matching degree relationship between the workstation type label of S1 and the AGV resource type of S21. The matching degree is quantified in the form of a percentage. The specific rules and reasons for setting them are as follows: The matching degree between heavy-duty AGVs and heavy-duty slow-paced workstation type tags is 100%, defined as a perfect match. This is because the two are perfectly matched in terms of load requirements and operating rhythm, achieving optimal replenishment efficiency after scheduling without requiring additional adjustments to AGV operating parameters. Similarly, the matching degree between light-duty AGVs and light-duty fast-paced workstation type tags is 100%, also defined as a perfect match. This is because the high-speed characteristics and light-duty capabilities of light-duty AGVs accurately match the high-frequency replenishment needs of light-duty fast-paced workstations, ensuring uninterrupted continuous production.

[0035] The matching degree between heavy-duty AGVs and light-duty fast-paced workstation type labels is 60%, defined as a partial matching level. The matching degree between light-duty AGVs and heavy-duty slow-paced workstation type labels is 70%, also defined as a partial matching level. The percentage value of the partial matching level is set based on a comprehensive assessment of the parameter adjustment costs, efficiency losses, and safety redundancy required when adapting different types of AGVs to unconventional workstations.

[0036] The solidification of this matching degree relationship allows the calculation results of the dynamic priority of the workstation in S1 to be quickly mapped to specific AGV resources, avoiding repeated determination of the matching relationship during the scheduling process. At the same time, the quantified matching degree provides a clear basis for parameter adjustment in subsequent matching scenarios, significantly improving the efficiency of scheduling decisions.

[0037] Real-time data acquisition of S23AGV status The system integrates UWB positioning technology with the Manufacturing Execution System (MES) data interface to collect real-time status information of three types of AGVs, providing dynamic data support for subsequent scheduling decisions based on S1 workstation priority. The specific collection rules and basis are as follows: The system collects the position status of AGVs. It employs UWB positioning technology for real-time AGV positioning, with a positioning accuracy strictly controlled within 0.1m. This 0.1m error range accurately distinguishes the specific stopping position of the AGV next to the production line, effectively preventing the failure of the nearest dispatch rule due to positioning deviation, and ensuring that the nearest AGV resource is matched to the high-priority workstations in S1. Positioning data is updated every 100ms; this update frequency ensures the real-time nature of the position information, keeping the deviation between the AGV position obtained by the scheduling system and the actual position within a controllable range.

[0038] The system collects the task status of AGVs. The system clearly categorizes AGV task status into three types: idle, in transit, and charging. For AGVs in transit, if the estimated completion time of their current task is no more than 30 seconds, the system marks them as soon to be idle. For AGVs charging, if their remaining charging time is no more than 5 minutes, the system marks them as soon to be available. These 30-second and 5-minute time thresholds are set based on the maximum tolerable waiting time of the core emergency workstations in S1. This setting maximizes the utilization of AGV resources, avoids excluding AGVs that are about to become available from the scheduling pool, and effectively improves resource utilization.

[0039] The system collects the AGV's power status. The system sets an AGV available power threshold of 30%. AGVs with power levels above 30% are marked as available, while those below 30% are forcibly guided to a designated charging area. This threshold is based on the minimum power requirement for an AGV to complete a single full resupply task. Actual testing shows that 30% power is sufficient to support an AGV making one round trip between the furthest workstation on the production line and the material warehouse, with an additional 5% reserve for emergencies. This effectively prevents task interruptions due to power outages and ensures the continuity of resupply tasks to high-priority workstations in S1.

[0040] By collecting data in real time from three types of states, the system constructs a dynamic profile of AGV resources, enabling subsequent scheduling decisions to combine the workstation priority of S1 with the real-time status of the AGV, effectively avoiding the scheduling inaccuracies caused by relying solely on static resource information in existing technologies.

[0041] S3 hierarchical scheduling decision based on dynamic priority like Figure 1As shown, S3 uses the dynamic priority level of the workstation output by S1 as the decision-making basis, and combines the AGV matching degree relationship fixed by S2 with real-time status data to execute differentiated AGV allocation logic. This step, through a hierarchical strategy of "high priority for re-adaptation, medium priority for balance, and low priority for compatibility", transforms the workstation priority attributes of S1 and the AGV resource attributes of S2 into dynamic scheduling instructions, effectively improving the accuracy of resource matching and the efficiency of production line response.

[0042] Precise AGV allocation at the core emergency workstation of S31 For core emergency workstations with dynamic priority higher than the first set threshold (i.e., a total dynamic priority score of no less than 28 points), the system executes a three-level allocation logic of "type locking - status filtering - optimal dispatch" to ensure that high-priority needs are guaranteed in terms of both adaptability and timeliness.

[0043] S311 Type Locking The system first performs a type locking operation. It calls the established matching relationships in S22 to filter out a set of AGVs from the entire AGV resource pool that match the current workstation's type label by 100%. If the current workstation is a heavy-load, slow-paced workstation as defined in S11, the system only locks the heavy-load AGVs defined in S21; if the current workstation is a light-load, fast-paced workstation as defined in S11, the system only locks the light-load AGVs defined in S21. This operation directly excludes partially matching and completely mismatched AGV types, avoiding the mismatch issues mentioned in S2, such as "heavy-load AGVs matching fast-paced workstations," from the source. This ensures a rigid fit between AGV resource types and workstation production needs, laying the foundation for efficient replenishment later.

[0044] S312 Status Filtering The system performs status filtering on the locked, fully matched AGV set. The system retrieves real-time AGV status data collected by S23, retaining only two types of AGVs: one type is AGVs whose task status is idle and whose battery level is not lower than 30%, allowing for immediate response; the other type is AGVs whose task status is in transit but whose current task's estimated completion time is no more than 30 seconds. The 30-second time limit is set as the first time threshold. This threshold is based on the maximum tolerable waiting time for core emergency workstations in S1. Production line testing shows that if the waiting time for core emergency workstations exceeds 30 seconds, material supply disruptions are highly likely. The 30-second time limit ensures that AGVs can quickly switch to new tasks, preventing production stoppages at workstations.

[0045] S313 Selective Dispatch The system performs optimal dispatch for AGVs that pass the status filtering. The system prioritizes calling the AGV position status data collected by S23, selecting the AGV closest to the current workstation, and issuing a replenishment task instruction to it. When multiple AGVs are at the same distance, the system further introduces a dual judgment rule: the first priority is to determine the battery status, selecting AGVs with a battery level above 50%. This value is set based on the fact that core emergency tasks often involve continuous replenishment needs, and high battery levels can prevent AGVs from returning to recharge midway. The second priority is to determine the historical reliability of the task, selecting AGVs with a task completion rate of no less than 98% in the past 7 days. This value is obtained by statistically analyzing the historical task records of AGVs collected by S23, which can reduce the risk of task interruption due to equipment failure. This multi-dimensional optimization method ensures both the timeliness of replenishment to core emergency workstations and improves the stability of task execution.

[0046] S32 Important Regular Workstation Routine Scheduling Decisions For important regular workstations with dynamic priorities between the first and second set thresholds (i.e., a total score of not less than 20 points and less than 28 points), the system executes a routine scheduling decision that prioritizes adaptation while also considering efficiency, balancing resource utilization and production line balance. The specific operation is as follows: The system first attempts to allocate perfectly matched AGV resources to this type of workstation. The allocation logic is consistent with the perfectly matched AGV screening, status filtering, and optimal dispatch logic in S311 to S313, but the first time threshold in the status filtering step is relaxed to 60 seconds. The reason for this threshold adjustment is that there is no immediate risk of supply interruption for important regular workstations. Appropriately relaxing the waiting time can avoid frequent AGV resource grabbing, ensure the scheduling priority of core emergency workstations, and balance response speed and resource utilization efficiency.

[0047] When there are insufficient fully matched AGV resources, the system filters partially matched AGVs with a matching degree of not less than 60%. At this time, there is no need to trigger the adjustment of AGV running parameters. The rule is set based on the fact that the production rhythm of this type of workstation is relatively slow, and the performance difference between partially matched AGVs and workstations can be absorbed by the workstation's own production buffer capacity, without the need to consume additional system computing power for parameter optimization.

[0048] The conventional scheduling decision, through the logic of prioritizing full matching and supplementing partial matching, meets the supply needs of important conventional workstations while avoiding excessive resource allocation to a single workstation, ensuring the scheduling balance of the entire production line and solving the resource waste problem in existing technologies.

[0049] S33 General Pending Workstation Compatibility Allocation and Parameter Adjustment For general workstations with a dynamic priority lower than the second set threshold (i.e., a total score lower than 20 points), the system executes a compatible matching and parameter adaptation allocation logic to ensure work safety while utilizing idle resources. The specific process is as follows: Based on the matching degree relationship of S22, the system prioritizes filtering AGV resources with a partial matching degree—heavy-load, slow-paced workstations prioritize light-load AGVs, while light-load, fast-paced workstations prioritize heavy-load AGVs. The core purpose of this filtering logic is to utilize the replenishment buffer time of low-priority workstations to utilize idle resources other than fully matched AGVs, thereby improving the overall AGV utilization rate of the system and avoiding empty runs of equipment.

[0050] The system triggers AGV operating parameter adjustments based on partial matching relationships. The adjustment rules and numerical basis for different matching scenarios are as follows: When a light-load AGV is assigned to a heavy-load, slow-paced workstation, the system reduces the operating speed of the light-load AGV from 0.8m / s to 0.5m / s, while limiting the maximum number of materials it can transport in a single trip to 2 sets. The speed value of 0.5m / s is based on the weight characteristics of the battery module of the electric vehicle. This speed can control the swaying amplitude during material transportation to within 1.5mm, ensuring the stability of heavy-load materials. The maximum number of 2 sets is calculated based on the rated load threshold of 30kg for the light-load AGV and the verified short-term overload limit under specific working conditions that are clearly permitted by the equipment manufacturer and closely monitored by the system. This allows for emergency transportation while ensuring operational safety.

[0051] When heavy-duty AGVs are assigned to light-load, fast-paced workstations, the system increases the operating speed of the heavy-duty AGVs from 0.6 m / s to 0.7 m / s, with a maximum of 12 groups transported per trip. The speed of 0.7 m / s is higher than the conventional speed of heavy-duty AGVs but does not exceed the mechanical structural tolerance, thus shortening the waiting time at fast-paced workstations. The maximum of 12 groups, calculated based on the 5 kg weight of a single light-duty battery as defined in S11, results in a total weight of 60 kg, which is within the minimum 50 kg load capacity of the heavy-duty AGVs, reducing the number of AGV round trips. Through parameter adjustments, some matched AGVs can adapt to the needs of low-priority workstations, solving the inefficiency and safety hazards caused by resource mismatch in existing technologies.

[0052] S4 priority mutation conflict resolution strategy like Figure 2 As shown, S4 is specifically designed to address scenarios where workstation priorities change rapidly during the S3 scheduling process. When a workstation priority rapidly jumps from a low level to a core emergency level, it can easily lead to resource contention and path conflicts caused by multiple AGVs responding simultaneously. S4 uses a progressive strategy to ensure timely response to high-priority demands while avoiding the scheduling chaos caused by the disorderly convergence of multiple devices in existing technologies.

[0053] S41 Priority Mutation Determination and Strategy Triggering The system first performs a priority mutation judgment operation to clarify the triggering conditions for the conflict resolution strategy. The system monitors the dynamic priority calculation results of each workstation in real time. When the priority jumps within a short period of time and crosses two levels, the conflict resolution strategy is immediately triggered. The specific judgment criteria are: the total dynamic priority score of the workstation jumps from less than 20 points to no less than 28 points within 10 seconds. Here, 20 points is the second set threshold in S1 that defines the general pending tasks and important routine levels, 28 points is the first set threshold that defines the important routine and core urgent levels, and 10 seconds is defined as the priority mutation time threshold.

[0054] The criteria for this decision are based on the following: the 10-second time limit effectively distinguishes between genuine priority shifts and data calculation fluctuations, preventing erroneous policy triggering due to system data update delays; the requirement to cross two levels clarifies the urgency of the shift, indicating a sudden and urgent production need at the workstation, necessitating a break from the regular scheduling queue's priority response. This difference in levels is the core cause of disordered responses from multiple AGVs, thus requiring a specific approach. This decision logic makes policy triggering more precise, ensuring no urgent needs are overlooked while avoiding ineffective consumption of scheduling resources.

[0055] Targeted pre-screening of candidate AGVs S42 After the system triggers the strategy, it immediately performs a pre-screening operation for candidate AGVs, narrowing down the candidate range through dual conditions to reduce the probability of conflict from the source. Based on the matching degree relationship fixed in S22, the system filters candidate resources that meet the conditions from all AGVs. The first screening condition is that the matching degree between the AGV and the current workstation type label is not less than 90%, and this value is set as the pre-screening threshold; the second screening condition is that the expected available time of the AGV is not more than 2 minutes, and this value is set as the second time threshold.

[0056] The reason for setting the pre-screening threshold of 90% is that the complete match level defined in S22 is 100%, and the minimum partial match level is 60%. The 90% threshold retains the high adaptability of near-complete matching, while moderately expanding the candidate range compared to the 100% complete match condition, avoiding the situation where no resources are available due to all fully matched AGVs being occupied. The second time threshold of 2 minutes is set based on the maximum tolerable waiting time of the core emergency workstations in S3. The workstations after the mutation have entered the core emergency level, and the 2-minute waiting limit can ensure the timeliness of replenishment, while reserving sufficient computing time for the system's subsequent sorting decisions.

[0057] During the screening process, the system prioritizes retrieving real-time status data of the AGVs collected by S23. AGVs that are charging but have no more than 2 minutes of remaining charging time, or those that are in transit but have an estimated completion time of no more than 2 minutes, are included in the candidate pool. This flexible screening rule ensures the immediate availability of candidate AGVs while maximizing the exploitation of potential resources, solving the problem of insufficient resource supply caused by only screening idle AGVs in existing technologies.

[0058] Dual sorting of candidate AGVs S43 The system performs a dual sorting of pre-screened candidate AGVs, determining the priority order through a rule that prioritizes timeliness and secondarily adaptability, ensuring that dispatch decisions are both fast and accurate. The sorting process relies entirely on the matching relationship between real-time data collected by S23 and S22, without introducing new data dimensions, thus improving decision-making efficiency.

[0059] The system first performs a preliminary sorting operation. It calls the AGV location status data obtained using UWB positioning technology in S23, calculates the actual passable path distance between each candidate AGV and the sudden change workstation, and sorts them from closest to furthest. The core purpose of this sorting rule is to ensure the timeliness of supply response; the closer the AGV is, the shorter the time it takes to reach the workstation, which can quickly alleviate the material shortage problem at the sudden change workstation, consistent with the core logic of dispatching the nearest AGV in S313.

[0060] The system performs a secondary sorting operation on candidate AGVs that are equidistant from the initial sorting. The system retrieves the fixed matching degree relationships from S22 and sorts the AGVs according to their matching degree with the current workstation type label, from highest to lowest. For example, among two candidate AGVs that are equidistant, the AGV with a 100% matching degree is prioritized over the AGV with a 92% matching degree. This sorting rule is set because, under the premise of the same timeliness, higher adaptability can reduce the cost of subsequent AGV operating parameter adjustments, improve the stability of replenishment operations, and avoid material losses due to insufficient adaptability.

[0061] The final dispatch of S44's only AGV Based on the combined results of preliminary and secondary sorting, the system determines the single AGV to be dispatched and issues it a precise task instruction containing the target workstation location, material type, and replenishment quantity. This task instruction is simultaneously synchronized to the MES system and the AGV scheduling system, achieving linkage between scheduling, production, and equipment execution information. The system also simultaneously sends a path-locking instruction to the AGV scheduling system to prevent other tasks from preempting planned paths. This final dispatch logic replaces the simultaneous response of multiple AGVs with a single AGV response, fundamentally solving the congestion problem caused by intersecting paths of multiple devices in existing technologies.

[0062] S5 Core Emergency Workstation Task Transfer Process like Figure 2 As shown, S5 specifically responds to the special scenario in S31 where AGVs are allocated to core emergency workstations with dynamic priorities higher than a first set threshold. The first set threshold is 28 points. When all fully matched AGVs are in task execution mode, and there are no AGVs about to become idle that meet the time threshold of S312, and simultaneously, there is an AGV currently executing a task whose current task priority is lower than the priority of the core emergency workstation, the system needs to occupy that AGV to prioritize meeting the core emergency needs. S5 achieves the dual goals of uninterrupted emergency supply and the ability to continue interrupted tasks through a standardized transfer process, solving the problems of material loss and process chaos caused by task preemption in existing technologies.

[0063] Positioning of S51 occupied AGV and selection of optimal transfer station The system first performs a positioning operation on the occupied AGV. The system calls upon the AGV position data collected using UWB positioning technology in S23 to accurately obtain the real-time coordinates of the occupied AGV, with positioning accuracy strictly controlled within 0.1m. This accuracy ensures that the coordinate error of subsequent path planning is within the allowable range, avoiding extended transfer time due to positioning deviations and ensuring that the waiting time for core emergency workstations is controllable.

[0064] The system performs optimal transfer station selection based on the positioning results. It retrieves the pre-set material transfer station layout data from the production line, calculates the actual passable distance from the current position of the occupied AGV to all available transfer stations, and selects the nearest available transfer station as the target transfer point. The system also sets a distance threshold of 50m. If the distance between the nearest available transfer station and the AGV exceeds 50m, the system immediately triggers a transfer station expansion prompt through both MES system pop-ups and audible / visual alarms, notifying on-site personnel to release the temporary storage area. The 50m threshold is set based on a matching calculation of the AGV's operating speed and the waiting limit of the core emergency workstation. For a heavy-duty AGV with a speed of 0.6m / s, a 50m journey takes approximately 83s. Adding the material unloading time, the total time can be controlled within 2 minutes, perfectly matching the maximum waiting time of the core emergency workstation in S42, ensuring that core emergency replenishment is not excessively delayed due to transfer links.

[0065] S52 transit route planning and material temporary storage execution The system plans dedicated transfer routes for occupied AGVs. The system generates routes that avoid the main production line aisles and high-traffic areas through the AGV scheduling system, with path planning response time strictly controlled within 2 seconds. This time limit prevents excessive waiting times at key emergency workstations due to lengthy path calculations, ensuring rapid activation of emergency procedures. Simultaneously, the system issues avoidance instructions to other AGVs along the path, specifying the target AGV's travel time and path range, ensuring that occupied AGVs reach the target transfer station at maximum safe speed.

[0066] The system guides the occupied AGV to complete the temporary storage of materials. After the AGV arrives at the transfer station, the system controls it to slowly unload the currently loaded materials according to a preset process. The infrared sensor at the transfer station detects the material placement status in real time, with a sensing accuracy of 0.5mm. This accuracy prevents material deviation from causing difficulties in subsequent material retrieval. Once the materials are in place, the sensor immediately sends a temporary storage completion signal to the MES system. The MES system binds a unique QR code identifier to this batch of materials, associating it with the workstation information and priority parameters corresponding to its original task, providing accurate data support for subsequent task continuation. If all transfer stations are found to be occupied during screening, the system activates a priority release mechanism, prioritizing the release of transfer stations storing general pending materials. The general pending material level corresponds to a priority standard below 20 points in S122. This rule ensures the priority of core urgent tasks while minimizing the impact on low-priority tasks.

[0067] Priority re-determination and scheduling sequence insertion for S53 interrupt tasks The system performs a priority reassessment operation on interrupted tasks. The system retrieves the original dynamic priority data of the workstation corresponding to the task before the interruption. If the priority had reached the core emergency level before the interruption, the system marks it with an emergency continuation flag; if it was at the important routine or general pending level, the original priority flag is retained. This reassessment logic ensures that high-priority interrupted tasks are not downgraded due to preemption, guaranteeing the stability of the overall production line rhythm and avoiding new scheduling imbalances caused by task preemption.

[0068] The system re-inserts interrupted tasks into the scheduling sequence. Based on the re-evaluated priority, the system inserts the interrupted task at the front of the corresponding scheduling queue. If it is an urgent follow-up task, its priority is second only to currently unexecuted core urgent tasks. The system also accurately records the task interruption time, material storage location, and identification information, generating a follow-up task order containing the material's unique identifier, original target workstation, and temporary storage transfer station number. When a matching AGV enters an idle state, the allocation process for this follow-up task is immediately triggered, reducing the average follow-up delay of interrupted tasks and improving task completion rate and production continuity.

[0069] S54 was taken over by an AGV for emergency task takeover. After the materials are temporarily stored, the system immediately clears the original task identifier of the occupied AGV, synchronously updates its status to "urgently available," and uploads it to the scheduling system. The system calls the real-time power detection data in S23. If the AGV's current power is below 40%, the system first controls it to stop at a fast charging station near the transfer station, completes a 30-second rapid power replenishment to raise the power to above 50%, and then issues a replenishment task for the core emergency workstation. If the power is sufficient, the AGV directly receives the task instruction and proceeds to the core emergency workstation according to the optimal path planning in S313. The 40% power threshold is calculated based on the energy consumption of a single emergency replenishment. The round trip distance from the transfer station to the farthest core emergency workstation is approximately 80m. The energy consumption for a heavy-duty AGV to complete this distance accounts for about 15% of the total power. The initial 40% power ensures that there is still sufficient power remaining after the task is completed, avoiding secondary interruptions to the task due to power outages. This design not only ensures the priority of meeting core emergency needs but also avoids process chaos caused by task preemption through standardized transfer procedures, achieving a balance between scheduling flexibility and production stability.

[0070] S6 System Self-Optimization Process like Figure 3 As shown, S6 focuses on the issues of task interruption rate, transfer station utilization, and AGV energy consumption exposed in the S5 task transfer process. Through a dynamic cycle of data collection, threshold triggering, parameter calibration, and verification and solidification, it continuously calibrates scheduling parameters and decision rules. This step transforms the emergency operation feedback from S5 into a basis for system parameter iteration, solving the problem of adaptability decay caused by the solidification of existing technical scheduling rules. At the same time, it strengthens the logical linkage with S3 hierarchical scheduling and S4 conflict resolution, enabling the entire scheduling system to dynamically adapt to changes in the production line.

[0071] Periodic collection of key performance indicators of S61 The system performs full-process key performance indicator (KPI) collection at a frequency of once every 5 minutes. This frequency is based on the production line cycle time; 5 minutes covers at least one complete AGV replenishment cycle (including material handling, transportation, and unloading), ensuring data real-time performance while avoiding excessive system computing power consumption from high-frequency collection. The collected indicators are divided into four categories, and the definitions, quantification methods, and collection criteria for each indicator are clearly defined as follows: The first category is the workstation response efficiency index, including the average waiting time for core emergency level, important routine level, and general pending level workstations. The average waiting time for core emergency level workstations is accurate to 0.1 minutes, which can capture subtle performance differences in the S313 optimal dispatch logic; the average waiting time for important routine level and general pending level workstations is accurate to 0.5 minutes, consistent with their production buffer characteristics. The system records the AGV task issuance time and material delivery confirmation time through the MES system, and the difference between the two is the waiting time. During data collection, the system is linked to the AGV status data of S23 to ensure that the waiting time calculation does not include abnormal time consumption such as AGV failures.

[0072] The second category is the AGV matching and adaptation index, which is the matching degree between the AGV type and the workstation type label. The matching degree is calculated as the ratio of the number of times a fully matching AGV is assigned to the total number of assignments, and the result is expressed as a percentage, accurate to 0.1%. This index directly reflects the collaborative efficiency of the S22 matching relationship solidification rules and the S3 hierarchical scheduling logic. When the matching degree is below 90%, it indicates that the S33 compatible assignment logic is over-enabled, and resource configuration optimization is needed to reduce the adaptation deviation.

[0073] The third category consists of conflict and interruption indicators, including the conflict rate in S4 priority mutation scenarios and the interruption rate in S5 task transfer scenarios. The conflict rate is the ratio of the number of times the conflict resolution strategy is triggered to the total number of times the core emergency-level workstations are required, and the interruption rate is the ratio of the number of times the task transfer process is triggered to the total number of times the core emergency-level workstations are required. Both are accurate to 0.1%. These indicators directly reflect the necessity and operational efficiency of S4 conflict resolution and S5 transfer processes, providing core data support for adjusting scheduling parameter thresholds.

[0074] The fourth category is resource utilization indicators, including the average idle time of heavy-duty AGVs and light-duty AGVs, the power consumption rate, and the turnover rate of the transfer station used in S5. AGV idle time is accurate to 1 minute, excluding idle periods for planned AGV charging during the statistics; the power consumption rate is in kWh / km, accurate to 0.01 kWh / km, and corresponds to the AGV load speed parameter in S21; the transfer station turnover rate is the number of times materials are temporarily stored and retrieved per unit time, accurate to 1 time / h, directly reflecting the efficiency level of the material storage process in S52.

[0075] S62 Optimized Threshold Determination and Dynamic Parameter Adjustment Based on historical production data and industry standards, the system sets improvement trigger thresholds for four types of indicators. When any indicator exceeds the threshold for three consecutive data collection cycles (15 minutes), parameter adjustment is immediately initiated. The specific adjustment rules, reasons for threshold settings, and related processes are as follows: If the average waiting time for core emergency workstations is ≥2 minutes, the system will prioritize adjusting the dynamic priority weight parameters of S122. The system will increase the workstation type suitability factor weight from 1.25 to 1.3, while keeping the process urgency factor weight unchanged at 1.0. This adjustment is based on the type locking logic of S311. Strengthening the suitability weight can pre-select more suitable AGV resources, reducing the time consumed in subsequent status filtering. The 2-minute threshold is set because the first time threshold of S312 is 30 seconds. Adding the AGV's movement time from the nearest warehouse to the farthest core emergency workstation (approximately 1.2 minutes), 2 minutes exceeds the maximum tolerance range for core emergency workstations, necessitating weight adjustment to improve the accuracy of priority calculation.

[0076] If the average waiting time at the light-load, fast-paced workstation is ≥3 minutes, the system adjusts some matching parameters of S22 and the operating parameters of S332. The system increases the matching degree between the heavy-load AGV and the light-load, fast-paced workstation from 60% to 65%, and simultaneously increases the operating speed of the heavy-load AGV serving the light-load workstation from 0.7 m / s to 0.75 m / s. The 3-minute threshold corresponds to an assembly efficiency of 30 sets per hour for the light-load workstation; exceeding this time will cause the material backlog to exceed the warning line of 5 boxes. Improving the matching degree increases the number of candidate AGVs, while adjusting the speed directly shortens the transportation time. Furthermore, 0.75 m / s does not exceed the mechanical structural tolerance limit of the heavy-load AGV. Actual measurements show that the AGV braking distance at this speed is still controlled within 0.35 m, meeting the production line safety requirements.

[0077] If the conflict rate at core emergency workstations is ≥5%, the system tightens the screening threshold in S42. The system adjusts the estimated available time threshold set in S42 from 2 minutes to 1.5 minutes, while increasing the pre-screening threshold from 90% to 95%. A 5% conflict rate means that AGV path congestion occurs once every 20 core emergency requests, affecting production line stability. Shortening the time threshold reduces the number of AGVs responding simultaneously, while increasing the pre-screening threshold ensures higher adaptability of candidate AGVs, reducing the probability of multi-device path intersections from the source.

[0078] If the task interruption rate of S5 is ≥3% or the turnover rate of the transfer station is ≤2 times / h, the system optimizes the relevant rules of the transfer station in S5. When the interruption rate exceeds the standard, the system issues an instruction to the field management module through the MES system to add a temporary transfer station. The temporary transfer station is located next to the auxiliary passage of the production line, no more than 30m away from the heavy-load slow-paced workstation, and its carrying capacity is the same as that of the fixed transfer station, which is 12 sets of light-load batteries or 2 sets of heavy-load batteries. When the turnover rate exceeds the standard, the system adjusts the transfer station release rule from "prioritize the release of general pending materials" to "prioritize the release of batches of general pending materials with a backlog time ≥10min". A 3% interruption rate corresponds to 3 task interruptions for every 100 core emergency requests, which does not meet the requirements of production continuity; the 10-minute backlog time threshold balances the storage needs of low-priority materials with the turnover efficiency of the transfer station, avoiding secondary scheduling waste caused by ineffective release.

[0079] Verification and parameter solidification of S63 optimization effect The system initiates the optimization effect verification process, setting the verification period to 7 days. The 7-day period is based on the production line's weekly production plan cycle, covering different load scenarios such as peak, stable, and off-peak orders, ensuring the verification results have universal applicability. During the verification process, the system monitors the target indicators in real time according to the S61 data collection rules and compares them with the baseline data from the 7 days prior to the adjustment.

[0080] The system employs a dual evaluation standard: if the improvement rate of the target indicator is ≥20% and does not cause other indicators to exceed the standard, the optimization is deemed effective; if the improvement rate is <10%, or if it leads to derivative problems such as increased waiting time at core emergency workstations or increased AGV failure frequency, the optimization is deemed ineffective. A 20% improvement rate is the industry standard for effective optimization of scheduling systems, effectively improving production line efficiency; the 10% lower limit prevents minor data fluctuations from being misjudged as effective optimization, ensuring the reliability of parameter adjustments.

[0081] After successful verification, the system will solidify the adjusted parameters into the scheduling core module, synchronously update the priority calculation weights of S122, the matching relationship table of S22, the conflict resolution screening threshold of S42, and the transit process rules of S5, and generate a parameter adjustment log. The log records the adjustment time, trigger indicator exceedance data, comparison of old and new parameters, and effect data, providing a basis for subsequent traceability and secondary optimization. If the verification is invalid, the system will immediately revert to the parameter configuration before the adjustment and trigger the root cause analysis process to investigate whether there are factors not included in the indicator association, such as AGV mechanical wear and tear or temporary changes in production line layout.

[0082] The implementation principle of the AGV scheduling and workstation collaborative control method for battery module assembly lines in this application is as follows: This application configures standardized type labels based on production attributes for workstations and establishes a matching degree relationship between them and predefined AGV resource types. It combines dynamic priority calculations based on multiple factors such as process urgency, workstation adaptability, and historical waiting time to execute hierarchical scheduling decisions: high-priority workstations are preferentially allocated fully matched AGVs to ensure adaptability and response speed; low-priority workstations are allowed partial matching and AGV operating parameter adaptive adjustments are triggered to accommodate idle resources. Simultaneously, conflict resolution, task transfer, and system self-optimization mechanisms are introduced to address dynamic changes and resource competition. This series of collaborative designs directly addresses problems in existing scheduling systems caused by the disconnect between workstation production characteristics and AGV capability attributes, such as "system-determined urgency but actual waiting time extended," resource mismatch, and frequent dynamic conflicts. It achieves precise dynamic scheduling of material handling in mixed-line battery assembly scenarios, thereby significantly improving the overall throughput, resource utilization, and production continuity of the production line.

[0083] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for AGV scheduling and workstation collaborative control in a battery module assembly line, characterized in that, Includes the following steps: Configure workstation type labels for each workstation on the assembly line, wherein the workstation type labels are defined according to the inherent production attributes of the battery modules assembled at the workstation; Based on the workstation type label, calculate the dynamic priority of each workstation; AGVs are predefined into multiple resource types, and each resource type has a predefined matching degree relationship with the workstation type label; Perform hierarchical scheduling decisions based on dynamic priorities: For workstations with dynamic priority higher than the first set threshold, AGVs that perfectly match the workstation type label are selected based on the matching degree relationship, and AGVs in a schedulable state are further selected from them for allocation. For workstations with dynamic priority lower than the second set threshold, AGVs that partially match the workstation type label are selected and allocated based on the matching degree relationship, and adaptive adjustments to the AGV operating parameters are triggered based on this partial matching relationship. For workstations with dynamic priorities between the first and second set thresholds, a standard scheduling decision is executed. The first set threshold is greater than the second set threshold.

2. The method according to claim 1, characterized in that, The calculation of the dynamic priority specifically includes: Based on the workstation type label, obtain the process urgency factor that represents the urgency of the current task; Obtain the workstation type fit factor, which characterizes the degree of fit between the requirements defined by the workstation type label and the AGV resource type; Obtain the historical waiting time factor based on the historical material replenishment performance of this workstation; By assigning a higher calculation weight to the workstation type suitability factor than to the process urgency factor, the process urgency factor, workstation type suitability factor, and historical waiting time factor are weighted and fused to generate the dynamic priority.

3. The method according to claim 1, characterized in that, The various resource types include a first resource type and a second resource type; The first resource type has a load capacity of not less than the first load threshold and its operating speed does not exceed the first speed threshold, and is associated with the workstation type label with heavy-load production requirements; The second resource type has a load capacity not exceeding the second load threshold and its operating speed is not lower than the second speed threshold, and is associated with a workstation type tag that has light-load, fast-paced production requirements.

4. The method according to claim 3, characterized in that, The predefined matching degree relationship is specifically as follows: The first resource type and the workstation type tag with the heavy production demand are at a perfect match level; The second resource type and the workstation type tag with the light-load, fast-paced production requirements are a perfect match. Among them, the first resource type and the workstation type label with the light-load fast-paced production needs, and the second resource type and the workstation type label with the heavy-load production needs, are both partially matched.

5. The method according to claim 1, characterized in that, The allocation of workstations with dynamic priorities higher than a first set threshold is achieved through the following steps: Perform matching locking, and based on the predefined matching degree relationship, filter out the set of AGVs whose resource type completely matches the current workstation type label from all AGVs; The matching and locked AGV set is filtered for status. Based on the real-time acquired task status information, only AGVs that are in an immediate responsive state or whose current task is expected to be completed in a time shorter than a first time threshold are retained. For AGVs that pass the status filter, the best dispatch is performed. Based on their spatial location information, the nearest AGV is selected for allocation. When the distance is the same, the dispatch decision is further made based on the logical rules of their energy status and the reliability of their task history.

6. The method according to claim 2, characterized in that, When the dynamic priority of a workstation is detected to increase rapidly and exceed the priority mutation threshold within a short period of time, a conflict resolution strategy is executed, which includes: Based on the predefined matching degree relationship, multiple candidate AGVs with a matching degree not lower than the pre-screening threshold and an expected available time shorter than the second time threshold are pre-screened from all AGVs; The candidate AGVs are initially sorted based on their spatial distance from the workstation. For candidate AGVs with the same spatial distance, a secondary sort is performed based on their matching degree with the current workstation type label; Based on the results of the preliminary and secondary sorting, a unique AGV is finally dispatched.

7. The method according to claim 1, characterized in that, The adaptive adjustment of the AGV operating parameters triggered based on this partial matching relationship specifically includes: When assigning a partially matched AGV to a workstation type tag with the aforementioned heavy-duty production requirements, its operating speed is reduced to a speed value suitable for heavy-duty production requirements, and the maximum amount of materials it can transport in a single trip is simultaneously limited.

8. The method according to claim 1, characterized in that, During the execution of the hierarchical scheduling decision based on dynamic priority, if it is necessary to allocate an AGV to a workstation with a dynamic priority higher than the first set threshold, thus occupying an AGV that is currently executing a task, then a task transfer process is triggered: Based on the real-time location information of the occupied AGV, it is determined and guided to the nearest idle material transfer station to unload and temporarily store the materials it is currently carrying; The task interrupted by the occupancy event will be reinserted into the task scheduling sequence based on the dynamic priority assigned to it before the interruption.

9. The method according to claim 6, characterized in that, It also includes system self-optimization steps, which include: Periodically collect and update several key performance indicators that reflect the system's scheduling efficiency; When any of the key performance indicators exceeds its preset improvement trigger threshold, the weight parameters associated with the dynamic priority calculation or the decision threshold associated with the hierarchical scheduling decision are dynamically adjusted according to the preset adjustment mapping relationship. Within each verification cycle after adjustment, the improvement effect of the key performance indicators is evaluated, and a decision is made based on the evaluation results as to whether to solidify the adjusted parameters or rules.

Images