Warehouse operation task scheduling system and method

Abstract

The invention provides a warehouse operation task scheduling system and method, and relates to the technical field of intelligent warehousing, and the method comprises the steps: automatically generating a replenishment task after a picking task is completed, integrating the center-of-gravity shift information of an inventory unit and the structural deformation information of a goods shelf, and intelligently deciding the shelving position and posture. According to the system, an active micro-vibration device and a sensor are arranged at key parts of a goods shelf structure, the modal response of the goods shelf structure is obtained, the deformation position and degree are recognized, meanwhile, three-dimensional modeling and gravity center analysis are conducted on an inventory unit, the stability of different goods shelf schemes is evaluated, and placement safety is ensured. And further combining historical shelf information and a mechanical model to predict the potential fatigue risk of the structure, and dynamically updating shelf health information. The system also supports an intelligent rollback and reselection mechanism when the storage location is unavailable, and the stability and flexibility of replenishment scheduling are improved. The method can be widely applied to an intelligent warehousing system in a high-density and complex material storage environment.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent warehousing, and in particular, to a warehouse operation task scheduling system and method. Background Art

[0002] With the rapid development of automated warehousing and intelligent logistics systems, the task scheduling technology in warehouse operations has gradually evolved from static rule-driven to dynamic perception and intelligent decision-making. In the prior art, the generation of replenishment tasks is usually triggered based on fixed rules or inventory thresholds, lacking the ability to dynamically perceive the actual state of the shelves. During the shelving process, the placement positions of inventory units are mostly determined by static bin mappings or simple location optimization algorithms, without considering the center-of-gravity offset characteristics of the inventory units themselves, nor making dynamic responses to the possible deformation risks of the shelf structures.

[0003] In addition, in the existing warehousing systems, the structural health status of the shelves is generally inferred indirectly through manual inspections or long-term monitoring data, and it is impossible to obtain and judge the status in a timely and accurate manner before the execution of replenishment tasks. When the inventory units have eccentric mass structures, or when the shelf structures exhibit deformation phenomena such as local softening or loose connections, it is easy to cause unstable placement, and even lead to problems such as cargo overturning, shelf deformation, or operation failure, posing relatively large safety hazards.

[0004] Therefore, there is an urgent need for a task scheduling method and system that are oriented to dynamic center-of-gravity perception and structural state feedback, to improve the shelving accuracy and operation safety, and to meet the dual requirements of modern intelligent warehouses for high reliability and flexible scheduling. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the present application provides a warehouse operation task scheduling system and method.

[0006] In a first aspect, the present application provides a warehouse operation task scheduling method, including: In response to the completion of the picking task for a target order, generating a replenishment task based on the picking task and a preset replenishment rule, and sending it to an execution device; wherein, the target order includes: inventory unit information to be picked. The generation of the replenishment task includes: Based on the inventory unit information, determining a first shelving bin for the inventory unit; obtaining the shelf health information of the first shelving bin, wherein the shelf health information includes: shelf deformation information. Obtaining the center-of-gravity offset information of the inventory unit corresponding to the inventory unit information; based on the center-of-gravity offset information and the shelf deformation information, adjusting the shelving position of the inventory unit on the shelf. Active vibration devices and sensors are arranged at the columns, crossbeams and connection nodes of the shelf. Among them, the active vibration device is used to apply vibration excitation to the shelf, and the sensor is used to obtain the response signal of the shelf to the vibration excitation, and perform time-frequency analysis on the response signal to determine the shelf deformation information.

[0007] As an optional implementation manner, determining the first storage location of the stock unit based on the stock unit information includes: Based on the stock unit information, determine whether there is a fixed storage location for the stock unit; In response to the stock unit having a fixed storage location, select at least one location from multiple fixed storage locations as the first storage location; In response to the stock unit not having a fixed storage location, obtain the overall location status information, and based on the overall location status information and the stock unit information, determine the locations that match the attributes of the stock unit, and select at least one location from them as the first storage location.

[0008] As an optional implementation manner, when applying vibration excitation to the shelf and obtaining the response signal, the shelf is in an empty state; The performing time-frequency analysis on the response signal to determine the shelf deformation information includes: When the shelf is in an empty state, apply a pulse excitation to the column or crossbeam of the shelf through the active vibration device; Use multiple sensors arranged at the columns, crossbeams and connection nodes of the shelf to collect the corresponding response signals; Perform time-frequency transformation on the response signal to obtain the multi-order modal distribution of the shelf; Compare the multi-order modal distribution with the modal distribution in the pre-stored healthy reference state to identify the modal curve differences that appear in the horizontal or vertical positions of the shelf; Based on the modal curve differences, determine the shelf deformation position in the shelf space coordinate system.

[0009] As an optional implementation manner, after determining the shelf deformation position, it further includes: Perform curvature difference operation on the modal curve corresponding to the shelf deformation position to obtain the modal curvature value at the shelf deformation position; Perform difference calculation on the modal curvature value and the reference curvature value in the healthy reference state to obtain a deformation quantification index; Based on the deformation quantification index, generate the shelf deformation information.

[0010] As an optional implementation manner, the generating the shelf deformation information based on the deformation quantification index further includes: Obtain the historical shelving information of the shelf; Based on the historical shelving information and the stock keeping unit information, obtain the predicted deformation position and predicted deformation degree of the shelf; Based on the predicted deformation position and the shelf deformation position, update the shelf deformation position; Based on the predicted deformation degree and the shelf deformation information, update the shelf deformation information.

[0011] As an alternative implementation, the obtaining of the center of gravity offset information of the stock keeping unit corresponding to the stock keeping unit information includes: Read the three-dimensional shape parameters and weight data of the stock keeping unit from the stock keeping unit information; Based on the three-dimensional shape parameters, determine the geometric center of the stock keeping unit; Based on the weight data, calculate the actual mass center position of the stock keeping unit in the three-dimensional coordinate system; Based on the actual mass center position and the geometric center, determine the center of gravity offset information.

[0012] As an alternative implementation, the shelf deformation information includes: deformation position and deformation degree; The adjusting of the shelving position of the stock keeping unit on the shelf based on the center of gravity offset information and the shelf deformation information includes: Based on the center of gravity offset information, determine the offset direction and offset amount; Within the placeable range of the first shelving bin, determine multiple candidate placement positions and placement postures; Based on the offset direction, offset amount and the deformation position, calculate the safety factor under each candidate placement position and placement posture, and sort the candidate placement positions; Select the candidate placement position and placement posture with the highest safety factor from the sorting result as the final shelving position;

[0013] Place the stock keeping unit at the final shelving position.

[0014] As an alternative implementation, after the adjusting of the shelving position of the stock keeping unit on the shelf based on the center of gravity offset information and the shelf deformation information, it further includes: In response to the stock keeping unit having the fixed shelving bin and there being no matching shelving position in the first shelving bin, select at least one bin from the remaining multiple fixed shelving bins as the second shelving bin; In response to the fact that the fixed storage location for the inventory unit does not exist and there is no matching storage location in the first storage location, at least one storage location is selected from the remaining multiple storage locations that match the attributes of the inventory unit as the second storage location.

[0015] As an optional implementation manner, the obtaining of the predicted deformation position and the predicted deformation degree of the shelf based on the historical storage information and the inventory unit information includes: Obtain the historical storage information; wherein, the historical storage information includes: the inventory unit information stored on the shelf in different time periods, the storage duration, and the centroid offset information corresponding to the inventory unit information; Based on the historical storage information, count the stress accumulation parameters of the shelf; Based on the inventory unit information and the stress accumulation parameters, use a preset shelf mechanical model to predict the potential deformation position and generate the corresponding predicted deformation coordinates; According to the predicted deformation coordinates and the local stiffness parameters of the shelf, calculate the predicted deformation degree corresponding to the predicted deformation position, and after comparing with the current shelf deformation information, output the predicted deformation position and the predicted deformation degree of the shelf.

[0016] In a second aspect, the present application provides a warehouse operation task scheduling system, including: A computing device, configured to, in response to the completion of the picking task for a target order, generate a replenishment task based on the picking task and a preset replenishment rule; determine a first storage location of the inventory unit based on the inventory unit information; obtain the shelf health information of the first storage location; obtain the centroid offset information of the inventory unit corresponding to the inventory unit information; and adjust the storage position of the inventory unit on the shelf based on the centroid offset information and the shelf deformation information; Sensors, arranged at the columns, crossbeams and connection nodes of the shelf, for obtaining response signals; Active micro-vibration devices, arranged at the columns, crossbeams and connection nodes of the shelf, for applying vibration excitation to the shelf; An execution device, configured to execute the replenishment task.

[0017] Compared with the prior art, after the picking task is completed, this application automatically generates a replenishment task, and before the replenishment operation is executed, it integrates the gravity center offset information of the stock keeping unit and the structural health information of the shelf, enabling a more secure, refined, and structurally adaptable shelving strategy. Compared with the prior art which only relies on static bin matching or rule-based location selection, this method introduces a shelf structure perception mechanism composed of an active micro-vibration device and sensors. By performing time-frequency analysis on the response signal, the deformation information of the shelf is obtained in real time. At the same time, combined with the geometric modeling of the stock keeping unit and the calculation of the mass center, the gravity center offset direction and amplitude are evaluated, so as to dynamically adjust the placement position and posture of the stock keeping unit in the replenishment scheduling. This method improves the perception ability and response ability of the shelf structure state in warehousing operations, significantly enhances the stability and safety of the shelving behavior, effectively reduces the risks such as overturning and abnormal stress caused by structural deformation or overweight items, and has higher adaptability and engineering practical value. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flowchart of a method for generating a replenishment task provided by an embodiment of this application; Figure 2 It is a schematic diagram of a scenario for warehouse operation task scheduling provided by an embodiment of this application; Figure 3 It is a flowchart of a method for determining shelf deformation information provided by an embodiment of this application; Figure 4 It is a schematic diagram of a deformed shelf structure provided by an embodiment of this application;

[0019] Figure 5 is Figure 4 a partial enlarged view of part A in DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments.

[0021] This application provides a method for warehouse operation task scheduling, including: In response to the completion of the picking task for the target order, based on the picking task and a preset replenishment rule, a replenishment task is generated and sent to the execution device; wherein, the target order includes: the stock keeping unit information to be picked. Refer to Figure 1 As shown, it is a flowchart of a method for generating a replenishment task provided by an embodiment of this application, including steps S101 to S102, wherein: S101: Based on the stock keeping unit information, determine the first storage location for the stock keeping unit; obtain the shelf health information of the first storage location, where the shelf health information includes: shelf deformation information; S102: Obtain the center of gravity offset information of the stock keeping unit corresponding to the stock keeping unit information; based on the center of gravity offset information and the shelf deformation information, adjust the storage position of the stock keeping unit on the shelf; wherein, sensors and active micro-vibration devices are arranged at the columns, beams and connection nodes of the shelf, and are used to apply vibration excitation to the shelf to obtain a response signal, and perform time-frequency analysis on the response signal to determine the shelf deformation information.

[0022] This application is applicable to an automated warehouse environment built based on a Warehouse Management System (WMS) and an Equipment Control System (ECS). This method is executed by a scheduling control module, which is used to automatically generate a replenishment task after the picking task is completed, and intelligently determine the storage position based on the shelf structure state and the physical characteristics of the stock keeping unit (SKU).

[0023] When a picking task for a certain target order recorded in the WMS is completed, the system triggers a scheduling process and reads the stock keeping unit information associated with the order. The stock keeping unit information includes, but is not limited to, the following fields: unique number (unit_id), three-dimensional size (size), weight (weight), material label (material_tag), fixed storage location flag (fixed_slot_flag), etc. The system determines whether to generate a replenishment task according to a preset replenishment rule set. The replenishment rules may include strategies such as triggering at the lower inventory limit, replenishing immediately after picking, and timed replenishment. If the replenishment rules are met, a corresponding replenishment task object is generated and added to the task queue of the WMS. The task is sent to the corresponding execution device, such as an Automated Guided Vehicle (AGV) or a robotic arm, through a communication network.

[0024] After the replenishment task is generated, the system calls the storage location matching module to retrieve the status information of the currently available storage locations in the WMS. Each storage location object includes fields such as number, location, spatial size, maximum load-bearing capacity, current load, shelf identification to which it belongs, and health status label. If there is a setting for a fixed storage area for the stock keeping unit, the system will screen out idle and matching storage locations from the fixed storage locations to form a candidate set of the first storage location; if there is no fixed storage location, a candidate set will be matched and generated from all idle storage locations according to the attributes of the stock keeping unit such as size, weight, and category, which is denoted as the first storage location.

[0025] To evaluate the structural status of the shelves where the candidate storage locations are located, active micro-vibration actuators (AMA) and vibration response sensor nodes (Accelerometer Sensor Node, ASN) are pre-deployed on the main structural components of the shelves. The above devices are respectively installed in the middle of the crossbeam, the bottom of the column, and the main connection nodes, and are used to apply vibration excitation and collect structural response signals.

[0026] Please refer to Figure 2 , Figure 2 which is a schematic diagram of a scenario for warehouse operation task scheduling provided by an embodiment of this application;

[0027] Among them, AMA and multiple ASNs are respectively arranged on the shelves, and the AGV is responsible for the replenishment execution of tasks.

[0028] When executing the health assessment process, the control system issues a start command to the AMA module through the Modbus TCP bus. The AMA module applies periodic sweep excitation or pulse excitation to the structure within a set frequency range (such as 1 Hz to 2 kHz). The excitation duration is 2 to 3 seconds. The ASN module synchronously collects the structural response signal data at a high-frequency sampling rate (for example, 2 kHz) and transmits it to the central edge computing controller through the bus. When the controller executes the time-frequency analysis module, it uses the short-time Fourier transform (STFT) or wavelet packet decomposition algorithm to perform frequency-domain expansion on the response signal and extract the modal response characteristics of the shelf structure.

[0029] The system compares the currently measured modal frequency vector with the modal frequency in the pre-stored healthy state by taking the difference, and identifies the structural parts with modal frequency offsets. In this way, the potential deformation area of the shelf in the local coordinate system can be determined, which constitutes the "structural health information" of the shelf and is associated with the storage location information for subsequent shelving decisions.

[0030] At the same time, the system performs the center of gravity offset analysis on the stock units to be replenished. First, it calls the geometric analysis module to construct a three-dimensional model based on the three-dimensional structure data and material weight data of the stock unit. The geometric center position is calculated by the following formula: ;

[0031] where represents the position vector of the geometric center of the stock unit in the three-dimensional coordinate system (unit: meter), V is the total volume of the stock unit (unit: ), is the spatial position vector of the volume integration variable. This formula is used to calculate the geometric center of an object with uniform density.

[0032] For inventory units with non-uniform mass distribution, the system introduces a density function to calculate the position of the center of mass : ;

[0033] where, is the center of mass vector (unit: meter), is the total mass (unit: kg), is the position at which the mass density is (unit: ). This formula is the classical centroid integral expression for a rigid body.

[0034] The system calculates the center of gravity offset vector as: ;

[0035] where, represents the offset of the center of mass relative to the geometric center. The system further converts it into polar coordinate form , where respectively represent the horizontal angle, vertical inclination angle of the offset direction and the offset modulus.

[0036] Based on the above two information sources, the system calls the optimized storage location module to screen and evaluate the available locations in the first storage location. By fusing the center of gravity offset vector with the shelf structure health information, the system performs attitude simulation and stability evaluation on the candidate placement locations, and outputs the structural safety factor of each candidate solution. Finally, the combination of the storage location and attitude with the maximum safety factor is selected as the final storage plan for the inventory unit.

[0037] Through the job task scheduling method provided in this embodiment, on the basis of ensuring the replenishment efficiency, the shelf structure state and the physical characteristics of the inventory unit can be fully considered, and the storage behavior can be dynamically adjusted, thereby improving the operation safety and structural response adaptability of the warehousing system.

[0038] As an alternative implementation, determining the first storage location of the inventory unit based on the inventory unit information includes: Based on the inventory unit information, determining whether there is a fixed storage location for the inventory unit; In response to the inventory unit having a fixed storage location, selecting at least one location from the multiple fixed storage locations as the first storage location; In response to the non - existence of a fixed storage location for the inventory unit, obtain the overall storage location status information, and based on the overall storage location status information and the inventory unit information, determine storage locations that match the attributes of the inventory unit, and select at least one storage location therefrom as the first storage location for putting away.

[0039] In specific implementation, the determination strategy for determining the first storage location for putting away can be adjusted according to the diversity of inventory units and the characteristics of dynamic changes in warehousing resources. This strategy aims to solve the problems of overly rigid or untimely response of the storage location assignment rules during the replenishment task execution in the existing warehouse, so as to improve the flexibility and intelligence of the storage path planning while ensuring the matching accuracy.

[0040] After the replenishment task is generated, the system calls the storage location matching module to process the inventory unit to be replenished currently. First, the system reads the fixed storage location flag field (fixed_slot_flag) in the inventory unit information. If the value of this field is True, it means that one or more exclusive fixed storage locations have been bound to this inventory unit in the WMS. For example, specific items that need to be refrigerated or moisture - proof, or high - frequency SKUs partitioned for sorting convenience. In this case, the system filters out the set of storage locations with the current status of being idle from the fixed storage location pool, and makes a preference according to the preset priority rules (such as the shortest handling path, historical stacking stability score, etc.), and selects at least one idle storage location as the first storage location for putting away.

[0041] If the inventory unit is not bound to a fixed storage location for putting away, that is, the value of this field is False, the system enters the dynamic matching mode. In this mode, the system first obtains the overall storage location status information table (Slot State Table), which records the status parameters of all valid storage locations in the current warehouse, including three - dimensional dimensions, maximum load - bearing capacity, current remaining space, historical stacking frequency, current shelf number, and the corresponding structural health status identifier. The system compares the main attribute fields (dimensions, weight, labels) of the inventory unit with the storage location parameters one by one. Through the spatial compatibility judgment (that is, judging whether the size of the storage location is larger than the size margin of the inventory unit) and the load - bearing capacity verification, the idle storage locations that meet the size and weight requirements are screened out as the candidate set.

[0042] Based on the candidate set, the system can also introduce a weighted scoring mechanism, comprehensively considering factors such as the length of the handling path, usage frequency, and shelf health status, to form a comprehensive scoring function. Exemplarily, the comprehensive scoring function can be: ;

[0043] where represents the comprehensive score of candidate storage location i; represents the size adaptation index (such as size matching margin) of candidate storage location i; Indicates the handling convenience index of candidate storage location i (such as the reciprocal of the distance from the main aisle); Indicates the current structural health status score of candidate storage location i (such as set according to the modal analysis results); , , Are weight parameters that satisfy , and can be set according to the actual application strategy.

[0044] The system sorts all candidate storage locations in descending order according to the scoring results, and selects one or more storage locations with the highest scores as the first storage locations for shelving, for subsequent processing by the shelving attitude optimization module.

[0045] Through this embodiment, when the system determines the first storage locations for shelving, it can not only make full use of the preset fixed configuration strategy to ensure the rigid requirements for the management of special items, but also dynamically adapt to the current storage location resources without fixed constraints, improve the storage location utilization rate and replenishment response speed, and further enhance the flexibility, scalability and engineering practicability of the replenishment scheduling system.

[0046] As an alternative embodiment, when applying a vibration excitation to the shelf to obtain a response signal, the shelf is in a vacant state; Please refer to Figure 3 , Figure 4 and Figure 5 , Figure 4 which is a schematic diagram of the deformation structure of a shelf provided by an embodiment of the present application; wherein, part A is the deformation area of the shelf.

[0047] Figure 5 is Figure 4 a partial enlarged view of part A in , showing the deformation generated in part A of the shelf.

[0048] Figure 3 which is a flowchart of a method for determining the deformation information of a shelf provided by an embodiment of the present application, including steps S201 to S205, where: The time-frequency analysis of the response signal to determine the deformation information of the shelf includes: S201: When the shelf is in a vacant state, apply a pulse excitation to the uprights or crossbeams of the shelf through the active micro-vibration device; S202: Use a plurality of sensors arranged at the uprights, crossbeams and connection nodes of the shelf to collect the corresponding response signals; S203: Perform a time-frequency transformation on the response signal to obtain the multi-order modal distribution of the shelf; S204: Compare the multi-order modal distribution with the modal distribution in the pre-stored healthy reference state to identify the modal curve differences that occur in the horizontal or vertical positions of the shelf; S205: Determine the shelf deformation position in the shelf space coordinate system based on the modal curve differences.

[0049] To ensure the accuracy of the shelf structure health assessment and improve the recognition of modal response characteristics, this application preferably conducts vibration detection when the shelf is in an empty state. This method aims to solve the problem of modal parameter drift caused by cargo load interference and improve the repeatability and reliability of the deformation recognition results by constructing a standardized detection condition.

[0050] When the shelf is marked as "detectable" (i.e., there is no stored inventory unit currently), the system starts the modal detection process. The control system sends a pulse excitation command to the AMA. In a specific implementation, the pulse excitation signal is a short-time broadband pulse, and the action time is usually set to 1 - 2 seconds, and the frequency range covers 1 Hz to 2 kHz, aiming to excite multiple order modal responses of the shelf structure. The excitation point is usually selected in the middle of the uprights or crossbeams of the shelf to improve the excitation efficiency and avoid the influence of boundary conditions.

[0051] Multiple ASNs are pre-deployed at the key structural nodes of the shelf, including the bottom of the uprights, the ends of the crossbeams, and the horizontal and vertical connection nodes. The above sensors synchronously collect response signals at a high sampling rate (such as 2 kHz) and transmit them to the edge control computing device in real time through the CAN bus or industrial Ethernet.

[0052] After receiving the multi-point response signals, the central controller executes the time-frequency analysis module, performs short-time Fourier transform or wavelet packet decomposition on the signals, extracts the frequency domain components of the multi-channel response signals, and generates the multi-order modal response curves of the shelf structure. The modal distribution includes the modal frequencies from the first order to the fifth order (or higher) and the corresponding modal vibration mode amplitudes.

[0053] The system compares the modal distribution data detected in the current period with the "modal distribution data in the healthy reference state" pre-stored in the database one by one. During the comparison process, the system identifies the abnormal modal response areas of the shelf structure in the horizontal (X-axis direction) or longitudinal (Z-axis direction) according to characteristics such as the modal frequency offset amplitude, the spatial phase difference of the vibration mode curves, and the change in response energy distribution, and then determines the possibility of its local deformation.

[0054] Finally, the system maps the spatial points corresponding to the identified modal curve differences to the local coordinate system of the shelf structure, determines the possible deformation positions of the shelf, serves as the position item in the "shelf deformation information", and conducts data docking with the subsequent on-shelf stability calculation module.

[0055] Through this embodiment, high-precision modal detection tasks can be carried out when the shelves are out of stock, the interference effect of dynamic loads on modal responses can be eliminated to the greatest extent, the accuracy and stability of deformation recognition can be effectively improved, and more reliable structural health reference data can be provided for subsequent shelving optimization.

[0056] Exemplarily, in a certain high-density automated warehouse, deformation monitoring is carried out on a group of standard light-load shelves. The shelf has a four-layer structure, and the main body is composed of steel columns, crossbeams and pin connection members. The width of each layer is 1.2 m, the depth is 0.8 m, and the height is 2.0 m. The maximum load of each layer is designed to be 300 kg. To ensure the detection accuracy, the detection task is set to be executed under the no-load state at night.

[0057] At the middle parts of two columns, the ends of four crossbeams, and four key connection nodes of each group of shelves, a total of 8 triaxial acceleration sensors (model ADXL356, sampling rate set to 2 kHz) are installed, and the sensors are connected to the edge acquisition controller through the CAN bus. The vibration excitation is applied by an electromagnetic active micro-vibration device (model MVA-10) installed at the center position of the crossbeam. A single-pulse excitation signal is used, the pulse duration is 0.5 s, the peak frequency is 1200 Hz, and the signal is uniformly triggered and sent by the controller.

[0058] After the data acquisition is completed, the controller preprocesses the 8-channel response signals, including filtering, denoising and normalization operations, and then performs time-frequency expansion on each channel signal by using the short-time Fourier transform to extract the first-order to third-order main modal frequency and amplitude characteristics of each sensor measurement point. The extracted modal frequencies are: the first order is 135 Hz, the second order is 428 Hz, and the third order is 912 Hz.

[0059] The system compares the current modal frequency data with the modal reference data recorded for this type of shelf at the factory state. It is found that there is a frequency drop of more than 10% near the first-order frequency at the crossbeam connection node numbered "R2-N2", and at the same time, the modal vibration mode energy distribution shows obvious asymmetry characteristics. Based on this, the system judges that there may be loosening of the crossbeam fixation or distortion of the column in this area, and marks the position of the abnormal area in the local coordinate system of the shelf, with the coordinates (X = 0.6 m, Z = 1.2 m).

[0060] Subsequently, the system synchronously writes the structural state identifier of the abnormal area into the shelf health information field in the WMS, marks this area as "to be avoided", for the replenishment task scheduling system to refer to when performing the selection of shelving positions, to ensure that no heavy-load shelving operations are carried out in this area.

[0061] As an alternative embodiment, after determining the deformation position of the shelf, it further includes: Perform curvature difference operation on the modal curve corresponding to the deformed position of the shelf to obtain the modal curvature value at the deformed position of the shelf; Perform difference calculation on the modal curvature value and the reference curvature value in the healthy reference state to obtain the deformation quantification index; Generate the shelf deformation information based on the deformation quantification index.

[0062] In a specific implementation, after identifying a potential deformed position of the shelf, the system further performs curvature analysis on the modal response data corresponding to this position to achieve quantitative calculation of the deformation degree, thereby generating complete shelf deformation information. This implementation aims to solve the technical defect of "only positioning but not quantifying" the deformation in the existing modal monitoring system, and improve the recognition sensitivity of the system to structural damage or plastic deformation and the accuracy of scheduling control.

[0063] After the controller completes the modal response frequency domain analysis and identifies the abnormal modal curve region, the system extracts the main modal response curve of the corresponding measurement point, and sets this curve as the displacement response function , where x represents the spatial coordinate (unit: m) along the transverse or longitudinal structural axis of the shelf. The system performs a second-order difference operation on this curve near the deformed area to calculate its modal curvature value, and the expression is as follows: ; Among them, represents the modal curvature value at position x, and the unit is ; is the displacement response function, that is, the normalized displacement value of the vibration mode at this point; : is the spacing between adjacent sensor layouts, and the unit is meters (m); the denominator is used to normalize the spatial gradient to ensure the physical consistency of the curvature calculation.

[0064] The system retrieves the modal response curve at the corresponding position of this type of shelf in the healthy reference state from the database, denoted as , and calculates its curvature in the same way . Subsequently, the system calculates the deformation quantification index : ; Among them, reflects the deviation of the current structure from the healthy state in modal curvature, and is used to quantitatively judge the degree of structural deformation.

[0065] To improve the robustness, the system can perform moving average or weighted average processing on the of multiple adjacent measurement points at the deformed position to generate the final deformation quantification index Based on the comparison of the index value with the threshold (e.g., exceeding 3σ standard deviation of the healthy state), the system identifies this position as a "structural anomaly point" and writes the deformation position coordinates and the corresponding quantitative index into the shelf deformation information field.

[0066] Through the above processing, not only can the accurate positioning of the spatial position of the shelf deformation be achieved, but also the degree value of the deformation can be generated on this basis for subsequent calls by modules such as safety factor estimation, scheduling avoidance path planning, and maintenance warning.

[0067] Exemplarily, when further curvature analysis is performed on the position of the abnormal area detected in the above-mentioned shelf (coordinates are X = 0.6m, Z = 1.2m), the system selects the displacement data (unit normalized) of the main modal response curves corresponding to this node and one adjacent node on each of its left and right sides (three points) as follows: ; Substitute into the modal curvature formula: ; Call the reference curvature value of this type of shelf in the healthy reference state in the database , then the deformation quantification index is: ; If the curvature deviation control limit in the normal state of the structure of this type of shelf according to the statistical model is , then the deformation quantification index at this position exceeds the tolerance range, and the system generates the following structural health information entry accordingly: Deformation position: X = 0.6m, Z = 1.2m; Deformation curvature difference: ; Deformation level: medium (the system can set the level threshold).

[0068] This entry is written into the structure state database as the shelf health information and is called by the subsequent shelf placement strategy adjustment module for dynamic avoidance and stacking attitude correction strategies.

[0069] As an alternative implementation, generating the shelf deformation information based on the deformation quantification index further includes: Obtain the historical shelf placement information of the shelf; Based on the historical shelf placement information and the stock unit information, obtain the predicted deformation position and predicted deformation degree of the shelf; Based on the predicted deformation position and the shelf deformation position, update the shelf deformation position; Based on the predicted deformation degree and the shelf deformation information, update the shelf deformation information.

[0070] As an alternative implementation, obtaining the predicted deformation position and predicted deformation degree of the shelf based on the historical shelf placement information and the stock unit information includes: Obtain the historical shelving information; wherein, the historical shelving information includes: inventory unit information shelved on the shelf at different time periods, shelving duration, and centroid offset information corresponding to the inventory unit information; Based on the historical shelving information, count the stress accumulation parameters of the shelf; Based on the inventory unit information and the stress accumulation parameters, use a preset shelf mechanical model to predict potential deformation positions and generate corresponding predicted deformation coordinates; According to the predicted deformation coordinates and the local stiffness parameters of the shelf, calculate the predicted deformation degree corresponding to the predicted deformation position, and after comparing with the current shelf deformation information, output the predicted deformation position and predicted deformation degree of the shelf.

[0071] Based on the current deformation position and modal curvature of the shelf obtained through real-time detection, the system further introduces historical shelving information and a mechanical prediction model to establish a prediction-calibration mechanism for improving the accuracy and timeliness of shelf deformation information. This implementation mainly solves the following problems: one is to give early warnings to structural areas that have not yet shown obvious modal anomalies; the second is to identify potential long-term fatigue deformation in combination with historical load behaviors; the third is to enhance the temporal continuity and dynamic response ability of shelf health information.

[0072] The system first calls the historical data module to extract the historical shelving information of the target shelf from the structural health database. The historical shelving information includes at least the following fields: timestamp, shelved inventory unit number, shelving duration, weight of the inventory unit, three-dimensional shape parameters, and centroid offset vector. Through time series aggregation, this information can constitute the comprehensive load distribution borne by different positions of the shelf in multiple time periods.

[0073] In engineering implementation, the shelf can be divided into several spatial cells (Voxels), and each shelving behavior is projected into the corresponding spatial unit to accumulate its equivalent stress influence. The system defines the "stress accumulation parameter" , which represents the equivalent cumulative load intensity experienced by the shelf at position x, and the calculation method is as follows: ; wherein, : the number of times the inventory unit acts on position x in the historical shelving record; : the weight (kg) of the qth shelved inventory unit; : the shelving duration (h) of the qth shelved inventory unit at this position; : the centroid offset coefficient of the qth shelved inventory unit, that is, the ratio of the offset vector modulus to the size, used to characterize the moment influence; The unit of

[0074] Based on the above stress accumulation parameters, the system conducts predictive analysis in combination with the structural mechanics model of the shelf. The structural mechanics model can be a parametric finite element model (FEM) or a neural network model trained based on historical data, and is used to simulate the stress-strain evolution path occurring at the structural nodes under long-term action.

[0075] The system calls the prediction model interface, takes the current stress distribution map and the local stiffness parameters of the shelf (which can be set through structural modeling or obtained through factory calibration) as inputs, and outputs the spatial position coordinates (predicted deformation positions) where the shelf may deform and the corresponding predicted deformation values (predicted deformation degrees). The above prediction results constitute the "shelf predicted deformation information".

[0076] The system performs spatial position matching and numerical comparison between the prediction results and the actual deformation information obtained through real-time modal analysis. If the two are adjacent or overlapping in space, the system uses a weighted fusion method to update the deformation position coordinates to improve the positioning accuracy. If the predicted position is completely different from the detected position, the system marks the predicted position as a "warning area" and adds it to the avoidance candidate list of the subsequent shelving strategy.

[0077] In addition, the system performs a fusion calculation on the predicted deformation degree and the detected modal curvature index, for example, using exponential moving average (EMA): ; where is the finally generated deformation quantification index, is the actually measured deformation quantification index, is the predicted deformation quantification index, is the fusion weight factor, which can be set according to the detection frequency and model confidence.

[0078] In the finally generated shelf deformation information, it will simultaneously include: the real-time measured modal curvature abnormal area, the predicted structural fatigue trend area, and the fused comprehensive deformation quantification index. This information can be used to dynamically update the shelf status label, guide the selection of shelving positions, or push maintenance tasks.

[0079] Exemplarily, the system performs a historical stress accumulation analysis on the right side area of the shelf numbered "R2-04". By retrieving the shelving records in the past 60 days, it is found that this area bears a type of metal component with a large center of gravity offset (the center of gravity offset coefficient is about 0.18) every day, with an average weight of 43 kg and an average single stay time of 26 hours.

[0080] The calculated stress accumulation parameters for this area are: ;

[0081] The system inputs this parameter into the shelf fatigue response prediction model (established based on the finite element regression curve), and the output shows that there is a potential deformation risk at the coordinates (X = 0.95m, Z = 1.6m), and the predicted curvature offset is .

[0082] It highly coincides spatially with the actual deformation quantification index obtained from the previous detection . The system performs weighted fusion processing and finally determines that the deformation degree of this area is , and updates it as the high-risk area label of the current shelf. At the same time, this area is marked in the prohibited placement area list for future replenishment task scheduling to avoid further expansion of structural overload.

[0083] As an alternative implementation method, to further improve the accuracy and acquisition efficiency of the center of gravity information of inventory units, when the system calculates the center of gravity offset information, the following data modeling and algorithm implementation means can be adopted.

[0084] First of all, the three-dimensional shape parameters included in the inventory unit information can be directly called by the WMS from the material warehousing filing data, or obtained through access to automatic measurement devices such as structured light scanners, industrial X-ray CT devices, and laser ranging systems. The shape data is generally represented by a three-dimensional mesh file in STL (stereolithography), OBJ, or PLY format. In the edge computing platform, the CAD parsing engine can be used to convert it into a point cloud model or a parametric geometric body.

[0085] After the three-dimensional model is constructed, the system calls the volume segmentation algorithm to discretize the overall structure in space. For example, the tetrahedral mesh generation method (Delaunay triangulation) is used to divide the inventory unit into N volume units. The system traverses and integrates the spatial centroid of each volume unit to calculate the overall geometric center: ;

[0086] where G is still the position vector (unit: meter) of the geometric center of the inventory unit defined above in the three-dimensional coordinate system, that is, by integrating the volume-weighted centroid of each volume unit, the geometric center vector of the overall inventory unit is obtained. represents the position vector of the w-th small volume unit (mesh unit), usually the geometric centroid position of this small volume unit (unit: meter). It should be noted that the definition of here does not conflict with the definition of r above. In the actual engineering implementation, the system spatially discretizes the integral volume V, and uses the position vector of the geometric center of each volume unit to approximately represent the spatial position vector r within the volume integration region to complete the numerical transformation from integration to summation.

[0087] Subsequently, the quality data can be obtained through the embedded weighing module or the standard weight field in the historical database. If the inventory unit materials are not evenly distributed, the system can call its material hierarchy structure and combine the mass density of each sub-structure with the geometric volume to perform weighted integration and calculate the actual center of mass: ;

[0088] where, is the total mass, represents the local mass density corresponding to the w-th volume unit (unit: kilograms per cubic meter, ). In addition, in the actual engineering implementation, the system adopts a spatial discretization method for the three-dimensional model, converting the center of mass integral calculation into a form of weighted summation of local volume units. However, the physical meaning of the obtained center of mass vector C is consistent with the above integral definition, and the unit is meters (m). Finally, through the vector difference: ;

[0089] the system obtains the center of gravity offset vector, and its modulus represents the offset amplitude, and the direction is determined by the vector direction. The system converts this vector into polar coordinate form ( ) for unified invocation by downstream modules.

[0090] To optimize the system efficiency, the system can cache the calculated center of gravity offset information into the inventory unit attribute field and reuse it in subsequent multiple schedules to avoid redundant modeling and repeated integral calculations. When there are multiple similar inventory units, the system can also introduce a representative modeling and parameter fitting mechanism to reduce the modeling cost.

[0091] It should be noted that for the convenience of implementation in the actual engineering system, the system performs spatial discretization processing on the three-dimensional model of the inventory unit, and uses volume segmentation and weighted summation instead of theoretical integration to realize the numerical calculation of the geometric center and the center of mass.

[0092] As an alternative implementation, the shelf deformation information includes: the deformation position and the deformation degree; Based on the center of gravity offset information and the shelf deformation information, adjusting the stocking position of the inventory unit on the shelf includes: Based on the center of gravity offset information, determining the offset direction and the offset amount; Within the placeable range of the first stocking bin, determining multiple candidate placement positions and placement postures; Based on the offset direction, offset amount and deformation position, calculate the safety factor under each candidate placement position and placement posture, and sort the candidate placement positions; Select the candidate placement position and placement posture with the highest safety factor from the sorting results as the final shelf position; The inventory unit is placed in the final shelf position.

[0093] Furthermore, before the inventory unit is put on the shelf, the system combines its center of gravity offset information with the structural deformation information of the shelf corresponding to the selected first shelf location to dynamically adjust the specific placement and posture of the inventory unit on the shelf, thereby improving the overall shelf stability and structural response safety. The main problem solved by this method is how to use the eccentricity of the inventory unit and the local softening characteristics of the shelf to coordinately perform refined position optimization to avoid the risk of structural overload, overturning or local extrusion.

[0094] In a specific implementation, the system first offsets the vector from the center of gravity of the inventory unit Extract its offset direction (polar coordinate angle ) and offset Combined with the structural health data of the shelf area corresponding to the target storage location, the deformation position coordinates and their corresponding deformation degree (such as modal curvature difference or predicted softening factor) are obtained to form a shelf deformation information set.

[0095] Then, the system builds a set of candidate locations within the placement range of the first shelf location. To this end, the system discretizes the placement area in space and generates several candidate points for placement in the horizontal (X), vertical (Y) and layer height (Z) directions of the shelf with a fixed step spacing. Combined with the placement methods allowed for each inventory unit (such as bottom down, inverted, sideways, etc.), each candidate placement position is assigned one or more acceptable placement postures to build a set of candidate shelf solutions. : ; in, : The spatial position coordinates of the Tth candidate placement point in the shelf coordinate system, in meters (m); : The rotation angle of the jth candidate posture around the X axis (horizontal axis), in degrees (°), indicating a flipped or supine state; : The rotation angle of the jth candidate posture around the Z axis (vertical axis), in degrees (°), indicating the direction; is the total number of candidate placement points, is the total number of candidate poses, the combined set of listing solutions Total Candidate listing options.

[0096] To quantitatively compare each candidate shelving plan, a structural safety factor evaluation formula is designed based on the principle of physical moment balance and is defined as follows: ; where : The structural safety factor (dimensionless) of candidate plan (T, j), the larger the value, the more stable; : The mass (kg) of the stock keeping unit, obtained from the inventory record or weighing module; : The gravitational acceleration constant, taking ; : The horizontal distance (m) from the candidate point to the load-bearing center of the shelf (geometric center of the support surface), reflecting the support strength; : The offset distance (m) between the center of gravity and the support surface in the current posture, the larger the value, the higher the risk of overturning; : The shelf structure softening factor corresponding to the candidate placement position, indicating the stiffness attenuation degree of the local structure under excitation, with the unit of Newton-meter (N·m), which can be obtained by calculating the modal curvature difference or modal frequency shift.

[0097] The above formula couples the center of gravity offset trend with the structural response characteristics of the shelf to form a shelving stability evaluation system for the dynamic structural state. When , it is regarded as unsafe; when , it is considered that this position has a high stability redundancy.

[0098] The system calculates for all candidate plans and sorts them in descending order. Finally, the combination with the highest safety factor is selected as the final shelving position and posture of the stock keeping unit, and instructions are sent to the AGV or robotic arm through the equipment control system to complete the shelving operation.

[0099] Through this implementation method, the stacking stability of the stock keeping unit under the shelf structure constraints can be significantly improved, and the physical risks caused by the superposition of the center of gravity offset and structural deformation can be minimized, thereby enhancing the engineering robustness and safety control ability of the automatic replenishment system.

[0100] Exemplarily, a smart warehouse needs to precisely stack a batch of eccentric cylindrical chemical reagent bottles. The mass of this type of stock keeping unit is 8.5 kg, the height is 0.4 m, the offset of the center of gravity relative to the central axis is 0.09 m, and the direction is the front upper side of the bottle mouth. Through the aforementioned center of gravity modeling module, the system obtains its offset vector , the offset direction , .

[0101] The system obtains that the first storage location for this stock keeping unit is the lower left area of shelf "R3-02", retrieves the deformation information of this shelf from the modal analysis module, and identifies that there is a structural softening area at the coordinate (X = 0.3m, Z = 0.4m) with a relatively high degree of deformation. The system generates 36 sets of position-pose combination schemes within the candidate range, calculates the safety factor for each set of schemes, and finds that when the offset direction of some poses faces the deformation area, its is in the same direction as the flexibility direction, resulting in a safety factor lower than 1.5.

[0102] Finally, the system selects a set of inverted placement schemes after rotation such that its heavy side is away from the deformation area and close to the center of gravity of the support surface, and the safety factor reaches 3.8. The system records this position and pose and uses it as the final storage instruction to complete the precise placement through the AGV platform.

[0103] As an alternative implementation, after adjusting the storage position of the stock keeping unit on the shelf based on the center of gravity offset information and the shelf deformation information, it further includes: In response to the stock keeping unit having the fixed storage location and there being no matching storage position in the first storage location, at least one location is selected from the remaining multiple fixed storage locations as the second storage location; In response to the stock keeping unit not having the fixed storage location and there being no matching storage position in the first storage location, at least one location is selected from the remaining multiple locations that match the attributes of the stock keeping unit as the second storage location.

[0104] After the system completes the adjustment of the storage position based on the center of gravity offset information and the shelf deformation information, if it is found that there is no available matching placement position or pose in the current first storage location, the system will initiate a dynamic fallback process to further search for other candidate locations to ensure the smooth execution of the replenishment task and avoid structural risk stacking. The main problem solved by this method is: when the current recommended location cannot meet the stable storage conditions due to severe local deformation of the shelf, large center of gravity offset, or other uncontrollable factors during actual operation, how the system can have the adaptive ability to re-plan the path and expand the selection space.

[0105] In a specific implementation, after the system attempts to perform the safety assessment within the first storage location (refer to the calculation process of the aforementioned safety factor ), if the safety factors of all candidate placement position and pose combinations are lower than the lower limit threshold set by the system (such as ), the system will consider that there is no matching storage position in the current first storage location and trigger the fallback logic.

[0106] The system first determines whether the current inventory unit has a fixed storage location (i.e., its fixed_slot_flag = True). If true, the system excludes the storage locations that have failed the current attempt from the predefined set of fixed storage locations. Among the remaining several fixed storage locations, it re-performs the first storage location selection logic (see the matching process above), forms a second candidate set of storage locations, and re-invokes modules such as shelf health analysis, center of gravity offset calculation, and safety scoring to complete a new round of storage location decision-making process.

[0107] If the inventory unit does not have a fixed storage location, the system enters the general storage location matching mode. It excludes the first storage location that has been attempted from the remaining idle storage locations that match the attributes of the inventory unit, and again screens and generates a new candidate set of storage locations as the second storage location according to criteria such as spatial dimensions, load-bearing range, availability status, and shelf health condition. And it executes the location optimization process in the same way.

[0108] To avoid infinite loops or frequent rollbacks, the system can set a maximum number of attempts, a rollback progression rule, or a task termination criterion (such as pushing a manual intervention flag after three consecutive failures). The entire rollback mechanism is implemented as a modular process in the system architecture, supporting dynamic loading and parallel invocation to ensure the execution stability and response flexibility of the replenishment task.

[0109] Through the introduction of this embodiment, the system has the elastic adaptation ability for real-world unpredictable working conditions, improving the engineering robustness of the scheduling logic and the task success rate, especially suitable for large-scale inventory management and intelligent warehousing scenarios with frequent fluctuations in the structural health state.

[0110] Based on the same inventive concept, an embodiment of the present application also provides a warehouse operation task scheduling system corresponding to a warehouse operation task scheduling method. Since the principle of solving problems by the system in the embodiment of the present application is similar to that of the above-mentioned method of a logistics charging engine based on rule generation in the embodiment of the present application, the implementation of the system can refer to the implementation of the method, and the repeated parts will not be described again. The system includes: A computing device, configured to, in response to the completion of the picking task for the target order, generate a replenishment task based on the picking task and a preset replenishment rule; determine the first storage location of the inventory unit based on the inventory unit information; obtain the shelf health information of the first storage location; obtain the center of gravity offset information of the inventory unit corresponding to the inventory unit information; and adjust the storage location of the inventory unit on the shelf based on the center of gravity offset information and the shelf deformation information; Sensors, arranged at the columns, crossbeams, and connection nodes of the shelf, for obtaining response signals; An active micro-vibration device, arranged at the columns, crossbeams, and connection nodes of the shelf, for applying vibration excitation to the shelf; An execution device for executing the replenishment task.

[0111] Exemplarily, a computing device, such as an industrial edge computing unit, an integrated server, or a scheduling master control module, is used to execute scheduling control logic. The specific functions include: after the target order picking task recorded in the WMS is completed, triggering the replenishment process based on the task completion signal; reading the inventory unit information and invoking the replenishment rule set to generate a replenishment task and allocate it to the execution queue; invoking the storage location matching logic to determine the first storage location for putting away based on the inventory unit attributes and the storage location status; invoking the modal recognition algorithm module to obtain the current structure state of the shelf and generate shelf deformation information; executing the center of gravity calculation module to extract the center of gravity offset of the inventory unit geometric model and quality data; integrating the center of gravity information and the shelf deformation information, executing the optimization of the putting away position and the attitude simulation, and outputting the final putting away instruction.

[0112] Exemplarily, a sensor is used to obtain the shelf structure response signal. This part consists of multiple high-sensitivity vibration response sensors (such as accelerometers, MEMS inertial modules), which are arranged at the key structure nodes of the shelf, including columns, beams, and the horizontal and vertical connection parts. Its functions include: during the application of the excitation, collecting multi-channel response signals at a high sampling rate; transmitting the signals to the computing device in real time for time-frequency analysis and modal parameter extraction; long-term monitoring of the shelf state changes to support predictive maintenance and health history modeling.

[0113] Exemplarily, an active micro-vibration device is deployed at the structure node positions of the shelf and is arranged in cooperation with the sensor. The active micro-vibration device can be an electromagnetic exciter, a piezoelectric drive module, etc., and has the ability of short-time pulse or swept-frequency excitation. Its functions include: receiving the control instruction from the scheduling system and applying the control excitation when the shelf is empty or within a specified time window; the excitation frequency range covers the first to third order modal response intervals of the shelf structure; working in cooperation with the sensor to achieve high-precision modal recognition and local flexibility inference.

[0114] Exemplarily, an execution device is used to specifically execute the replenishment task and the putting away action. This device can include but is not limited to: Automated Guided Vehicle (AGV), Robot Arm, stacker, pallet lifting device, etc. Its functions include: receiving the replenishment task instruction and the putting away parameters sent by the computing device; accurately completing the transportation and placement of the inventory unit from picking up to the specified location; supporting the linkage with the WMS and ECS to achieve the closed-loop execution of the task and the status feedback.

[0115] Each module realizes data interaction through industrial buses (such as CAN, Modbus TCP) and wireless communication protocols (such as Wi-Fi 6 or LoRa), and is coordinated by a unified system management software. This system can run on a local server or an edge cloud platform, and has functions such as remote upgrade, module reconstruction, and algorithm strategy adaptation.

[0116] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this application can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of this application.

Claims

1. A warehouse operation task scheduling method, characterized in that: include: In response to the completion of the picking task for the target order, a replenishment task is generated based on the picking task and the preset replenishment rule, and is sent to the execution device; wherein the target order includes: inventory unit information of the expected picking; The generating replenishment task comprises: Based on the inventory unit information, determine a first shelf storage location of the inventory unit; obtain shelf health information of the first shelf storage location, wherein the shelf health information includes: shelf deformation information; Acquire the center of gravity offset information of the inventory unit corresponding to the inventory unit information; adjust the shelf position of the inventory unit on the shelf based on the center of gravity offset information and the shelf deformation information; Active micro-vibration devices and sensors are arranged at the columns, beams and connection nodes of the shelf, wherein the active micro-vibration device is used to apply vibration excitation to the shelf, and the sensor is used to obtain the response signal of the shelf to the vibration excitation, and time-frequency analysis is performed on the response signal to determine the deformation information of the shelf.

2. A warehouse operation task scheduling method according to claim 1, characterized in that: The determining, based on the inventory unit information, a first shelf location of the inventory unit comprises: Based on the inventory unit information, determining whether the inventory unit has a fixed shelf location; In response to the inventory unit having a fixed shelf storage location, selecting at least one storage location from a plurality of the fixed shelf storage locations as a first shelf storage location; In response to the absence of a fixed shelving location for the inventory unit, total location status information is obtained, and based on the total location status information and the inventory unit information, locations that match the attributes of the inventory unit are determined, and at least one location is selected as the first shelving location.

3. A warehouse operation task scheduling method according to claim 1, characterized in that: When applying vibration excitation to the shelf and obtaining a response signal, the shelf is in an empty state; The performing time-frequency analysis on the response signal to determine the shelf deformation information includes: When the shelf is empty, applying pulse excitation to the columns or beams of the shelf through the active micro-vibration device; Collect corresponding response signals using multiple sensors arranged at the columns, beams and connection nodes of the shelf; Performing time-frequency transformation on the response signal to obtain multi-order modal distribution of the shelf; Comparing the multi-order modal distribution with the modal distribution in a pre-stored healthy reference state to identify modal curve differences occurring in the lateral or longitudinal position of the shelf; Based on the modal curve difference, the shelf deformation position is determined in the shelf space coordinate system.

4. A warehouse operation task scheduling method according to claim 3, characterized in that: After determining the deformation position of the shelf, the method further includes: Performing a curvature difference operation on the modal curve corresponding to the deformation position of the shelf to obtain the modal curvature value at the deformation position of the shelf; Performing differential calculation on the modal curvature value and the reference curvature value under the healthy reference state to obtain a deformation quantitative index; The shelf deformation information is generated based on the deformation quantification index.

5. A warehouse operation task scheduling method according to claim 4, characterized in that: The step of generating the shelf deformation information based on the deformation quantification index further comprises: Get the historical shelf information of the shelf; Based on the historical shelf information and the inventory unit information, obtaining a predicted deformation position and a predicted deformation degree of the shelf; Based on the predicted deformation position and the shelf deformation position, updating the shelf deformation position; Based on the predicted deformation degree and the shelf deformation information, the shelf deformation information is updated.

6. A warehouse operation task scheduling method according to claim 1, characterized in that: The obtaining of the center of gravity offset information of the inventory unit corresponding to the inventory unit information includes: Reading three-dimensional shape parameters and weight data of the inventory unit from the inventory unit information; Based on the three-dimensional shape parameters, determining a geometric center of the inventory unit; Based on the weight data, calculating the actual mass center position of the inventory unit in a three-dimensional coordinate system; The center of gravity offset information is determined based on the actual mass center position and the geometric center.

7. A warehouse operation task scheduling method according to claim 6, characterized in that: The shelf deformation information includes: deformation position and deformation degree; The adjusting the shelf position of the inventory unit on the shelf based on the center of gravity offset information and the shelf deformation information includes: Based on the center of gravity offset information, determining an offset direction and an offset amount; Determine a plurality of candidate placement positions and placement postures within the placement range of the first shelf storage location; Based on the offset direction, offset amount and deformation position, calculate the safety factor under each candidate placement position and placement posture, and sort the candidate placement positions; Select the candidate placement position and placement posture with the highest safety factor from the sorting results as the final shelf position; The inventory unit is placed in the final shelf position.

8. A warehouse operation task scheduling method according to claim 2, characterized in that: After adjusting the shelf position of the inventory unit on the shelf based on the center of gravity offset information and the shelf deformation information, the method further includes: In response to the existence of the fixed shelf storage location for the inventory unit and no matching shelf location in the first shelf storage location, selecting at least one storage location from the remaining plurality of the fixed shelf storage locations as a second shelf storage location; In response to the inventory unit not having the fixed shelving location and no matching shelving position in the first shelving location, at least one location is selected from the remaining multiple locations that match the attributes of the inventory unit as the second shelving location.

9. A warehouse operation task scheduling method according to claim 5, characterized in that: The obtaining of the predicted deformation position and predicted deformation degree of the shelf based on the historical shelf information and the inventory unit information includes: Acquire the historical shelf information; wherein the historical shelf information includes: inventory unit information, shelf time, and center of gravity offset information corresponding to the inventory unit information of the shelves in different time periods; Based on the historical shelf information, statistically analyzing stress accumulation parameters on the shelf; Based on the inventory unit information and the stress accumulation parameter, using a preset shelf mechanics model, predicting the potential deformation position and generating corresponding predicted deformation coordinates; The predicted deformation degree corresponding to the predicted deformation position is calculated according to the predicted deformation coordinates and the local stiffness parameters of the shelf, and after comparing with the current shelf deformation information, the predicted deformation position and predicted deformation degree of the shelf are output.

10. A warehouse operation task scheduling system, used to implement a warehouse operation task scheduling method according to any one of claims 1 to 9, characterized in that: include: A computing device, for generating a replenishment task based on the picking task and a preset replenishment rule in response to completion of the picking task for the target order; Based on the inventory unit information, determine the first shelf location of the inventory unit; Obtain shelf health information of the first shelf storage location; Obtaining center of gravity offset information of the inventory unit corresponding to the inventory unit information; Based on the center of gravity offset information and the shelf deformation information, adjusting the shelf position of the inventory unit on the shelf; Sensors are placed on the columns, beams and connection nodes of the shelves to obtain response signals; Active micro-vibration devices are arranged on the columns, beams and connection nodes of the shelves to apply vibration excitation to the shelves; An execution device is used to execute the replenishment task.

Images