Warehousing space utilization rate optimization method and system based on digital twinning

Abstract

The invention relates to the technical field of digital twinning, in particular to a storage space utilization rate optimization method and system based on digital twinning, and the method comprises the steps: determining three-dimensional visual modeling input data of a storage digital twinning body according to an obtained environment data set of a storage space and an operation state data set of storage equipment; performing feature extraction processing on the three-dimensional visual modeling input data to generate a storage feature set, and generating a three-dimensional visual model of the storage digital twin based on the storage feature set; calling a spatial layout simulation analysis algorithm, performing joint simulation processing on shelf distribution characteristics, equipment moving path characteristics and goods access frequency characteristics in the three-dimensional visual model, and generating a spatial layout simulation result set; and generating a warehousing operation adjustment strategy set according to the path conflict simulation result and the space idle simulation result in the space layout simulation result set, and indicating a warehousing control system to perform space layout optimization processing based on the warehousing operation adjustment strategy set.

Description

Technical Field

[0001] The present invention relates to the field of digital twin technology, and more specifically, to a warehouse space utilization optimization method and system based on digital twin. Background Art

[0002] In the field of warehouse space management, with the rapid development of the modern logistics industry, the scale of warehousing continues to expand, the variety of goods is increasing, and the complexity of warehouse management is also increasing. Efficient warehouse space management plays a key role in improving logistics efficiency and reducing operating costs. Rational utilization of storage space, smooth operation of equipment, and optimized operational processes are all important factors in ensuring efficient warehousing operations.

[0003] However, existing warehouse space management technologies are often unable to fully obtain environmental data of the warehouse space and operating status data of the warehouse equipment, resulting in a lack of in-depth understanding of the actual situation of the warehouse; when constructing the warehouse model, there is a lack of effective feature extraction methods, and the model is difficult to accurately reflect the actual situation of the warehouse; for key factors such as shelf distribution, equipment movement paths and cargo access frequency, it is impossible to conduct a comprehensive analysis, making it difficult to discover potential problems in a timely manner; and in the face of problems such as path conflicts and idle space, the existing technology lacks a scientific and effective strategy generation mechanism, making it difficult to effectively guide the warehouse control system, thus failing to achieve the optimization of the warehouse space layout, resulting in low warehouse operation efficiency and serious resource waste. Summary of the Invention

[0004] In view of this, the present invention provides a method and system for optimizing storage space utilization based on digital twins.

[0005] An embodiment of the present invention provides a warehouse space utilization optimization method based on digital twins, which is applied to a warehouse management system. The method includes: determining the three-dimensional visualization modeling input data of the warehouse digital twin based on the acquired environmental data set of the warehouse space and the operating status data set of the warehouse equipment; performing feature extraction processing on the three-dimensional visualization modeling input data to generate a warehouse feature set, and generating a three-dimensional visualization model of the warehouse digital twin based on the warehouse feature set; calling a spatial layout simulation analysis algorithm to jointly simulate the shelf distribution characteristics, equipment movement path characteristics and cargo access frequency characteristics in the three-dimensional visualization model to generate a spatial layout simulation result set; generating a warehouse operation adjustment strategy set based on the path conflict simulation results and space idle simulation results in the spatial layout simulation result set, and instructing the warehouse control system to perform spatial layout optimization processing based on the warehouse operation adjustment strategy set.

[0006] The present invention also provides a warehouse management system, comprising: a memory for storing program instructions and data; a processor for coupling with the memory and executing instructions in the memory to implement the above method.

[0007] The present invention also provides a computer storage medium comprising instructions, which implement the above method when executed on a processor.

[0008] The embodiment of the present invention determines the modeling input data by combining the storage space environment data and the equipment operation status data, providing a data basis for generating an accurate three-dimensional visualization model of the warehouse digital twin; through feature extraction processing, the three-dimensional visualization model can more realistically reflect the actual situation of the warehouse; then, joint simulation processing can be used to comprehensively analyze the characteristics such as shelf distribution, equipment movement path and frequency of goods access, and explore potential problems; further, the operation adjustment strategy set generated based on the path conflict and space idleness simulation results can effectively guide the warehouse control system to optimize the spatial layout, improve the utilization rate of warehouse space, improve equipment operation efficiency, reduce operation conflicts, and make warehouse operations more efficient and orderly. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0010] Figure 1 A schematic flow chart of the steps of a method for optimizing warehouse space utilization based on digital twins provided in an embodiment of the present invention.

[0011] Figure 2 This is a structural block diagram of a warehouse management system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0012] The technical solutions of the present invention will be described below in conjunction with the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation methods described in the following exemplary embodiments do not represent all implementation methods consistent with the present invention. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present invention. It should be noted that the terms "first", "second", etc. in the specification of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0013] See also Figure 1 , Figure 1 This is a flow chart of a warehouse space utilization optimization method based on digital twins provided in an embodiment of the present invention. The method is applied to a warehouse management system and may further include steps 110 to 140.

[0014] Step 110: Determine the three-dimensional visualization modeling input data of the warehouse digital twin based on the acquired storage space environment data set and storage equipment operation status data set.

[0015] In power battery storage scenarios, the environmental conditions of the storage space and the operating status of the storage equipment are crucial for building an accurate 3D visualization model of the warehouse digital twin. Environmental factors can impact the performance and storage safety of power batteries, while the operating status of the equipment impacts the efficiency and quality of storage operations. Therefore, to obtain the data required for modeling, warehouse management systems integrate multiple data collection methods.

[0016] Based on this, in an optional embodiment, determining the input data for the three-dimensional visual modeling of the warehouse digital twin based on the acquired environmental data set of the warehouse space and the operating status data set of the warehouse equipment includes: Step 111: Collecting an environmental data set uploaded by multiple sensor nodes in the storage space, wherein the environmental data set includes environmental temperature distribution characteristics, humidity distribution characteristics, and light intensity distribution characteristics.

[0017] Within a power battery warehouse, numerous sensor nodes are distributed, labeled A, B, and C. Node A is located in the northeast corner of the warehouse, while node B is in the middle of a row of shelves. These nodes continuously monitor and upload environmental data. Regarding ambient temperature distribution characteristics, sensor nodes in different locations measure different temperatures. For example, node A measures temperature Ta, while node B measures temperature Tb. Since different areas of the warehouse are affected differently by factors such as the external environment and ventilation system, Ta and Tb may differ. Regarding humidity distribution characteristics, node C measures humidity Hc, while node D measures humidity Hd. Locations near warehouse walls or corners may have higher humidity. The same applies to light intensity distribution characteristics. Sensor nodes near lighting fixtures measure different light intensities than those farther away. For example, node E measures light intensity LiE, while node F measures light intensity LiF. The temperature, humidity, and light intensity data from these different nodes are integrated to form an environmental data set.

[0018] Step 112: Acquire an operating status data set of the storage equipment, wherein the operating status data set includes equipment movement trajectory characteristics, equipment energy consumption characteristics, and equipment fault record characteristics.

[0019] Alternatively, operating status data for warehouse equipment such as automated guided vehicles (AGVs) and forklifts is key information. Taking AGVs as an example, different AGVs are identified by M1, M2, and so on. In one example, the movement trajectory of AGV M1 within the warehouse is recorded by a series of coordinate points, from coordinate point (x1, y1) to (x2, y2), then to (x3, y3), and so on. These coordinate points constitute the device's movement trajectory signature. Regarding the device's energy consumption signature, the energy consumption of AGV M1 during operation is related to factors such as travel distance and load weight. For example, if AGV M1 has a load weight of W1 and a travel distance of D1, its energy consumption is E1. The device fault record signature records information related to device faults. For example, if AGV M2 fails at time t with fault type F2, it could be a motor or sensor failure. Based on this, the movement trajectory, energy consumption, and fault record data of these different devices are collected to form a set of operating status data.

[0020] It is worth mentioning that in order to deal with the noise of the equipment movement trajectory in the actual application process, the existing Kalman filter algorithm can be combined to smooth the AGV coordinate sequence in step 112: through the state equation X k =AX k-1 +Bu k and the observation equation Z k =HX k +ω k Iteratively correct coordinate deviations to improve trajectory accuracy.

[0021] Step 113: performing data preprocessing on the environment data set and the operation status data set to generate a preprocessed environment data set and a preprocessed operation status data set.

[0022] Environmental data sets may contain missing or abnormal data. For example, if a sensor node fails to upload temperature data at a certain moment, for example, node G has no temperature data at time t1, the temperature of node G at time t1 can be estimated using an interpolation algorithm by analyzing the temperature data of neighboring nodes (such as nodes F and H) at similar times. Abnormal data, such as the light intensity measured at node I significantly deviating from the normal range, may indicate a sensor failure and require correction or elimination. For operational status data sets, device trajectory data may contain noise. For example, in the trajectory record of AGV M3, some coordinate points may be deviated due to positioning errors, requiring filtering algorithms to remove noise. Device energy consumption data may contain errors due to the accuracy of the measuring equipment and require calibration to ensure data accuracy. It can be understood that after the above preprocessing operations, a preprocessed environmental data set and a preprocessed operational status data set are obtained.

[0023] When implementing the above technical solution, technical personnel in the relevant field can supplement and optimize the environmental data in step 113 based on the linear interpolation method in the prior art. For example, when the temperature data of node G is missing, the missing value can be estimated by using the measurement values TaF and TaH of adjacent nodes F and H at time t1, using the formula TaG=(TaF+TaH) / 2, thereby ensuring data integrity.

[0024] Step 114: Perform spatiotemporal alignment processing on the pre-processed environment data set and the pre-processed operating status data set to generate a spatiotemporal correlation data sequence, and use the spatiotemporal correlation data sequence as input data for three-dimensional visualization modeling of the warehouse digital twin.

[0025] The purpose of spatiotemporal alignment is to accurately align environmental data and operational status data across time and space. For example, at time t2, AGV M4 is located at coordinates (x4, y4). The environmental data at that location (such as temperature T4, humidity H4, and light intensity Li4) must be correlated with AGV M4's operational status data (such as energy consumption E4 and movement speed V4). By matching timestamps and spatial coordinate information, the two preprocessed data types are integrated into a spatiotemporal correlation data sequence. This sequence comprehensively reflects the warehouse environment and equipment operation at different times and locations, serving as input data for the 3D visualization modeling of the warehouse digital twin.

[0026] Step 120: Perform feature extraction processing on the three-dimensional visualization modeling input data to generate a warehouse feature set, and generate a three-dimensional visualization model of the warehouse digital twin based on the warehouse feature set.

[0027] After acquiring the spatiotemporal correlation data sequence, valuable features can be extracted to construct a warehouse feature set, and then a three-dimensional visualization model can be generated to provide an intuitive display and analysis basis for warehouse management. Based on this, in an optional embodiment, the feature extraction processing of the three-dimensional visualization modeling input data to generate a warehouse feature set includes: Step 121: Calling a multi-scale spatiotemporal feature extraction model to perform time dimension sliding window segmentation on the spatiotemporal correlation data sequence to generate multiple time segment subsequences.

[0028] Among them, the multi-scale spatiotemporal feature extraction model processes the spatiotemporal correlation data sequence according to the set time window parameters. The time window size is set to T1, and the time interval of each sliding is T2. Starting from the starting position of the data sequence, the data intercepted by the first time window forms the first time segment subsequence S1, which contains data from time t0 to t0+T1; then the window slides T2, and the data intercepted by the second time window forms the second time segment subsequence S2, which contains data from time t0+T2 to t0+T1+T2, and so on. Different time segment subsequences cover warehouse activity information in different time periods. For example, a subsequence may contain data during the period of concentrated equipment operation, and another subsequence may contain data during the period of stable environmental parameters.

[0029] Step 122: Perform spatial topological structure analysis on each time segment subsequence to extract spatial topological association features, wherein the spatial topological association features reflect the spatial distance relationship and the goods stacking relationship between adjacent shelves.

[0030] For each time segment subsequence, the spatial topology of the warehouse can be analyzed. Taking the shelves as an example, shelves R1, R2, etc. have spatial positional relationships with each other. The spatial distance between adjacent shelves R1 and R2 can be calculated using coordinates. The coordinates of R1 are (xR1, yR1, zR1), and the coordinates of R2 are (xR2, yR2, zR2). The distance between them is d = √[(xR2-xR1)] 2 + (yR2-yR1) 2 + (zR2-zR1) 2 Regarding the stacking relationship of goods, within a certain time segment, the stacking of goods on shelf R1 may be carried out according to preset rules, such as layered stacking, with power batteries of model P1 placed on the first layer and power batteries of model P2 placed on the second layer. By analyzing the above spatial relationships and the stacking of goods, spatial topological association features are extracted.

[0031] Step 123: performing device operation mode recognition on each time segment subsequence, and extracting device operation cycle features, wherein the device operation cycle features include device movement acceleration change features and device turning frequency features.

[0032] For each time-segment subsequence of device operation data, the device's operating mode is identified and relevant features are extracted. Taking an AGV as an example, its acceleration variation characteristics are analyzed. In time-segment subsequence S3, AGV M5 starts from a standstill and gradually increases its acceleration. During time period t3-t4, the acceleration changes from a1 to a2. By analyzing and calculating the speed and time data (acceleration changes can be analyzed by the ratio of speed change to time change), the acceleration variation is determined. Regarding the device's turning frequency characteristics, AGV M5 may perform multiple turning operations during the same time segment. The time and angle information for each turn is recorded, and the relationship between the number of turns and the duration of the time segment is calculated to determine the device's turning frequency characteristics.

[0033] Step 124: performing feature fusion processing on the spatial topology association feature and the equipment operation cycle feature to generate a spatiotemporal fusion feature vector, and merging the spatiotemporal fusion feature vectors of all time segment subsequences into the storage feature set.

[0034] The spatial topology association features and the equipment operation cycle features are fused. For example, the feature value representing the distance between adjacent shelves is weightedly fused with the feature value of the change in equipment movement acceleration. The feature value of the distance between adjacent shelves in the spatial topology association features is Df, and the feature value of the change in equipment movement acceleration is Af. The weight coefficients w1 and w2 (w1+w2=1) are set, and the fused feature value is F1=w1*Df+w2*Af (one of the exemplary fusion methods, those skilled in the art will know how to make adaptive adjustments according to needs in actual application). Similar fusion operations are also performed on the cargo stacking relationship features and the equipment turning frequency features. Based on this, all the above fused features / feature values are combined into a spatiotemporal fusion feature vector, and then the spatiotemporal fusion feature vectors generated by all time segment subsequences are merged together to form a warehousing feature set.

[0035] In addition, for the feature fusion weight setting in step 124, the values of w1 and w2 can be automatically determined based on historical data through machine learning (such as random forest feature importance analysis). For example, the shelf distance feature and the equipment acceleration feature are input into the training model, and the weight coefficient is allocated according to the impact of the feature on warehousing efficiency, thereby avoiding the subjectivity of manual experience.

[0036] In an embodiment of the present invention, the core algorithm type of the multi-scale spatiotemporal feature extraction model is based on the hierarchical characteristics of spatiotemporal data, combining time series analysis, spatial graph convolution and feature fusion mechanisms to achieve joint modeling of the dynamic warehouse environment and equipment behavior. In the time dimension, the model uses a sliding window segmentation algorithm to divide the spatiotemporal correlation data sequence into multiple time segment subsequences. The window size T1 and the sliding step size T2 must be set to match the warehouse operation cycle (such as the equipment inspection cycle and the peak period of cargo storage and retrieval) to ensure that the subsequences cover the complete operation unit. Each time segment subsequence serves as input for subsequent processing. Its data structure contains multimodal fields such as timestamp, spatial coordinates, environmental parameters and equipment status, forming a tensor form to adapt to the deep learning framework.

[0037] At the spatial topology analysis layer, the model incorporates a graph neural network (GNN) to encode the spatial relationships between shelves. A topological graph is constructed with shelves as nodes and adjacent relationships as edges. Graph convolutions aggregate k-hop neighbor information to extract spatial distance features and stacking pattern characteristics of shelf groups. For example, the GNN layer converts the 3D coordinate difference between shelves R1 and R2 in step 122 into edge attributes and embeds them into a high-dimensional vector. Node features are generated based on the stacking rules (such as tiered storage strategies) used to generate node features. The final output is a spatial topology correlation feature vector reflecting the global warehouse layout. The output of this layer is strictly bound to the time segment subsequence, ensuring that the spatial features are correlated with dynamic changes within the time window.

[0038] Furthermore, the device operation mode recognition layer utilizes a combined architecture of a temporal convolutional network (TCN) and an attention mechanism to process time-dependent data such as device movement trajectories and energy consumption series. The TCN's multi-layer dilated convolutional structure captures the long-term and short-term dependencies between periodic features such as device acceleration changes and turning frequency. For example, after filtering the AGV's acceleration curve through the convolution kernel in step 123, its peaks and fluctuation patterns are extracted as cyclical features of the device's operation. The attention mechanism further assigns higher weights to key time points (such as periods of rapid acceleration and frequent turning), enhancing the model's sensitivity to abnormal operating conditions. The output of this layer is dimensionally aligned with the spatial topological features, providing temporal semantic information for subsequent fusion.

[0039] The feature fusion layer then uses a cross-modal attention mechanism to dynamically assign weights to spatial topological association features and equipment operation cycle features. Specifically, this layer calculates the similarity matrix of the two feature vectors to generate spatial-temporal attention weights, which are then used to weightedly concatenate the features. For example, during periods of high-frequency access to goods, the model automatically enhances the weight of equipment movement trajectory features; whereas in areas where environmental parameters suddenly change, the contribution of spatial topological features is emphasized. The fused spatiotemporal fusion feature vector encompasses both static factors such as shelf distribution and ambient thermal conditions, as well as dynamic indicators such as equipment behavior and operational efficiency. This is consistent with the weighted fusion logic described in step 124, and the weight coefficients are optimized through end-to-end training, avoiding the subjectivity of manual settings. The final output warehouse feature set serves as the input to the 3D reconstruction module, driving the visualization model to accurately reflect the geometric structure of the warehouse entity, equipment movement, and environmental status.

[0040] In another optional embodiment, generating a three-dimensional visualization model of the warehouse digital twin based on the warehouse feature set includes: Step 125: Input the spatiotemporal fusion feature vector into a three-dimensional reconstruction module, and reconstruct the three-dimensional geometric structure of the storage space according to the spatial topological association features in the spatiotemporal fusion feature vector.

[0041] In this step, the spatiotemporal fusion feature vector is input into the 3D reconstruction module, and the module operates according to the spatial topological association features therein. For example, the position and shape of the shelves in 3D space are determined based on information such as the spatial distance relationship between adjacent shelves and the stacking relationship of goods. If the spatial topological association features indicate that shelves R1 and R2 are parallel and adjacent, with a distance of d1, then in the 3D reconstruction, the 3D models of the two shelves will be constructed according to this relationship. The stacking relationship of goods determines how the goods are placed on the shelves, such as placing different types of power batteries in layers or areas. The 3D reconstruction module will construct a virtual model of the goods at the corresponding position, thereby gradually reconstructing the 3D geometric structure of the storage space.

[0042] Step 126: Rendering a real-time animation of the device movement trajectory in the three-dimensional geometric structure based on the device operation cycle feature in the spatiotemporal fusion feature vector.

[0043] In this step, the device's operational cycle characteristics within the spatiotemporal fusion feature vector are used to render a real-time animation of the device's movement trajectory within the constructed 3D geometry. For example, the device's position and motion state at different moments can be determined based on the characteristics of its acceleration and turning frequency. For example, if the AGV M6 has an acceleration of a3 and a turning frequency of f3 within a time period, the position coordinates of the AGV M6 at each point in time are calculated based on these characteristics and related common techniques. The AGV M6's movement trajectory is then plotted within the 3D geometry according to these coordinates, and its motion is displayed through animation, allowing users to intuitively visualize the device's operation within the warehouse.

[0044] Step 127: Generate an environmental thermal layer by overlaying it in the three-dimensional geometric structure based on the ambient temperature distribution characteristics and humidity distribution characteristics in the spatiotemporal correlation data sequence.

[0045] Optionally, an environmental thermal map is generated by overlaying it on the three-dimensional geometric structure based on the ambient temperature and humidity distribution characteristics in the spatiotemporal correlation data sequence. For example, for the temperature distribution feature, the temperature data at different locations is mapped to the corresponding positions in three-dimensional space. If the temperature in a certain area of the warehouse is higher, the high temperature is represented by a darker color at the corresponding three-dimensional model location in that area, and the lower temperature areas are represented by a lighter color. The humidity distribution feature is also processed in a similar manner, and the humidity distribution is displayed through color changes or other visualization methods, thereby overlaying a thermal map reflecting the ambient temperature and humidity on the three-dimensional geometric structure.

[0046] Step 128: Overlay and render the three-dimensional geometric structure, the real-time animation, and the environmental thermal layer to generate a three-dimensional visualization model of the warehouse digital twin.

[0047] This step aims to comprehensively process the 3D geometry, real-time animations of equipment movement trajectories, and environmental thermal maps. First, the real-time animation and environmental thermal maps are precisely overlaid on the 3D geometry according to their respective position and time information. The overlaid model is then processed using a rendering algorithm, adjusting parameters such as lighting, color, and transparency to make the entire model more realistic and intuitive. The resulting 3D visualization model of the warehouse digital twin comprehensively displays the warehouse's spatial layout, equipment operation, and environmental conditions, providing powerful visualization support for warehouse management and optimization.

[0048] Step 130: calling a spatial layout simulation analysis algorithm to perform a joint simulation process on the shelf distribution characteristics, equipment movement path characteristics, and cargo access frequency characteristics in the three-dimensional visualization model to generate a set of spatial layout simulation results.

[0049] After generating the 3D visualization model, a spatial layout simulation analysis algorithm is used to jointly simulate key features in the model to evaluate and optimize the warehouse space layout. Based on this, in a preferred embodiment, the spatial layout simulation analysis algorithm is called to jointly simulate the shelf distribution characteristics, equipment movement path characteristics, and cargo access frequency characteristics in the 3D visualization model to generate a set of spatial layout simulation results, including: Step 131: constructing a shelf location topology map based on the shelf distribution characteristics, and marking the spatial coordinates and cargo storage capacity of each shelf in the shelf location topology map.

[0050] Optionally, a shelf location topology map is constructed based on the shelf distribution characteristics in the 3D visualization model. The shelf distribution characteristics include the location information of each shelf, such as shelf R3 located at coordinates (x3, y3, z3). In the topology map, each shelf is represented by a graphical node, and adjacent shelves are connected by lines to represent their topological relationship. For each node, its spatial coordinates (x, y, z) and cargo storage capacity are annotated. For example, the cargo storage capacity of shelf R3 is C3, meaning it can store C3 power batteries. The resulting topology map shows the spatial location relationship and storage capacity of the shelves.

[0051] Step 132: Identify path intersection nodes in the equipment movement path characteristics, and calculate the potential conflict probability of each path intersection node based on the cargo access frequency characteristics.

[0052] Within the device movement path characteristics, the movement paths of different devices are analyzed. For example, the movement paths of AGV M7 and AGV M8 may intersect, and these intersections are referred to as path intersection nodes. During a specific time period, multiple devices operate within the warehouse. The probability of potential collisions is calculated by counting the frequency of different devices appearing at these intersection nodes and the frequency of cargo access. If a path intersection node has a high number of device appearances per unit time and a high frequency of cargo access, the potential collision probability at that node is relatively high. The potential collision probability for each path intersection node is determined using a commonly used statistical analysis algorithm.

[0053] In the calculation of the path conflict probability in step 132, the Poisson distribution model in the prior art can be used, combining the number of equipment passes per unit time λ and the frequency of goods access μ, according to the formula P=1-e -λμΔt Calculate the node conflict probability, where Δt is the time window length, to achieve quantitative evaluation.

[0054] Step 133: Based on the potential conflict probability and the cargo storage capacity, simulate the mobile path conflict scenarios of the equipment under different access frequencies to generate a path conflict simulation result.

[0055] Optionally, the potential conflict probability and cargo storage capacity information can be combined to simulate the movement of devices under different access frequencies. If the potential conflict probability at a certain path intersection node is high and the cargo storage capacity in that area is limited, as the access frequency increases, devices at that node may experience congestion, collisions, and other conflicts. By simulating different access frequency scenarios (such as low-frequency access, medium-frequency access, and high-frequency access), observing the operation of devices at path intersections, and recording information such as the time, number of conflicts, and impact range, path conflict simulation results are generated.

[0056] Step 134: Generate a space idleness simulation result based on the area of the region not covered by the equipment movement path in the shelf location topology diagram and the time distribution characteristics of the frequency of goods access.

[0057] Optionally, the system analyzes the areas uncovered by the equipment movement paths in the shelf location topology and calculates their areas. For example, the boundaries of uncovered areas can be determined using information from the nodes and lines in the topology, allowing for the calculation of their areas. Furthermore, the system considers the temporal distribution of the frequency of access to goods. For example, if access to goods is low during certain time periods, the corresponding shelf areas are more likely to be idle. Combining these two factors, the system generates spatial idleness simulation results, such as determining which areas are idle during different time periods and the degree of idleness.

[0058] Step 135: Merge the path conflict simulation results and the space idleness simulation results into the space layout simulation result set.

[0059] In this step, the path conflict simulation results and the space idleness simulation results are integrated together to form a spatial layout simulation result set. This set comprehensively includes information on equipment movement path conflicts and space idleness, providing detailed data support for the subsequent formulation of warehouse operation adjustment strategies.

[0060] Step 140: Generate a warehouse operation adjustment strategy set according to the path conflict simulation results and the space idle simulation results in the spatial layout simulation result set, and instruct the warehouse control system to perform spatial layout optimization processing based on the warehouse operation adjustment strategy set.

[0061] In practical applications, the information in the spatial layout simulation result set can be used to formulate warehouse operation adjustment strategies to optimize the warehouse space layout. Based on this, in an optional embodiment, the warehouse operation adjustment strategy set is generated based on the path conflict simulation results and space idleness simulation results in the spatial layout simulation result set, including: Step 141: replanning paths for frequently conflicting nodes in the path conflict simulation results to generate a device movement path optimization strategy, wherein the device movement path optimization strategy includes adding backup path nodes and adjusting path priorities.

[0062] Frequently conflicting nodes are identified from the path conflict simulation results. For example, node N1 has a high frequency of conflicts in multiple simulations. Path replanning is performed for these nodes: (1) Alternate path nodes are added, with nodes N1a and N1b added near node N1 as alternate path nodes; (2) Path priorities are adjusted. For different equipment paths passing through node N1, priorities are reallocated based on factors such as the urgency of the equipment's tasks. For example, a higher priority is given to equipment paths performing urgent cargo access tasks to ensure that they can pass through node N1 first, thereby generating an equipment movement path optimization strategy.

[0063] Step 142: Generate a shelf migration strategy based on the area distribution in the idle space simulation result. The shelf migration strategy includes migrating infrequently accessed goods to edge areas and releasing space in the central area.

[0064] In this step, the area distribution in the idle space simulation results is analyzed to determine which areas have the most idle space. For example, the corners of the warehouse have a large amount of idle space, while the central area is frequently accessed. Based on this situation, a shelf migration strategy is developed to move shelves storing less frequently accessed goods to the edge areas. For example, shelf R4, which stores goods that are less frequently accessed, can be moved from the center to the edge areas, freeing up space in the center and improving its utilization efficiency.

[0065] Step 143: combining the device movement path optimization strategy and the shelf migration strategy to generate an operation timing adjustment strategy, wherein the operation timing adjustment strategy includes staggered scheduling of frequently accessed devices and infrequently accessed devices.

[0066] Optionally, combine equipment movement path optimization strategies with shelf migration strategies to develop a strategy for adjusting operational timing. For example, frequently accessed equipment, such as the AGV M9, can be scheduled to operate during designated time periods to avoid concurrent operation with less frequently accessed equipment in certain areas. Less frequently accessed equipment, such as the forklift H1, can be scheduled to operate during alternate time periods to achieve staggered scheduling. This reduces conflicts between equipment and improves overall warehouse efficiency.

[0067] Step 144: sorting the equipment movement path optimization strategy, the shelf migration strategy, and the operation timing adjustment strategy by strategy priority to generate the warehouse operation adjustment strategy set.

[0068] In this step, the importance of each strategy is determined based on the warehouse's actual needs and goals. For example, if path conflicts within the current storage space are severe and impacting overall operational efficiency, the equipment movement path optimization strategy may be given the highest priority. If space utilization is low, with a large amount of idle space, the shelf migration strategy may be given a relatively high priority. The operation sequence adjustment strategy can be prioritized based on equipment operating conditions and cargo access patterns. By comprehensively considering various factors, these three strategies are rationally ranked to form a set of warehouse operation adjustment strategies.

[0069] In yet another optional embodiment, instructing the warehouse control system to perform spatial layout optimization based on the warehouse operation adjustment strategy set includes: Step 145: Convert the device movement path optimization strategy into a device control instruction set, where the control instruction set includes a path node coordinate sequence and device speed control parameters.

[0070] The new path and adjusted path priority determined in the equipment movement path optimization strategy are converted into a set of equipment control instructions that the warehouse control system can recognize and execute. Taking the AGV as an example, the optimized path passes through nodes P1, P2, and P3. The path node coordinate sequence is (xP1, yP1, zP1), (xP2, yP2, zP2), and (xP3, yP3, zP3). At the same time, the equipment speed control parameters are set according to the conditions of different path segments and the equipment performance requirements. For example, a higher speed V1 can be set for sections with wide paths and no interference from other equipment; a lower speed V2 can be set for intersections or narrow areas to ensure safe and efficient operation of the equipment. The above path node coordinate sequence and equipment speed control parameters together constitute the equipment control instruction set.

[0071] Step 146: Generate a shelf adjustment task queue according to the shelf migration strategy, wherein the task queue includes a target shelf identifier, migration target coordinates, and migration priority.

[0072] This step determines the shelves that need to be migrated based on the shelf migration strategy. For example, if shelves R5 and R6 need to be migrated, the target shelf identifiers are R5 and R6. For each target shelf, its migration target coordinates are determined. For example, shelf R5 needs to be migrated to the coordinates (x5n, y5n, z5n), and shelf R6 needs to be migrated to the coordinates (x6n, y6n, z6n). The migration priority is determined based on the urgency of the migration or the impact on warehouse operations. For example, if the goods stored on shelf R5 are about to expire and need to be moved as soon as possible for inventory operations, its migration priority is higher than that of shelf R6. The target shelf identifiers, migration target coordinates, and migration priority information are organized into a shelf adjustment task queue so that the warehouse control system can execute migration tasks in sequence.

[0073] Step 147: Split the job timing adjustment strategy into multiple independent time window control signals, each time window control signal is associated with the operation authority of the target device group and shelf group.

[0074] Optionally, the operation timing adjustment strategy involves the operation arrangement of different equipment and shelves in different time periods. It is divided into multiple independent time window control signals, and each time window corresponds to a time period. For example, time window T1 is from 9 am to 10 am. Within this time window, specific equipment groups, such as AGV group A and stacker group B, and specific shelf groups, such as shelf areas R7-R10, are associated. It is stipulated that within this time window, AGV group A can only perform cargo storage and retrieval operations on shelf area R7-R10, and stacker group B can only perform related operations in this area. In this way, the operating permissions of each equipment group and shelf group in different time periods are clarified to avoid operation conflicts.

[0075] Step 148: Synchronously send the device control instruction set, the shelf adjustment task queue, and the time window control signal to the execution unit of the warehouse control system to trigger a real-time space reconstruction operation.

[0076] In this step, the generated device control instruction set, shelf adjustment task queue, and time window control signal are simultaneously sent to the warehouse control system's execution unit. Upon receiving this information, the execution unit controls the device's path and speed according to the instruction set, executes shelf migration tasks based on the task queue, and allocates device and shelf operation permissions based on the time window control signal. Through these operations, the warehouse space layout changes in real time based on the adjustment strategy, triggering real-time spatial reconstruction operations to optimize the warehouse space layout.

[0077] By applying the embodiments of the present invention, by integrating multiple data to determine the modeling input, a three-dimensional visualization model can be accurately constructed, thereby deeply exploring potential problems and generating effective strategies, thereby optimizing the warehouse space layout and ultimately effectively improving warehouse operation efficiency and resource utilization.

[0078] In an alternative technical solution, after triggering the real-time spatial reconstruction operation, the method further includes: Step 200: continuously collect the optimized storage space environment data set and equipment operation status data set; input the optimized storage space environment data set and equipment operation status data set into the three-dimensional visualization model of the warehouse digital twin to generate a visualization comparison diagram of the optimization effect; extract the path conflict reduction rate and the space utilization improvement rate in the visualization comparison diagram of the optimization effect; when the path conflict reduction rate is lower than the preset threshold or the space utilization improvement rate does not reach the expected target, jump to calling the space layout simulation analysis algorithm, and jointly simulate the shelf distribution characteristics, equipment movement path characteristics and cargo access frequency characteristics in the three-dimensional visualization model to generate a space layout simulation result set, and execute to generate a warehouse operation adjustment strategy set according to the path conflict simulation results and space idle simulation results in the space layout simulation result set.

[0079] After triggering the real-time spatial reconstruction operation, the warehouse environment and equipment operating conditions will change. Continuously collect data on the optimized warehouse space environment, including temperature, humidity, and light intensity, as well as equipment operating status data such as equipment movement trajectories, energy consumption, and fault records. This newly collected data is input into the warehouse digital twin's 3D visualization model, which generates a visual comparison chart of the optimization results based on the new data. In this comparison chart, the path conflict reduction rate is calculated by analyzing the changes in path conflict points. For example, comparing the number of path conflict points before and after optimization within the same time period, with N1 being the number before optimization and N2 being the number after, the path conflict reduction rate can be calculated as: (N1 - N2) / N1. The space utilization improvement rate is calculated by calculating the change in the ratio of the effectively utilized area to the total area of the warehouse space. Preset thresholds and expected targets are set. If the path conflict reduction rate falls below the preset threshold or the space utilization improvement rate does not meet the expected target, the optimization results are unsatisfactory. At this point, the spatial layout simulation analysis algorithm is called back to perform joint simulation processing on the shelf distribution characteristics, equipment movement path characteristics, and cargo access frequency characteristics in the three-dimensional visualization model to generate a new set of spatial layout simulation results. According to the subsequent steps, a set of warehouse operation adjustment strategies is generated and the optimization operation is performed again.

[0080] For the closed-loop optimization triggering condition in step 200, a dual-threshold joint judgment rule can be set according to existing industry standards. When the path conflict reduction rate is less than 30% and the space utilization improvement rate is less than 15%, re-simulation is triggered to avoid misjudgment of a single indicator.

[0081] In another alternative technical solution, after generating the storage operation adjustment strategy set, the method further includes: Step 300: Generate a policy verification test case set based on the warehouse operation adjustment policy set, the test case set including different equipment load scenarios and cargo access pressure scenarios; load the test case set in the three-dimensional visualization model of the warehouse digital twin, and perform a virtual stress test; obtain the system response delay data and space allocation anomaly data in the virtual stress test results; when the system response delay data exceeds the tolerance threshold or the space allocation anomaly data reaches the warning level, adjust the weight of the warehouse operation adjustment policy set, generate the adjusted policy version and update it to the warehouse control system.

[0082] In this embodiment, a set of policy adjustments is made based on the warehouse operation, and a variety of different test scenarios are designed to generate a set of policy verification test cases. For example, equipment load scenarios are set, including light load, medium load, and heavy load conditions, to simulate the operation of goods of different quantities and weights stored on the equipment. The cargo access pressure scenario can set different conditions such as high-frequency access, medium-frequency access, and low-frequency access to test the effectiveness of the policy under different cargo demands. The above test case set is loaded into the three-dimensional visualization model of the warehouse digital twin, and a virtual stress test is performed. During the test, the system response delay data is recorded, that is, the time interval from the issuance of the operation instruction to the start of the equipment to perform the operation; and the space allocation abnormality data is monitored, such as whether there is a situation where the equipment cannot reach the specified location or the cargo storage location is unreasonable. The tolerance threshold and warning level are set. When the system response delay data exceeds the tolerance threshold or the space allocation abnormality data reaches the warning level, it indicates that there is a problem with the current policy. Adjust the weights of each strategy in the warehouse operation adjustment strategy set, such as increasing the weight of the equipment movement path optimization strategy and reducing the weight of the operation timing adjustment strategy, etc., generate the adjusted strategy version, and update it to the warehouse control system to improve the warehouse operation strategy.

[0083] In some independent embodiments, after instructing the warehouse control system to perform spatial layout optimization based on the warehouse operation adjustment strategy set, the method further includes any one of step 400, step 500, and step 600: Step 400: Collect the optimized storage equipment operation energy consumption data set and task completion time data set in real time, normalize the storage equipment operation energy consumption data set and task completion time data set based on the preset energy consumption-time weight coefficient, and generate an equipment operation performance indicator set; compare the equipment operation performance indicator set with the historical optimal performance threshold set, and if the current equipment operation performance indicator is lower than the corresponding threshold, adjust the path priority weight and shelf migration frequency weight in the storage operation adjustment strategy set based on the gradient descent algorithm; regenerate the equipment movement path optimization strategy and shelf migration strategy according to the adjusted weights to obtain an updated strategy set, and synchronize the updated strategy set to the decision unit of the warehouse control system to trigger incremental spatial layout iterative optimization.

[0084] Optionally, real-time data collection is performed on optimized warehouse equipment operating energy consumption data, recording the energy consumption of each device during operation. For example, the energy consumption of AGV M10 during a task is E10. Task completion time data is also collected, such as the time it takes AGV M10 to complete the task, T10. These two sets of data are normalized based on preset energy-time weighting coefficients. For example, the energy weighting coefficient is set to wE and the time weighting coefficient is set to wT (wE + wT = 1). A set of equipment operating performance indicators is generated using commonly used calculation methods (e.g., combining energy and time data according to weights). This set is compared with a set of historical optimal performance thresholds, which records the best performance data for equipment in the past. If the current equipment operating performance indicator falls below the corresponding threshold, it indicates that the equipment's operating efficiency needs to be improved. Based on the gradient descent algorithm, the path priority weights and shelf migration frequency weights in the warehouse operation adjustment strategy set are adjusted. For example, if equipment consumes high energy and takes a long time to complete tasks on a certain path, the priority weight of that path is appropriately lowered. If shelf migration significantly impacts overall performance, the shelf migration frequency weight is adjusted. Based on the adjusted weights, the equipment movement path optimization strategy and shelf migration strategy are regenerated to form an updated strategy set, which is synchronized to the decision-making unit of the warehouse control system. Upon receiving the updated strategy set, the decision-making unit triggers incremental spatial layout iterative optimization to further improve the warehouse space layout.

[0085] Step 500: Obtain the optimized shelf space occupancy distribution map and equipment path coverage distribution map, perform pixel-level grid matching analysis on the shelf space occupancy distribution map and the equipment path coverage distribution map, identify the utilization saturation characteristics of the overlapping area and the idle characteristics of the non-overlapping area; construct a multi-objective optimization function based on the utilization saturation characteristics and the idle characteristics, the multi-objective optimization function includes maximizing the weighted sum of the path coverage and the shelf space occupancy rate, and the constraints include the minimum turning radius of the equipment and the upper limit of the shelf load; call the particle swarm optimization algorithm to solve the multi-objective optimization function, generate a Pareto optimal solution set, and sort the solutions in the solution set according to the path conflict suppression rate and space utilization improvement rate of each solution; map the highest-ranked Pareto optimal solution to shelf position coordinate adjustment instructions and equipment path node update instructions, and send them to the warehouse control system for real-time calibration.

[0086] In this embodiment, an optimized shelf space occupancy distribution map and equipment path coverage distribution map are obtained. The shelf space occupancy distribution map shows the proportion of each shelf area occupied by goods, and the equipment path coverage distribution map shows the coverage range of the equipment's operation path in the warehouse. Pixel-level grid matching analysis is performed on these two distribution maps, and the distribution maps are divided into small grid units, and the situation of each grid unit in the two maps is compared. In the overlapping area, it is analyzed whether there is a utilization saturation feature. For example, if the shelf space in a certain grid unit is completely occupied by goods, and the equipment frequently passes through the area, it can be considered that the utilization of the area is saturated. In the non-overlapping area, idle features are identified, such as the shelf space of some grid units is not fully utilized, and the equipment rarely passes through.

[0087] Based on the above characteristics, a multi-objective optimization function is constructed. The goal is to maximize the weighted sum of path coverage and shelf space occupancy. Specifically, the goal is to ensure that the equipment path covers more area while fully utilizing shelf space. Constraints are set, such as a minimum turning radius for the equipment to ensure that the equipment does not become unable to operate normally due to a small turning radius, and an upper limit on the shelf load to ensure that the shelves are not damaged by overloaded cargo. A particle swarm optimization algorithm is used to solve this multi-objective optimization function. The particle swarm optimization algorithm simulates the foraging behavior of a flock of birds to find the optimal solution. After the algorithm runs, a Pareto-optimal solution set is generated. Each solution in the Pareto-optimal solution set meets the multi-objective optimization requirements to varying degrees. Solutions are ranked based on their path conflict suppression rate and space utilization improvement rate. Solutions with higher path conflict suppression rates and higher space utilization improvement rates are ranked higher. The highest-ranked Pareto-optimal solution is mapped into shelf position coordinate adjustment instructions and equipment path node update instructions. For example, based on the optimal solution, a shelf needs to be moved to a new coordinate position, or nodes need to be added or removed from a certain equipment path. These instructions are issued to the warehouse control system for real-time calibration, further optimizing the warehouse space layout.

[0088] Among them, for the particle swarm optimization algorithm in step 500, the standard multi-objective optimization framework of the prior art can be referred to, and the fitness function is set to F=α*Coverage+β*Occupancy (α+β=1), and the device turning radius r≥r in the constraint condition min Determined by AGV mechanical parameters, shelf load W≤W max Limited by the shelf specifications, the Pareto optimal solution is searched by setting the typical parameter combination of the number of particles N = 100 and the number of iterations T = 500.

[0089] Step 600: Monitor the real-time vibration amplitude data and shelf tilt angle data of the storage equipment during the execution of the optimization strategy. When the vibration amplitude exceeds the safety threshold or the tilt angle deviates from the preset tolerance range, generate an abnormal equipment operation alarm signal; locate the abnormal equipment and associated shelves according to the alarm signal, freeze the current operation instructions and start the emergency simulation module, and simulate the abnormal diffusion path and potential collision area in the three-dimensional visualization model; recalculate the path conflict simulation results and the space idle simulation results based on the simulation results, and generate an emergency adjustment strategy set, which includes equipment emergency braking instructions, shelf reinforcement instructions and backup path activation instructions; insert the emergency adjustment strategy set into the first priority of the original storage operation adjustment strategy queue, triggering the abnormal response unit of the storage control system to perform safety isolation and recovery operations.

[0090] During the optimization process of the storage equipment, the real-time vibration amplitude data of the equipment and the tilt angle data of the shelf are continuously monitored. For example, the data obtained by the sensors installed on the equipment and the shelf is that the vibration amplitude of device A is V1 and the tilt angle of shelf GS is θ1. The safety threshold and the preset tolerance range are set. When the vibration amplitude of the equipment exceeds the safety threshold, such as V1 is greater than the set safety vibration amplitude threshold Vth, or the tilt angle of the shelf deviates from the preset tolerance range, such as θ1 exceeds the set angle range [θ min ,θ max], generating an abnormal equipment operation alarm signal. Based on the alarm signal, the system's positioning function is used to determine the abnormal equipment and the associated shelves. For example, device A and shelf GS are located by equipment number and shelf ID. The currently executing operation instructions are immediately frozen to prevent the abnormal situation from further deteriorating. The emergency simulation module is launched to simulate the diffusion path and potential collision areas of the abnormal situation in the 3D visualization model of the warehouse digital twin. For example, based on the running direction and speed of device A and the tilt state of shelf GS, the areas where collisions may occur and the range of the potential impact of the abnormal situation are simulated. Based on the simulation results, the path conflict simulation results and the space idleness simulation results are recalculated. If the simulation finds that a certain area may increase path conflicts due to equipment abnormalities and shelf tilt, the path conflict simulation results are updated; if some areas cannot be used normally due to abnormal conditions, the space idleness simulation results are adjusted. Based on the new simulation results, an emergency adjustment strategy set is generated, including an equipment emergency braking instruction to immediately stop device A; a shelf reinforcement instruction to reinforce shelf GS; and an alternative path activation instruction to activate an alternative path to avoid the affected area. The emergency adjustment strategy set is inserted into the first priority of the original warehouse operation adjustment strategy queue. After the abnormal response unit of the warehouse control system receives the strategy set, it performs safety isolation and recovery operations to ensure that the warehouse operation can continue safely and stably.

[0091] In the timing coordination of the emergency strategy and the original strategy in step 600, the emergency instruction can be marked as the highest priority ISR based on the real-time operating system priority preemption mechanism of the existing technology, and the current time window control signal can be paused in the execution unit and the emergency braking instruction can be inserted. After the exception is resolved, the original queue execution can be resumed.

[0092] In addition, to address the issue of dimensional unification throughout the entire process, the AGV coordinates (meters) in step 112, the temperature (degrees Celsius) in step 113, the shelf migration distance (meters) in step 142, and the energy consumption (kilowatt-hours) in step 400 can be standardized with reference to the International System of Units. Sensor calibration (such as using the ISO / IEC 17025 standard) can be performed at the data acquisition end to ensure spatial reference consistency between environmental data and equipment status data, thereby eliminating modeling errors caused by dimensional mixing.

[0093] As a result, technicians in this field can clarify the algorithm implementation details, quantify evaluation indicators and ensure system synergy, and ultimately achieve high-precision modeling and optimization of warehouse digital twins.

[0094] In summary, the embodiment of the present invention determines the modeling input data by combining the storage space environment data and the equipment operation status data, providing a data basis for generating an accurate three-dimensional visualization model of the warehouse digital twin; through feature extraction processing, the three-dimensional visualization model can more realistically reflect the actual situation of the warehouse; then, joint simulation processing can be used to comprehensively analyze the characteristics such as shelf distribution, equipment movement path and cargo access frequency to explore potential problems; further, the operation adjustment strategy set generated based on the path conflict and space idle simulation results can effectively guide the warehouse control system to optimize the spatial layout, improve the utilization rate of warehouse space, improve equipment operation efficiency, reduce operation conflicts, and make warehouse operations more efficient and orderly.

[0095] Further, Figure 2 The structure block diagram of the warehouse management system 300 is shown, which includes: a memory 310 for storing program instructions and data; a processor 320 for coupling with the memory 310 and executing the instructions in the memory 310 to implement the above method.

[0096] Furthermore, a computer storage medium is provided, comprising instructions, which implement the above method when executed on a processor.

[0097] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A warehouse space utilization optimization method based on digital twins, characterized in that: include: Determine the input data for the 3D visualization modeling of the warehouse digital twin based on the acquired storage space environmental data set and storage equipment operating status data set; Performing feature extraction processing on the three-dimensional visualization modeling input data to generate a warehouse feature set, and generating a three-dimensional visualization model of the warehouse digital twin based on the warehouse feature set; Invoking a spatial layout simulation analysis algorithm to perform a joint simulation process on shelf distribution characteristics, equipment movement path characteristics, and cargo access frequency characteristics in the three-dimensional visualization model to generate a set of spatial layout simulation results; A warehouse operation adjustment strategy set is generated according to the path conflict simulation results and the space idle simulation results in the spatial layout simulation result set, and a warehouse control system is instructed to perform spatial layout optimization processing based on the warehouse operation adjustment strategy set.

2. The method according to claim 1, characterized in that The step of determining the input data for the three-dimensional visualization modeling of the warehouse digital twin based on the acquired storage space environment data set and storage equipment operation status data set includes: Collecting an environmental data set uploaded by a plurality of sensor nodes in the storage space, wherein the environmental data set includes environmental temperature distribution characteristics, humidity distribution characteristics, and light intensity distribution characteristics; Acquire an operating status data set of the storage equipment, wherein the operating status data set includes equipment movement trajectory characteristics, equipment energy consumption characteristics, and equipment fault record characteristics; Performing data preprocessing on the environment data set and the operation status data set to generate a preprocessed environment data set and a preprocessed operation status data set; The pre-processing environment data set and the pre-processing operation status data set are subjected to spatiotemporal alignment processing to generate a spatiotemporal correlation data sequence, and the spatiotemporal correlation data sequence is used as input data for three-dimensional visualization modeling of the warehouse digital twin.

3. The method according to claim 2, characterized in that The performing feature extraction processing on the three-dimensional visual modeling input data to generate a storage feature set includes: Calling a multi-scale spatiotemporal feature extraction model to perform time dimension sliding window segmentation on the spatiotemporal correlation data sequence to generate multiple time segment subsequences; Performing spatial topological structure analysis on each time segment subsequence to extract spatial topological correlation features, wherein the spatial topological correlation features reflect the spatial distance relationship between adjacent shelves and the stacking relationship of goods; Performing device operation mode recognition on each time segment subsequence to extract device operation cycle features, wherein the device operation cycle features include device movement acceleration change features and device turning frequency features; The spatial topology association feature and the equipment operation cycle feature are subjected to feature fusion processing to generate a spatiotemporal fusion feature vector, and the spatiotemporal fusion feature vectors of all time segment subsequences are merged into the storage feature set.

4. The method according to claim 3, characterized in that Generating a three-dimensional visualization model of the warehouse digital twin based on the warehouse feature set includes: Inputting the spatiotemporal fusion feature vector into a three-dimensional reconstruction module, and reconstructing the three-dimensional geometric structure of the storage space according to the spatial topological correlation features in the spatiotemporal fusion feature vector; Rendering a real-time animation of the device movement trajectory in the three-dimensional geometric structure based on the device operation cycle characteristics in the spatiotemporal fusion feature vector; Generate an environmental thermal layer by superimposing it on the three-dimensional geometric structure according to the environmental temperature distribution characteristics and humidity distribution characteristics in the spatiotemporal correlation data sequence; The three-dimensional geometric structure, the real-time animation and the environmental thermal layer are layered and rendered to generate a three-dimensional visualization model of the warehouse digital twin.

5. The method according to claim 1, wherein The calling of the spatial layout simulation analysis algorithm performs a joint simulation process on the shelf distribution characteristics, equipment movement path characteristics, and cargo access frequency characteristics in the three-dimensional visualization model to generate a set of spatial layout simulation results, including: Constructing a shelf location topology map based on the shelf distribution characteristics, and marking the spatial coordinates and cargo storage capacity of each shelf in the shelf location topology map; Identifying path intersection nodes in the equipment movement path characteristics, and calculating a potential conflict probability of each path intersection node based on the cargo access frequency characteristics; Simulating, based on the potential conflict probability and the cargo storage capacity, a mobile path conflict scenario of the device under different access frequencies, and generating a path conflict simulation result; Generate a space idleness simulation result based on the area of the area not covered by the equipment movement path in the shelf position topology map and the time distribution characteristics of the frequency of goods access; The path conflict simulation results and the space idleness simulation results are combined into the space layout simulation result set.

6. The method according to claim 5, characterized in that The step of generating a warehouse operation adjustment strategy set according to the path conflict simulation results and the space idleness simulation results in the spatial layout simulation result set includes: Replanning paths for frequently conflicting nodes in the path conflict simulation results to generate a device movement path optimization strategy, wherein the device movement path optimization strategy includes adding backup path nodes and adjusting path priorities; Generating a shelf migration strategy based on the area distribution of the idle space simulation results, wherein the shelf migration strategy includes migrating infrequently accessed goods to edge areas and releasing space in the central area; Combining the equipment movement path optimization strategy and the shelf migration strategy, generating an operation timing adjustment strategy, wherein the operation timing adjustment strategy includes staggered scheduling of frequently accessed equipment and infrequently accessed equipment; The equipment movement path optimization strategy, the shelf migration strategy, and the operation timing adjustment strategy are sorted by strategy priority to generate the warehouse operation adjustment strategy set.

7. The method according to claim 6, characterized in that The instructing the warehouse control system to perform spatial layout optimization based on the warehouse operation adjustment strategy set includes: Converting the device movement path optimization strategy into a device control instruction set, wherein the control instruction set includes a path node coordinate sequence and a device speed control parameter; Generate a shelf adjustment task queue according to the shelf migration strategy, wherein the task queue includes a target shelf identifier, a migration target coordinate, and a migration priority; Splitting the job timing adjustment strategy into multiple independent time window control signals, each time window control signal is associated with the operation authority of the target device group and shelf group; Synchronously sending the device control instruction set, the shelf adjustment task queue, and the time window control signal to the execution unit of the warehouse control system to trigger a real-time space reconstruction operation; After triggering the real-time space reconstruction operation, the method further includes: Continuously collect optimized storage space environment data sets and equipment operation status data sets; Inputting the optimized storage space environment data set and equipment operation status data set into the three-dimensional visualization model of the warehouse digital twin to generate a visualization comparison chart of the optimization effect; Extracting the path conflict reduction rate and space utilization improvement rate from the optimization effect visualization comparison diagram; When the path conflict reduction rate is lower than a preset threshold or the space utilization improvement rate does not reach the expected target, the process jumps to calling the space layout simulation analysis algorithm, performs joint simulation processing on the shelf distribution characteristics, equipment movement path characteristics and cargo access frequency characteristics in the three-dimensional visualization model, generates a space layout simulation result set, and executes to generate a warehouse operation adjustment strategy set based on the path conflict simulation results and space idle simulation results in the space layout simulation result set.

8. The method according to claim 1, characterized in that After generating the storage operation adjustment strategy set, the method further includes: Generate a strategy verification test case set based on the warehouse operation adjustment strategy set, wherein the test case set includes different equipment load scenarios and cargo storage and access pressure scenarios; Loading the test case set into the three-dimensional visualization model of the warehouse digital twin and performing a virtual stress test; Obtain system response delay data and space allocation exception data from virtual stress test results; When the system response delay data exceeds the tolerance threshold or the space allocation abnormality data reaches the warning level, the weight of the warehouse operation adjustment strategy set is adjusted, and an adjusted strategy version is generated and updated to the warehouse control system.

9. A warehouse management system, characterized in that: include: Memory, used to store program instructions and data; A processor, coupled to a memory, and configured to execute instructions in the memory to implement the method according to any one of claims 1 to 8.

10. A computer storage medium, characterized in that The method comprises instructions which, when executed on a processor, implement the method according to any one of claims 1 to 8.

Images