CN-121544181-B - Digital twin warehouse management method and system based on three-dimensional real-time modeling

CN121544181BCN 121544181 BCN121544181 BCN 121544181BCN-121544181-B

Abstract

The invention belongs to the technical field of digital warehouse management, and particularly relates to a digital twin warehouse management method and system for three-dimensional real-time modeling. Initializing storage environment parameters, collecting and accessing multi-type real-time data, constructing a three-dimensional state model based on point cloud and vector data, evaluating modeling errors and executing fitting calibration, and finally forming an iteratively updatable digital twin model. The system comprises modules such as environment initialization, data acquisition, time synchronization, three-dimensional modeling, error measurement and correction and the like, and can realize dynamic mapping of a space structure, logistics behaviors and equipment states. By introducing a unified time source synchronization mechanism and a residual error optimization correction strategy, the modeling precision, response efficiency and historical traceability of the digital twin system are improved, and the method is suitable for intelligent warehouse management under a high dynamic scene.

Inventors

LI TAOTAO
SHAN KAIYUAN
SHEN MENGJIE
ZHU WEN
Zuo Minna

Assignees

杭州硕泰科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260119

Claims (6)

1. The digital twin warehouse management method for three-dimensional real-time modeling is characterized by comprising the following steps of: S1, initializing warehouse environment parameters, namely, establishing association mapping corresponding to initial state information and an actual warehouse environment, wherein the association mapping comprises a three-dimensional position of a warehouse entity, a shelf number, equipment coordinates, an entrance coordinate and a loading and unloading operation area; S2, acquiring and accessing multi-type real-time data, wherein the multi-type real-time data comprise data of cargo warehouse entry, warehouse exit, transportation, storage, unloading and equipment running states, and aligning and correcting the acquisition time stamp through a synchronous mechanism based on a unified time source; S3, constructing a three-dimensional real-time state model, wherein the model is based on point cloud and vector data, and is combined with a dynamic data set subjected to history state restoration processing to fit the current modeling state to form a space structure constraint condition; s4, modeling error evaluation and fitness judgment, namely calculating modeling errors of the three-dimensional state model by using an error measurement mode after modeling is completed, and judging whether the modeled fitness meets the standard or not according to a preset threshold; S4.1, if the modeling error is within a set threshold, updating a mapping data cache, and synchronously writing the mapping data cache into a modeling database to serve as a digital model snapshot at the current moment; S4.2, if the modeling error exceeds a set threshold, counting a distribution function of the modeling error under different data sources, and identifying two source areas with the largest error as key areas of reverse correction, wherein parameters of the distribution function comprise a comprehensive calculation result of a space coordinate difference value, a time delay and a data fitting residual error, and are ordered according to area weights so as to lock a key area with the largest influence on the error; S5, performing reverse fitting and historical residual error calibration, namely extracting historical input data and modeling state records of the heavy point region in a set time window, utilizing residual errors between a current modeling result and the historical records, adjusting the position and attribute information of a model node through the minimum residual error, generating a corrected modeling result and outputting the corrected modeling result; And S6, outputting the correction model and executing error verification iteration, namely taking the modeling result after residual error minimization optimization as the latest three-dimensional model at the current moment, and carrying out comparison check on the space position, the time stamp and the three-dimensional index of the article identification again with the current state of the real warehousing system.
2. The method of claim 1, wherein in step S1, the initialization operation is triggered based on a start signal of a warehouse system, and comprises automatic loading identification of equipment coordinates, static mapping marks of shelf numbers, and system preset correction of access coordinates and loading and unloading operation areas.
3. The method of claim 1, wherein in step S2, the multi-type real-time data includes at least one set of control instruction data from sensor data, RFID read-write data and automated handling equipment, and is written into the middleware cache interface in a unified format.
4. The method for digital twin warehouse management based on three-dimensional real-time modeling as defined in claim 1, wherein in step S6, if all indexes meet the preset error threshold requirement, the state snapshot is updated and then the next round of data access and modeling flow is entered, so that dynamic real-time iteration of the digital twin system is realized.
5. The method of claim 1, wherein in step S5, the inverse fitting and the historical residual calibration are performed based on a set fixed time segment, the time segment is 3 minutes long, the time segments are processed according to a preset step sequence, and modeling residual of the modeling node is recalculated in each segment to form a residual comparison table before and after optimization.
6. A digital twin warehouse management system modeled in three dimensions in real time for implementing the method according to any of claims 1 to 5, comprising: The environment initialization module is used for collecting initial warehouse environment parameters and establishing a corresponding relation with an actual scene; The data acquisition module is used for acquiring data, including unified acquisition, screening and deduplication operations on cargo circulation, equipment scheduling, environmental parameters, sensor sampling and system events; The time synchronization module is used for establishing a time sequence, performing time alignment on all accessed data sources, and performing synchronization processing and data association through unifying the time stamp dimension; The three-dimensional modeling module is used for receiving the time-series data and generating a real-time three-dimensional modeling scene through calculation rules; the error measurement module is used for carrying out error calculation on time deviation, space deviation and object deviation between the modeling model and the actual scene, and outputting residual errors to feed back modeling fitting degree; the correction module is used for comparing the precision of key areas in the modeling model based on the error measurement result and dynamically correcting the areas with modeling errors exceeding a preset threshold value; and the cache database is used for storing the access data set, the corrected modeling state set and binding historical state information associated with the time stamp.

Description

Digital twin warehouse management method and system based on three-dimensional real-time modeling Technical Field The invention belongs to the technical field of digital warehouse management, and particularly relates to a digital twin warehouse management method and system for three-dimensional real-time modeling. Background In the modern logistics and warehousing industry, with the wide application of technologies such as industrial Internet of things, edge computing, artificial intelligence and the like, the digital twin technology gradually permeates into each link of warehousing management. The digital twin can reflect the state and behavior characteristics of physical entities in real time by constructing high-fidelity virtual mapping of the real space, supports dynamic prediction, intelligent scheduling and full life cycle management, and becomes one of important support technologies of intelligent storage. In this context, the three-dimensional modeling capability becomes one of the key bases of visual interaction of the digital twin platform, and determines the response accuracy of the management system to physical space and event changes. The existing digital twin warehousing system mainly builds digital mapping of the warehousing environment in a static modeling or periodical updating mode. Typical technical routes include generating static warehouse layout diagrams based on BIM (building information modeling), and local state updating in combination with sensing data such as RFID, cameras or AGVs. However, such methods generally have problems of delayed response, coarse update granularity, incapability of accurately reflecting dynamic changes, and the like. For example, the situations such as the change of the temporary stacking position of the materials, the temporary adjustment of the forklift path, the goods shelf reconstruction and the like often cannot be synchronized to the system in time, so that a system scheduling strategy is invalid or an operator misjudges. In addition, the current partial three-dimensional modeling method relies on manual modeling or laser scanning, has long modeling period and high updating cost, and is not suitable for high-frequency fluctuation scenes. The lack of the capability of fusion modeling between task dynamics, physical states and operation behaviors also limits the capability of the existing system for fine management and real-time optimization of complex tasks. Therefore, it is needed to construct a digital twin warehouse management scheme with three-dimensional real-time modeling capability to adapt to the requirements of intelligent decision making and collaborative execution in a high-frequency dynamic environment. Disclosure of Invention Aiming at the problems, the invention aims to provide a digital twin warehouse management method for three-dimensional real-time modeling, which comprises the following steps: S1, initializing warehouse environment parameters, namely, establishing association mapping corresponding to initial state information and an actual warehouse environment, wherein the association mapping comprises a three-dimensional position of a warehouse entity, a shelf number, equipment coordinates, an entrance coordinate and a loading and unloading operation area; S2, acquiring and accessing multi-type real-time data, wherein the multi-type real-time data comprise data of cargo warehouse entry, warehouse exit, transportation, storage, unloading and equipment running states, and aligning and correcting the acquisition time stamp through a set of synchronous mechanism based on a uniform time source; s3, constructing a three-dimensional real-time state model, wherein the model is based on point cloud and vector data, and fitting the current modeling state by combining a dynamic data set subjected to history state reduction processing to form a space structure constraint condition; s4, modeling error evaluation and fitness judgment, namely calculating modeling errors of the three-dimensional state model by using an error measurement mode after modeling is completed, and judging whether the modeled fitness meets the standard or not according to a preset threshold; S5, performing reverse fitting and historical residual error calibration, namely extracting historical input data and modeling state records of the model nodes in a set time window from the heavy point region, performing parameter reverse pushing and modeling optimization by utilizing residual errors between the current modeling result and the historical records, adjusting the position and attribute information of model nodes by the minimum residual errors, generating corrected modeling results and outputting the corrected modeling results. The step S4 specifically comprises the following steps of S4.1, updating a mapping data cache if modeling errors are within a set threshold value, and synchronously writing the mapping data cache into a modeling database to serve as a digital model