Warehousing digital twinning system

Abstract

The invention belongs to the technical field of digital twinning three-dimensional modeling, and particularly relates to a storage digital twinning system. The system comprises a digital twin model construction module, a real-time positioning module, an electronic fence module, a visual angle display module and a data analysis module. The real-time positioning module collects data by using a UWB positioning unit, coordinates are fused through a positioning data processing unit, and a real-time track is generated. The digital twin model building module is based on a three-dimensional modeling technology and realizes real-time mapping with space coordinates of a physical warehouse. The electronic fence module can define an area, detect a trigger and respond. The visual angle display module supports multi-visual angle switching and first-person visual angle interaction. And the data analysis module calculates the working path efficiency of the forklift and the personnel and visually displays the working path efficiency. According to the method, the defects of an existing storage digital twin system in the aspects of model construction precision, real-time positioning precision, stability and the like are overcome, and the intelligent and efficient storage management level can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of digital twin three-dimensional modeling, and particularly relates to a warehousing digital twin system. Background Art

[0002] Warehousing is an activity of storing and preserving items using warehouses and related facilities and equipment. It not only includes storing goods in warehouse space but also covers a series of tasks such as inbound and outbound management of goods, inventory counting, and maintenance. In the modern logistics system, warehousing plays a key buffering and regulating role.

[0003] With the rapid development of global intelligent logistics and Industry 4.0 technologies, warehousing management is transforming from traditional manual experience-driven to digital and intelligent. Digital twin technology provides a new solution for warehousing facility monitoring and operation process optimization by constructing a virtual mapping of physical entities. However, the digital twin systems currently applied to the warehousing field are not yet perfect, and there are deficiencies in aspects such as the accuracy of model construction, the accuracy and stability of real-time positioning, and multi-module collaborative operations, unable to fully meet the needs of modern warehousing intelligent and efficient management. Summary of the Invention

[0004] The present invention aims at the technical problems existing in the above background and proposes a warehousing digital twin system.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows: It includes a digital twin model construction module, a real-time positioning module, an electronic fence module, a perspective display module, and a data analysis module;

[0006] Among them, the real-time positioning module: includes a UWB positioning unit and a positioning data processing unit;

[0007] The positioning data processing unit is used to fuse the position coordinate data with the spatial coordinates of the digital twin model, display the positions of forklifts and personnel in real time in the digital twin model, and generate real-time trajectories;

[0008] The specific operation method for fusing the position coordinate data with the spatial coordinates of the digital twin model in the positioning data processing unit is as follows:

[0009] Define the state vector , where represents the current position coordinates, represents the current velocity component, are the motion direction angle and angular velocity respectively, is the device state vector;

[0010] Combine the real-time motion to dynamically adjust the state transition matrix , and the calculation formula is: , where represents the basic transition matrix, is the correction matrix, which is determined by the device state and the environmental factor ;

[0011] Perform position prediction , where represents the control input, is the control input matrix, represents the noise-driven matrix, represents the process noise;

[0012] The predicted covariance matrix , where represents the noise-driven matrix, represents the process noise covariance;

[0013] The measurement vector is defined as , and a credibility-weighted measurement prediction is performed on it, and the calculation method is: , where m represents the number of sensors, h is the measurement function, represents the credibility of the sensor;

[0014] Then the interactive multi-model Kalman gain , where is the measurement matrix, which maps the state space to the measurement space;

[0015] Perform state update , where, is the fused measurement vector, and abnormal sensor data is removed through residual threshold detection. If , it is marked as abnormal, where is the standard deviation of the sensor measurement error;

[0016] Update the covariance , where, represents the identity matrix, is the credibility decay matrix, which dynamically increases the covariance according to the number of sensor anomalies. If there are 3 consecutive anomalies, then .

[0017] Preferably, the digital twin model construction module: constructs a digital twin model of the warehouse based on three-dimensional modeling technology and maps it in real time with the spatial coordinates of the physical warehouse;

[0018] The electronic fence module: includes a fence area setting unit, a fence trigger detection unit, and a fence response unit;

[0019] ​Fence area setting unit: used to demarcate an electronic fence area in the digital twin model, and the electronic fence area includes but is not limited to a three-dimensional shelf area and a dangerous goods storage area;

[0020] Fence trigger detection unit: used to monitor the position data of forklifts, personnel and materials in real time, and generate a fence trigger event when it detects entry into or exit from the electronic fence area;

[0021] Fence response unit: used to trigger an alarm mechanism or a device control instruction according to the fence trigger event;

[0022] View display module: includes a multi-view switching unit and a first-person view interaction unit;

[0023] Multi-view switching unit: used to support the user to switch views in the digital twin model; the switched views include switching to a global bird's-eye view, a regional plan view, a first-person roaming view, and a hotspot view, and the hotspot view includes key area views of a preset shelf area, an entrance / exit, and an equipment parking area;

[0024] First-person view interaction unit: used to receive the user's interaction operations, realize simulating the walking path of a person in the first-person roaming view of the digital twin model, and update the view screen in real time;

[0025] Data analysis module: used to analyze real-time positioning data, calculate the operation efficiency of forklifts and the working path efficiency of personnel, and feedback the results to the digital twin model for visual display.

[0026] Preferably, the operation of real-time mapping between the digital twin model and the physical warehouse in the digital twin model construction module is as follows:

[0027] Construct a physical three-dimensional space coordinate system , collect key position coordinate data ;

[0028] Establish a digital twin space coordinate system , collect coordinate data at the same position as the physical space ;

[0029] Perform coordinate model mapping, and achieve initial alignment through an improved affine transformation matrix T. The calculation method is: , where represents the initial error, and the specific form of T is , where R represents a rotation matrix, S is a scale factor matrix, K is a shear factor matrix, and t is a translation vector;

[0030] The Kalman filter algorithm is introduced to iteratively optimize and improve the affine transformation matrix and the initial error, so that the mean square error converges to a set threshold, and a constructed data twin model is obtained.

[0031] Preferably, the UWB positioning unit is used to collect the position coordinate data of forklifts, personnel and materials in real time, including positioning tags and positioning base stations;

[0032] The positioning tags are installed on forklifts, personnel and materials;

[0033] The positioning base stations are installed at key positions in the warehouse and are used to transmit and receive UWB signals to determine the position coordinate data of the positioning tags.

[0034] Preferably, after the UWB positioning unit and before the positioning data processing unit, a vision sensor and an inertial measurement unit are further introduced, which are used to identify the appearance features and relative positions of forklifts and personnel, and to obtain the acceleration and angular velocity of movement.

[0035] Preferably, the specific implementation steps of the first-person perspective interaction unit in the perspective display module are as follows:

[0036] Define the origin displacement of the perspective as the position of the virtual character's head , the perspective adjustment direction vector , the upward vector ;

[0037] Construct the perspective matrix V to convert the world coordinate system to the camera coordinate system ;

[0038] Perform spherical interpolation on the continuous perspective matrix V to avoid perspective jitter, and the calculation method is: , where represents the frame interval time normalization parameter;

[0039] And adjust the field of view FW according to the moving speed of the virtual task, and the calculation method is: , where is the field of view expansion coefficient.

[0040] Preferably, the specific implementation steps of the LA function in the conversion of the world coordinate system to the camera coordinate system are as follows: First, generate the right vector , construct the rotation matrix , and calculate the final view matrix through the translation matrix .

[0041] Preferably, the specific implementation of the data analysis module is as follows:

[0042] Extract the position coordinate sequence from the real-time position data , where each is the coordinate of a positioning point;

[0043] Generate the shortest path coordinate sequence from the task start point to the end point based on the Dijkstra algorithm , and extract key points ;

[0044] Comprehensively calculate the efficiency index through the geometric similarity, path redundancy, and key point coverage of the actual path and the optimal path. The calculation method is:

[0045] , where the three items are the path length efficiency item, the key point coverage, and the redundant path redundancy respectively. Among them, in the redundant path redundancy is an indicator function, which is recorded as 1 when the distance between two points is less than the threshold , reflecting the number of times the path is walked repeatedly.

[0046] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0047] 1. In terms of model construction, through three-dimensional modeling combined with dynamic error correction technology, using an improved affine transformation matrix and a Kalman filtering algorithm, the real-time and accurate mapping between the digital twin model and the spatial coordinates of the physical warehouse is realized, solving the problem of large mapping errors in traditional models.

[0048] 2. The real-time positioning module uses a UWB positioning unit, a vision sensor, and an inertial measurement unit to work together to provide a high-frequency and high-precision position data stream for the digital twin model. When the positioning data processing unit fuses coordinate data, a variety of algorithms are used to improve the fusion accuracy and stability.

[0049] 3. The electronic fence module realizes intelligent monitoring of key areas in the warehouse, and formulates differentiated strategies according to different area types. The perspective display module provides multi-perspective switching and a smooth first-person perspective interaction experience. The data analysis module can effectively evaluate the operation efficiency and visualize the display, helping to manage the warehouse intelligently and efficiently. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0051] Figure 1 It is a schematic structural diagram of a warehouse digital twin system. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] To better understand the above objects, features, and advantages of the present invention, the present invention will be further described below with reference to the accompanying drawings and embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0053] In the following description, many specific details are set forth in order to provide a thorough understanding of the present invention. However, the present invention may be practiced in other ways different from those described herein. Therefore, the present invention is not limited to the limitations of the specific embodiments disclosed in the following specification.

[0054] Embodiments, the digital twin system for the warehousing field is not yet perfect, and there are deficiencies in aspects such as the accuracy of model construction, the accuracy and stability of real-time positioning, and multi-module collaborative operations, and cannot fully meet the needs of modern warehousing intelligent and efficient management. The present invention proposes a warehousing digital twin system. The overall structure is as Figure 1 shown, the warehousing digital twin system includes a digital twin model construction module, a real-time positioning module, an electronic fence module, a perspective display module, and a data analysis module.

[0055] The digital twin model construction module constructs a digital twin model of the warehouse based on three-dimensional modeling technology and maps it in real time with the spatial coordinates of the physical warehouse. This module solves the problem that the mapping error between the traditional digital twin model and the physical warehouse spatial coordinates is large and cannot be accurately matched. Through three-dimensional modeling and dynamic error correction technology, it realizes the real-time and accurate mapping of the virtual model and the physical space. Specifically, taking the northwest corner of the physical space of the warehouse as the origin, a right-handed three-dimensional coordinate system is established The total station is used to collect the coordinate data of key positions such as the corners of the shelves, the center lines of the aisles, and the reference points of the equipment , and a 1:1 scale three-dimensional model of the warehouse is constructed based on the Unity 3D modeling tool, and a twin coordinate system is established , ensuring that the scale and orientation of the virtual space are exactly the same as those of the entity, and synchronously collecting the coordinate data of the corresponding key positions in the model . Perform coordinate model mapping, and achieve initial alignment through the improved affine transformation matrix T. The calculation method is: , where represents the initial error, and the specific form of T is , where R represents the rotation matrix, S is the scale factor matrix, K is the shear factor matrix, and t is the translation vector; the Kalman filter algorithm is introduced to iteratively optimize the improved affine transformation matrix and the initial error, so that the mean square error of the error converges to the set threshold, and the constructed data twin model is obtained.

[0056] The real-time positioning module includes a UWB positioning unit and a positioning data processing unit. The UWB positioning unit is used to collect the position coordinate data of forklifts, personnel and materials in real time, including positioning tags and positioning base stations; the positioning tags are installed on forklifts, personnel and materials; the positioning base stations are installed at key positions in the warehouse and are used to transmit and receive UWB signals to determine the position coordinate data of the positioning tags. Specifically, high-precision positioning is achieved through the cooperation of the positioning tags and the base stations. The positioning tags integrate UWB modules to transmit ultra-wideband signals. The positioning base stations are deployed in a honeycomb pattern on the warehouse ceiling and columns. After receiving the tag signals, they analyze the time of arrival and the time difference of flight of the signals, and use the trilateration algorithm to calculate the three-dimensional coordinates: geometric equations are constructed by measuring the distances with at least 3 base stations, and the coordinate values are solved by the least squares method. It provides a high-frequency and high-precision position data stream for the digital twin model to meet the real-time positioning requirements of warehousing dynamic operations. After the UWB positioning unit and before the positioning data processing unit, a vision sensor and an inertial measurement unit are also introduced to identify the appearance features and relative positions of forklifts and personnel, and to obtain the acceleration and angular velocity of movement.

[0057] The positioning data processing unit is used to fuse the position coordinate data with the spatial coordinates of the digital twin model, display the positions of forklifts and personnel in real time in the digital twin model, and generate real-time trajectories. Among them, the specific operation method for fusing the position coordinate data with the spatial coordinates of the digital twin model in the positioning data processing unit is as follows: First, define a state vector including position, speed, direction angle and equipment status, and dynamically adjust the state transition matrix in combination with real-time motion. Then, perform position prediction and calculate the covariance matrix, weight the multi-sensor measurement vectors according to credibility, and detect and remove noise through the residual threshold. Use the interactive multi-model Kalman gain to update the state with the fused measurement vector, and dynamically adjust the covariance according to the number of sensor anomalies. Finally, achieve the precise fusion of the position coordinates and the spatial coordinates of the digital twin model. Specifically, define the state vector , where represents the current position coordinates, represents the current speed component, are the direction angle of motion and the angular velocity respectively, is the equipment status vector; dynamically adjust the state transition matrix in combination with real-time motion, and the calculation formula is: , where represents the basic transition matrix, is the correction matrix, which is determined by the equipment status and the environmental factor ; perform position prediction , where represents the control input, is the control input matrix, represents the noise-driven matrix, represents the process noise; the prediction covariance matrix , where represents the noise-driven matrix, represents the process noise covariance; the measurement vector is defined as , and a credibility-weighted measurement prediction is performed on it, calculated as: , where m represents the number of sensors, h is the measurement function, represents the credibility of the sensor; then the interactive multiple model Kalman gain , where is the measurement matrix that maps the state space to the measurement space; a state update is performed , where, is the fused measurement vector, and abnormal sensor data is removed through residual threshold detection. If is marked as abnormal when, where is the standard deviation of the sensor measurement error; when updating the covariance , where, represents the identity matrix, is the credibility decay matrix, and the covariance is dynamically increased according to the number of sensor anomalies. If there are 3 consecutive anomalies, then .

[0058] The electronic fence module includes a fence area setting unit, a fence trigger detection unit, and a fence response unit; Fence area setting unit: used to demarcate the electronic fence area in the digital twin model, and the electronic fence area includes but is not limited to the three-dimensional shelf area and the dangerous goods storage area; Fence trigger detection unit: used to monitor the position data of forklifts, personnel, and materials in real time, and generate a fence trigger event when it detects entry into or exit from the electronic fence area; Fence response unit: used to trigger an alarm mechanism or a device control instruction according to the fence trigger event. Specifically, the electronic fence module realizes intelligent monitoring of key areas in the warehouse through the cooperation of the three units. First, the fence area setting unit supports visual demarcation of the electronic fence in the digital twin model, covering various types such as the three-dimensional shelf area and the dangerous goods storage area, and can customize the area shape and trigger rules, and generate differentiated fence strategies by associating with the device status. Secondly, the fence trigger detection unit accesses the target coordinates output by the positioning data processing unit in real time, and dynamically monitors the position changes of forklifts, personnel, and materials through spatial geometric algorithms. When the target coordinates meet the fence entry or exit conditions, a trigger event including the target ID, area type, and timestamp is generated. Finally, the fence response unit classifies and processes the trigger events according to the preset strategy: for the trigger event in the prohibited entry area, immediately send a braking instruction to the target device and push an audible and visual alarm to the management terminal; for the trigger event in the speed-limited operation area, automatically adjust the device operation parameters and highlight the movement trajectory in the digital twin model; for the trigger event in the permission control area, first verify the target permission ID, and start a voice warning and record an exception log when there is no permission.

[0059] The perspective display module includes a multi-perspective switching unit and a first-person perspective interaction unit. Multi-perspective switching unit: used to support the user to switch perspectives in the digital twin model; the switching perspectives include switching to the global bird's-eye view, the regional plan view, the first-person roaming view, and the hot-spot view, and the hot-spot view includes the key area views of the preset shelf area, the entrance and exit, and the equipment parking area. First-person perspective interaction unit: used to receive the user's interaction operations, and realize simulating the walking path of a person in the digital twin model using the first-person roaming view, and updating the perspective screen in real time. The specific implementation steps of the first-person perspective interaction unit in the perspective display module are as follows: define the origin displacement of the perspective to the virtual character's head position , the perspective adjustment direction vector , the upward vector ; construct the perspective matrix V to convert the world coordinate system to the camera coordinate system ; perform spherical interpolation on the continuous perspective matrix V to avoid perspective jitter, and the calculation method is: , where represents the frame time normalization parameter; and adjust the field of view range FW according to the moving speed of the virtual task, and the calculation method is: , where is the field of view expansion coefficient. Among them, the specific implementation steps of converting the world coordinate system into the LA function in the camera coordinate system are as follows: First, generate the right vector , construct the rotation matrix , and calculate the final view matrix through the translation matrix . .

[0060] The data analysis module is used to analyze the real-time positioning data, calculate the forklift operation and personnel operation path efficiency, and feedback the results to the digital twin model for visual display. Specifically, extract the position coordinate sequence from the real-time position data , where each is the positioning point coordinate; generate the shortest path coordinate sequence from the task start point to the end point based on the Dijkstra algorithm , and extract the key points ; comprehensively calculate the efficiency index through the geometric similarity, path redundancy and key point coverage of the actual path and the optimal path. The calculation method is as follows:

[0061] , where the three items are the path length efficiency item, the key point coverage, and the redundant path redundancy respectively. Among them, in the redundant path redundancy is the indicator function, which is recorded as 1 when the distance between two points is less than the threshold , reflecting the number of times the path is walked repeatedly. Feed the efficiency index back to the digital twin model. Display the actual and optimal paths in different colors, and the transparency reflects the redundancy. Mark the key points, and flash the prompt for those not covered.

[0062] As described above, it is only a preferred embodiment of the present invention, and it is not a limitation of the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical content of the technical solution of the present invention, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention still belong to the protection scope of the technical solution of the present invention.

Claims

1. A warehouse digital twin system, characterized in that: It includes digital twin model building module, real-time positioning module, electronic fence module, perspective display module and data analysis module; Among them, the real-time positioning module includes a UWB positioning unit and a positioning data processing unit; The positioning data processing unit is used to fuse the position coordinate data with the spatial coordinates of the digital twin model, display the positions of the forklift and personnel in real time in the digital twin model, and generate a real-time trajectory; The specific operation method of fusing the position coordinate data with the spatial coordinates of the digital twin model in the positioning data processing unit is: Define the state vector ,in Represents the current position coordinates, represents the current velocity component, are the motion direction angle and angular velocity respectively, is the device state vector; Dynamically adjust the state transfer matrix in combination with real-time motion , the calculation formula is: ,in represents the basic transfer matrix, is the correction matrix, which is composed of the device status and environmental factors Decide; Make location prediction ,in represents the control input, is the control input matrix, represents the noise driving matrix, represents process noise; Prediction covariance matrix ,in represents the noise driving matrix, represents the process noise covariance; The measurement vector is defined as , and perform a credibility weighted measurement prediction on it, the calculation method is: , where m represents the number of sensors and h is the measurement function. Represents the reliability of the sensor; Then the interactive multi-model Kalman gain ,in is the measurement matrix, which maps the state space to the measurement space; Make a status update ,in, is the fused measurement vector, and the abnormal sensor data is removed through residual threshold detection. is marked as an exception, where is the standard deviation of the sensor measurement error; In the updated covariance ,in, represents the identity matrix, is the credibility attenuation matrix, which dynamically increases the covariance according to the number of sensor anomalies. If there are three consecutive anomalies, .

2. A warehouse digital twin system according to claim 1, characterized in that: The digital twin model construction module: constructs a digital twin model of the warehouse based on three-dimensional modeling technology, and maps it with the spatial coordinates of the physical warehouse in real time; Electronic fence module: including fence area setting unit, fence trigger detection unit, and fence response unit; A fence area setting unit: used to define an electronic fence area in the digital twin model, wherein the electronic fence area includes but is not limited to a three-dimensional shelf area and a dangerous goods storage area; Fence trigger detection unit: used to monitor the location data of forklifts, personnel and materials in real time, and generate fence trigger events when entering or leaving the electronic fence area is detected; Fence response unit: used to trigger an alarm mechanism or device control instruction according to the fence triggering event; Perspective display module: including multi-perspective switching unit and first-person perspective interaction unit; Multi-perspective switching unit: used to support users to switch perspectives in the digital twin model; the switching perspectives include switching the global bird's-eye view, the regional plane view, the first-person roaming view and the hotspot view. The hotspot view includes the preset shelf area, entrance and exit, and equipment parking area key area perspectives; First-person perspective interaction unit: used to receive user interaction operations, implement the simulation of the walking path of personnel in the first-person roaming perspective of the digital twin model, and update the perspective screen in real time; Data analysis module: used to analyze real-time positioning data, calculate the efficiency of forklift operation and personnel work paths, and feed the results back to the digital twin model for visual display.

3. A warehouse digital twin system according to claim 2, characterized in that: The operation of real-time mapping between the digital twin model and the physical warehouse in the digital twin model construction module is: Constructing a solid three-dimensional space coordinate system , collect key location coordinate data ; Establishing the digital twin space coordinate system , collect coordinate data of the same position as the physical space ; Coordinate model mapping is performed and initial alignment is achieved by improving the affine transformation matrix T, which is calculated as follows: ,in, Represents the initial error, and the specific form of T is , where R represents the rotation matrix, S is the scale factor matrix, K is the shear factor matrix, and t is the translation vector; The Kalman filter algorithm is introduced to iteratively optimize and improve the affine transformation matrix and initial error so that the mean square error converges to the set threshold and a constructed data twin model is obtained.

4. A warehouse digital twin system according to claim 1, characterized in that: The UWB positioning unit is used to collect the position coordinate data of forklifts, personnel and materials in real time, including positioning tags and positioning base stations; The positioning tags are installed on forklifts, personnel and materials; The positioning base station is installed at a key position of the warehouse and is used to transmit and receive UWB signals to determine the position coordinate data of the positioning tag.

5. A warehouse digital twin system according to claim 4, characterized in that: After the UWB positioning unit, visual sensors and inertial measurement units should be introduced before the positioning data processing unit to identify the appearance characteristics and relative positions of forklifts and personnel, and to obtain the acceleration and angular velocity of the motion.

6. A warehouse digital twin system according to claim 2, characterized in that: The specific implementation steps of the first-person perspective interaction unit in the perspective display module are: Define the perspective origin displacement virtual character head position , the viewing angle adjustment direction vector , the up vector ; Construct the perspective matrix V to transform the world coordinate system into the camera coordinate system ; Spherical interpolation is performed on the continuous viewing matrix V to avoid viewing jitter. The calculation method is: ,in represents the inter-frame temporal normalization parameter; And adjust the field of view FW according to the virtual task movement speed, the calculation method is: ,in is the field of view expansion coefficient.

7. A warehouse digital twin system according to claim 6, characterized in that: The specific implementation steps of the LA function in converting the world coordinate system to the camera coordinate system are as follows: first, generate the right vector , construct the rotation matrix , through the translation matrix Calculate the final view matrix .

8. A warehouse digital twin system according to claim 2, characterized in that: The specific implementation of the data analysis module is: Extracting location coordinate sequences from real-time location data , where each is the coordinate of the anchor point; Generate the shortest path coordinate sequence from the starting point to the end point of the task based on the Dijkstra algorithm , and extract key points ; The efficiency index is calculated by comprehensively calculating the geometric similarity between the actual path and the optimal path, the path redundancy and the key point coverage. The calculation method is: , where the three items are path length efficiency, key point coverage and repeated path redundancy. is an indicator function. When the distance between two points is less than the threshold 1, which is the number of times the reaction path is repeated.

Images