Bidirectional mapping debugging method and system for digital twin and physical entity of transfer robot

Abstract

The invention relates to a bidirectional mapping debugging method and system for a digital twinborn body and a physical entity of a transfer robot, and the method comprises the steps: constructing a digital twinborn scene of the transfer robot, and building an initial mapping relation between the digital twinborn body and the physical entity; generating candidate paths based on the digital twin scene parameters, and outputting candidate path parameters; performing virtual rehearsal and feasibility verification on the candidate path parameters to obtain optimal path parameters; establishing a real-time data transmission link between the digital twin and the physical entity, and collecting actual operation parameters of the physical robot and real-time environment parameters of the physical scene; and comparing the optimal path parameter in the digital twin scene with the actual operation parameter of the physical robot, identifying the parameter deviation and analyzing the deviation source. The technical problems that in an existing transfer robot debugging technology, the physical trial and error efficiency is low, simulation and physical scenes are disjointed, path generation lacks multi-target optimization, virtual and real debugging collaboration is insufficient, and the production changing debugging period is long are solved.

Description

Technical Field

[0001] This application relates to the field of debugging technology for handling robots, specifically to a two-way mapping debugging method and system for a handling robot between a digital twin and its physical entity. Background Technology

[0002] In practical applications of handling robots, path debugging is a crucial step in ensuring stable equipment operation. Traditional debugging methods mainly rely on physical trial and error or basic simulation, which have many technical bottlenecks. The physical trial and error mode requires engineers to manually and repeatedly adjust path parameters, making each debugging process cumbersome and time-consuming. At the same time, it lacks a pre-judgment mechanism, resulting in a high failure rate of robots colliding with shelves or goods. It can only adapt to a single type of goods, and a complete process debugging is required when changing products, which cannot meet the needs of rapid handling of multiple types of goods.

[0003] While the basic simulation debugging mode introduces virtual simulation, there is a serious disconnect between the simulation scenario and the physical scenario. It fails to consider the impact of dynamic environmental changes and cargo status changes on the path, resulting in significant deviations in the virtual path during actual vehicle operation. This type of mode lacks a closed-loop optimization mechanism, separating simulation from actual vehicle debugging. It cannot provide real-time feedback of real-vehicle issues to the virtual end for parameter optimization. Furthermore, the path generation algorithm is simplistic, focusing solely on minimizing path length without considering constraints such as energy consumption, cargo protection, and equipment lifespan. The generated paths are of poor practicality and require secondary manual adjustments.

[0004] Industry data shows that traditional debugging methods result in a high proportion of debugging time for handling robots in the total downtime of the equipment, and the production capacity loss caused by production change debugging is significant, which has become a core bottleneck restricting the improvement of logistics efficiency. There is an urgent need for a debugging solution that can achieve virtual-physical collaboration, multi-objective optimization, and rapid iteration. Summary of the Invention

[0005] This application mainly addresses the technical problems in existing handling robot debugging technologies, such as low efficiency of physical trial and error, disconnect between simulation and physical scene, lack of multi-objective optimization in path generation, insufficient coordination between virtual and real debugging, and long production changeover debugging cycle.

[0006] This application provides a bidirectional mapping and debugging method for a handling robot between a digital twin and its physical entity. The bidirectional mapping and debugging method for the handling robot includes: Construct a digital twin scenario of the handling robot, generate digital twin scenario parameters, and establish an initial mapping relationship between the digital twin and the physical entity; Generate at least one set of candidate paths that satisfy multiple constraints based on the parameters of the digital twin scenario, and output the candidate path parameters; In a digital twin scenario, candidate path parameters are virtually simulated and their feasibility verified to select the optimal path parameters. Establish a real-time data transmission link between the digital twin and the physical entity, send the optimal path parameters to the physical robot, and simultaneously collect the actual operating parameters of the physical robot and the real-time environmental parameters of the physical scene, and feed them back to the digital twin scene; By comparing the optimal path parameters in the digital twin scenario with the actual operating parameters of the physical robot, parameter deviations are identified and their root causes are analyzed. Based on the root causes of the deviations, the parameters of the digital twin scenario and the motion parameters of the physical robot are dynamically adjusted to achieve bidirectional mapping optimization between the digital twin and the physical entity, and the debugging is completed.

[0007] In the aforementioned bidirectional mapping and debugging method for the digital twin and physical entity of a handling robot, as a preferred embodiment, the steps of constructing a digital twin scene of the handling robot, generating digital twin scene parameters, and establishing an initial mapping relationship between the digital twin and the physical entity include: Real-time environmental parameters of the physical scene are collected and imported into the digital twin scene. The virtual environment parameters are dynamically adjusted to form the environmental sub-parameters in the digital twin scene parameters. Import the cargo model and calculate the cargo center of gravity parameters. Simulate the impact of cargo state changes on the center of gravity parameters to form cargo sub-parameters in the digital twin scene parameters. Import the hardware parameters of the handling robot as constraint sub-parameters for path generation in the digital twin scenario, forming the device sub-parameters in the digital twin scenario parameters; Environmental sub-parameters, cargo sub-parameters, and equipment sub-parameters are integrated into digital twin scene parameters. The digital twin scene parameters are periodically compared with the real-time environmental parameters of the physical scene and the hardware parameters of the physical robot to correct deviations in the digital twin scene parameters and maintain dynamic consistency.

[0008] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of the handling robot, as a preferred embodiment, the step of collecting real-time environmental parameters of the physical scene and importing them into the digital twin scene, dynamically adjusting the virtual environment parameters to form environmental sub-parameters of the digital twin scene, includes: Temperature and humidity data and ground friction data of the physical scene are collected by temperature and humidity sensors and ground friction sensors deployed in the physical scene, and wind speed data of the physical scene is obtained by wind speed acquisition equipment. The data are integrated to form real-time environmental parameters of the physical scene. The real-time environmental parameters of the physical scene are automatically synchronized to the digital twin scene. The friction coefficient parameter of the virtual environment is adjusted based on temperature and humidity data, and the air resistance parameter of the virtual environment is adjusted based on wind speed data, forming the environmental sub-parameters of the digital twin scene. Regularly compare the environmental sub-parameters of the digital twin scene with the real-time environmental parameters of the physical scene, set a deviation threshold, and adjust the environmental sub-parameters of the digital twin scene when the deviation between the environmental sub-parameters and the real-time environmental parameters exceeds the threshold so that the deviation between the environmental sub-parameters and the real-time environmental parameters is controlled within a preset range.

[0009] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of the handling robot, as a preferred embodiment, the step of generating at least one set of candidate paths satisfying multiple constraints based on the digital twin scene parameters and outputting candidate path parameters includes: The device sub-parameters and cargo sub-parameters in the digital twin scene parameters are extracted as constraints; among them, the device sub-parameters are robot hardware parameters, and the cargo sub-parameters are cargo center of gravity parameters, and a multi-objective optimization function is constructed. The multi-objective optimization function takes path length, robot energy consumption, cargo protection requirements and equipment life constraints as the core optimization dimensions. The four core optimization dimensions correspond to the basic requirements of path planning, energy consumption control requirements, cargo safety requirements and equipment wear control requirements, respectively. Based on the cargo weight and fragility characteristics in the cargo sub-parameters, the weight ratio of each dimension of the multi-objective optimization function is dynamically allocated to achieve the adaptation of the optimization function to the cargo attributes; A multi-objective optimization algorithm is used to solve the function, generating multiple sets of initial path parameters. Combined with the constraint sub-parameters in the digital twin scenario parameters, i.e., the hardware rating values, the initial path parameters that do not meet the constraints are eliminated, and finally at least one set of candidate path parameters is output.

[0010] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of the handling robot, as a preferred embodiment, the construction of the multi-objective optimization function includes: The expression for the multi-objective optimization function is: F=ω1×L+ω2×E+ω3×a+ω4×T; Where F is the comprehensive value of the optimization function, L is the path length in the candidate path parameters, and ω1 is the weight of the path length; E is the path energy consumption calculated based on the robot kinematic model and candidate path parameters, and ω2 is the weight of the path energy consumption; a is the maximum path acceleration in the candidate path parameters, and ω3 is the weight of the maximum path acceleration; T is the ratio of the maximum joint torque in the candidate path parameters to the sub-parameters of the digital twin scene device, where the digital twin scene device sub-parameters are the rated values ​​of the joint torque, and ω4 is the weight of the ratio T; The weights are adjusted based on the cargo type in the cargo sub-parameters: When handling heavy cargo, if the cargo weight parameter is greater than or equal to the set value, the weight of ω2 is increased to prioritize energy consumption control; for fragile cargo, the weight of ω3 is increased to prioritize acceleration control; for light cargo, ω1 is maintained as the dominant weight to prioritize path length control. Set constraint thresholds for each optimization dimension: L satisfies the path planning range in the digital twin scenario parameters, E is lower than the robot's single-cycle energy consumption limit in the digital twin scenario device sub-parameters, a does not exceed the maximum allowable acceleration for cargo protection in the cargo sub-parameters, and T does not exceed the set ratio to ensure that the joint torque does not exceed the rated value in the device sub-parameters.

[0011] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of the handling robot, as a preferred embodiment, the step of virtually pre-simulating and verifying the feasibility of candidate path parameters in the digital twin scenario to obtain the optimal path parameters includes: In a digital twin scenario, candidate path parameters are loaded to simulate typical abnormal scenarios. Typical abnormal scenarios include sudden personnel intrusion, cargo center of gravity parameter shift, and temporary equipment sub-parameter failure. The candidate path parameters are then run. Paths are selected using two core indicators: pass rate and average deviation. The pass rate is the percentage of times a candidate path parameter successfully avoids obstacles without exceeding the parameter limit in all abnormal scenarios. The average deviation is the average deviation between the planned position and the actual running position of the candidate path parameter. Candidate path parameters with high pass rates and low average deviations are given priority. Extract the motion parameters of key nodes from the candidate path parameters. Key nodes are turning points and shelf docking points. The motion parameters are joint angles, speeds, and torques. Verify whether the motion parameters are within the hardware rated range of the device sub-parameters in the digital twin scenario. If the motion parameters of a node exceed the rated range, automatically adjust the motion parameters of that node and re-verify the candidate path parameters until all node motion parameters are within the rated range, and obtain the optimal path parameters.

[0012] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of a handling robot, as a preferred embodiment, the establishment of a real-time data transmission link between the digital twin and the physical entity, the distribution of optimal path parameters to the physical robot, and the synchronous collection of the physical robot's actual operating parameters and the real-time environmental parameters of the physical scene, fed back to the digital twin scene, includes: Dual transmission links are deployed on the physical robot. The first transmission link is used to transmit the optimal path parameters and the actual operating parameters of the physical robot, including position coordinates and joint torques. The second transmission link is used to transmit the real-time environmental parameters of the physical scene and high-definition images of the physical scene, including temperature, humidity, friction, and wind speed. Operators connect to the digital twin platform through a visual interactive device, overlaying the optimal path parameters in the digital twin scene onto the physical scene in a preset form, marking the stuttering points and deviation points in the operation of the physical robot in real time, forming problem marking information, and associating it with the corresponding actual operating parameter timestamps; The actual operating parameters of the physical robot, the real-time environmental parameters of the physical scene, and the problem marking information are synchronized to the digital twin scene through dual transmission links, updating the real-time data in the digital twin scene and ensuring real-time alignment between the virtual and physical ends.

[0013] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of a handling robot, as a preferred embodiment, the step of comparing the optimal path parameters in the digital twin scenario with the actual operating parameters of the physical robot, identifying parameter deviations and analyzing their root causes, and dynamically adjusting the digital twin scenario parameters and the motion parameters of the physical robot based on the root causes of the deviations to achieve bidirectional mapping optimization and complete the debugging includes: The optimal path parameters in the digital twin scenario are compared with the actual operating parameters of the physical robot at set time intervals. The optimal path parameters include the planned position, planned velocity, and planned acceleration. The actual operating parameters are the actual position, actual velocity, and actual acceleration. The parameter deviation values ​​are calculated. When the parameter deviation exceeds the set threshold, deviation analysis is triggered, which retrieves the actual operating parameters, real-time environmental parameters of the physical scene, and digital twin scene parameters corresponding to the deviation node. The actual operating parameters include joint torque and motion speed; the real-time environmental parameters of the physical scene include temperature, humidity, and friction. A fusion model is used to identify the root cause of the deviation. Adjust parameters based on the root cause of the deviation: If the deviation originates from torque exceeding the threshold, i.e., the joint torque in the actual operating parameters exceeds the rated value of the sub-parameters of the digital twin scene device, a motion speed adjustment command is issued through the transmission bus to correct the motion parameters of the physical robot; if the deviation originates from environmental changes, i.e., the deviation between the real-time environmental parameters of the physical scene and the environmental sub-parameters of the digital twin scene exceeds the limit, the environmental sub-parameters of the digital twin scene are updated; continuously monitor the adjusted parameter deviation value until the deviation returns to the set range, and the debugging is completed.

[0014] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of the handling robot, as a preferred embodiment, the step of retrieving the actual operating parameters corresponding to the deviation node, the real-time environmental parameters of the physical scene, and the digital twin scene parameters, and using a fusion model to identify the root cause of the deviation, includes: Record multi-source data of deviation nodes with set precision: actual operating parameters, real-time environmental parameters of the physical scene, and digital twin scene parameters, while associating the timestamp of the problem marker information; among them, the actual operating parameters include the actual value of joint torque, the actual value of movement speed, and the parameter deviation value; the real-time environmental parameters of the physical scene include the temperature and humidity, ground friction, and wind speed when the deviation occurs; the digital twin scene parameters include the environmental sub-parameters and equipment sub-parameters at the corresponding time. In the fusion model, the input data is first roughly classified based on the comparison results between the digital twin scene parameters and the real-time environmental parameters of the physical scene, as well as the deviation values ​​between the actual operating parameters and the optimal path parameters. The preliminary judgment includes environmental deviation, equipment parameter, or path planning categories. Based on similar deviation cases in historical debugging data, the coarse classification results are further refined by associating similar deviation cases with the correspondence between actual operating parameters, real-time environmental parameters and deviation root causes. The specific deviation root causes and confidence levels are output. Environmental deviations caused by changes in ground friction can be used as one of the specific deviation root causes, and the confidence level is determined according to the calculation results. Establish a multi-source data log to store digital twin scene parameters, real-time environmental parameters of the physical scene, actual operating parameters of the physical robot, parameter deviation values, deviation analysis results, and parameter adjustment records. The log includes timestamps and data hash values ​​for subsequent deviation backtracking and debugging optimization reference.

[0015] This application also provides a bidirectional mapping and debugging system for a handling robot's digital twin and its physical entity, the system comprising: A digital twin scene construction module is used to collect real-time environmental parameters of the physical scene, cargo center of gravity parameters and robot hardware parameters, generate digital twin scene parameters, construct a digital twin scene of the handling robot, and establish an initial mapping relationship between the digital twin and the physical entity. The candidate path generation module is used to construct a multi-objective optimization function based on the equipment sub-parameters and cargo sub-parameters in the digital twin scenario parameters, solve to generate multiple sets of initial path parameters, and filter and output at least one set of candidate path parameters. The virtual pre-simulation and screening module is used to perform virtual pre-simulation and feasibility verification on candidate path parameters in a digital twin scenario, and to screen the optimal path parameters by comparing the verification results. The real-time data transmission module is used to establish a real-time data transmission link between the digital twin and the physical entity, issue optimal path parameters, and synchronously collect and feedback the actual operating parameters of the physical robot and the real-time environmental parameters of the physical scene. The bidirectional mapping optimization module is used to compare the optimal path parameters with the actual operating parameters, identify parameter deviations and analyze their root causes, and adjust the environmental sub-parameters, equipment sub-parameters and physical robot motion parameters in the digital twin scene parameters based on the root causes to achieve bidirectional mapping optimization and complete the debugging.

[0016] The method and system for bidirectional mapping debugging of digital twins and physical entities of handling robots according to the above embodiments construct a high-fidelity digital twin scene and dynamically integrate multi-dimensional parameters of environment, goods, and equipment to ensure the consistency between the virtual scene and the physical scene, significantly reducing path deviation. It generates customized paths adapted to different goods characteristics through multi-objective optimization functions, taking into account path length, energy consumption, goods protection, and equipment lifespan, thus improving path practicality. It simulates abnormal scenarios through virtual pre-simulation, eliminating risks such as collisions and parameter exceeding limits in advance, reducing the failure rate of actual vehicle debugging. It achieves real-time virtual-physical linkage through dual transmission links and visual interaction, quickly locating operational problems. Through deviation root cause analysis and bidirectional parameter adjustment, it forms a closed-loop optimization process, significantly shortening the debugging cycle. This application effectively solves the problems of low debugging efficiency, poor accuracy, and weak adaptability in traditional debugging, significantly improving the efficiency of single debugging and the speed of debugging for multiple types of goods, reducing equipment collision risks and goods damage rates, extending equipment lifespan, providing reliable technical support for the efficient and stable operation of handling robots, and promoting the intelligent and flexible development of the logistics handling field. Attached Figure Description

[0017] Figure 1 A flowchart (I) of the bidirectional mapping debugging method between the digital twin and the physical entity of the handling robot provided in the embodiments of this application; Figure 2 A flowchart (II) of the bidirectional mapping debugging method between the digital twin and the physical entity of the handling robot provided in the embodiments of this application; Figure 3 A flowchart (III) of the bidirectional mapping debugging method between the digital twin and the physical entity of the handling robot provided in the embodiments of this application; Figure 4 This is a schematic diagram of the principle of the bidirectional mapping and debugging system between the digital twin and the physical entity of the handling robot provided in the embodiments of this application. Detailed Implementation

[0018] The present application will now be described in further detail with reference to the accompanying drawings and specific embodiments. Similar elements in different embodiments are referred to by related similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the present application. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present application are not shown or described in the specification. This is to avoid obscuring the core parts of the present application with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0019] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0020] The serial numbers assigned to components in this document, such as "first" and "second," are used only to distinguish the described objects and have no sequential or technical meaning. The terms "connection" and "linkage" used in this application, unless otherwise specified, include both direct and indirect connections (linkages).

[0021] Please refer to Figure 1 To address the technical problems of low efficiency in physical trial and error, disconnect between simulation and physical scene, lack of multi-objective optimization in path generation, insufficient coordination between virtual and physical debugging, and long debugging cycles in existing handling robot debugging technologies, this application provides a bidirectional mapping debugging method between a digital twin and a physical entity of a handling robot. By constructing a high-fidelity digital twin scene, generating multi-constraint candidate paths, real-time virtual-physical linkage, and dynamic optimization, an efficient, accurate, and flexible debugging process is achieved, adapting to the needs of handling various types of goods. The bidirectional mapping debugging method between the digital twin and the physical entity of the handling robot includes the following steps: Step S101: Construct a digital twin scene of the handling robot, generate digital twin scene parameters, and establish an initial mapping relationship between the digital twin and the physical entity.

[0022] In some embodiments, constructing a digital twin scenario requires comprehensive integration of multi-dimensional parameters of the physical scene, goods, and robot to ensure a high-fidelity mapping between the virtual and physical worlds. Real-time environmental data is collected by sensors deployed at key locations in the physical scene: temperature and humidity sensors capture changes in ambient temperature and humidity; ground friction sensors monitor differences in friction coefficients across different areas; and wind speed sensors record airflow patterns. This data is automatically synchronized to the digital twin platform, dynamically adjusting corresponding parameters of the virtual environment to form environmental sub-parameters. A 3D model of the goods is imported, and the center of gravity coordinates are automatically resolved using a center of gravity calculation algorithm. Simultaneously, the impact of changes in the number of stacked layers and stack offset on the center of gravity is simulated to generate dynamically updated goods sub-parameters. The core hardware parameters of the handling robot, including joint torque ratings, maximum movement speed, acceleration threshold, and motor power, are imported as hard constraints for path generation, forming equipment sub-parameters. The environmental sub-parameters, goods sub-parameters, and equipment sub-parameters are integrated into complete digital twin scenario parameters, establishing an initial mapping relationship between the digital twin and the physical entity at the parameter level. To maintain dynamic consistency, the system periodically compares the parameters of the digital twin scene with the measured data of the physical scene and the real-time parameters of the robot hardware. When the deviation exceeds the preset threshold, the system automatically triggers parameter correction to ensure that the virtual scene can reflect the changes of the physical scene in real time.

[0023] Step S102: Generate at least one set of candidate paths that meet multiple constraints based on the digital twin scene parameters, and output the candidate path parameters.

[0024] In some embodiments, path generation is based on digital twin scene parameters to construct an optimization system that considers multiple objectives. Hardware constraints from equipment sub-parameters and characteristic requirements from cargo sub-parameters are extracted to establish four optimization dimensions: path length, robot energy consumption, cargo protection, and equipment lifespan, thus constructing a multi-objective optimization function. The weights of each dimension in the function are dynamically allocated according to cargo type: heavy cargo prioritizes increasing energy consumption weight to control energy consumption; fragile cargo prioritizes increasing acceleration weight to reduce impact; and light cargo maintains path length weight to ensure transportation efficiency. Simultaneously, constraint thresholds are set for each dimension: path length must adapt to the spatial layout of the physical scene; energy consumption must not exceed the robot's single-cycle endurance limit; acceleration must be within the allowable range for cargo protection; and joint torque must not exceed the equipment's rated value. A multi-objective optimization algorithm is used to solve the function, generating multiple sets of initial path parameters. These are then rigorously screened using the hardware rated values ​​from the equipment sub-parameters, eliminating paths with parameters exceeding limits or inability to avoid obstacles. Finally, at least one set of candidate path parameters that satisfy all constraints is output, providing a foundation for subsequent screening.

[0025] Step S103: Perform virtual pre-simulation and feasibility verification on the candidate path parameters in the digital twin scenario, and select the optimal path parameters.

[0026] In some embodiments, virtual pre-simulation and verification aim to eliminate path defects in advance and ensure the practicality of the optimal path. Candidate path parameters are loaded into a digital twin scenario to simulate typical anomalies that may occur in logistics scenarios, including sudden personnel intrusion, cargo center of gravity shift, and temporary robot joint malfunctions, comprehensively testing the adaptability of candidate paths. Path performance is evaluated using two core indicators: pass rate and average deviation. The pass rate reflects the path's ability to successfully avoid obstacles in abnormal scenarios without exceeding parameter limits, while the average deviation reflects the degree of consistency between the planned path location and the actual operating location. Candidate paths with high pass rates and low average deviations are prioritized. Key nodes in the path are further extracted, such as turning points, shelf docking points, and narrow passages, to verify whether the joint angles, movement speeds, torques, and other parameters of these nodes meet the hardware rated range in the equipment sub-parameters. If a node parameter is found to exceed the limit, the node's motion parameters are automatically adjusted, such as extending the movement time or reducing acceleration, and virtual pre-simulation verification is performed again until all node parameters meet the requirements, ultimately selecting the optimal path parameters.

[0027] Step S104: Establish a real-time data transmission link between the digital twin and the physical entity, send the optimal path parameters to the physical robot, and simultaneously collect the actual operating parameters of the physical robot and the real-time environmental parameters of the physical scene, and feed them back to the digital twin scene.

[0028] In some embodiments, the real-time data transmission link is a key support for realizing virtual-physical collaboration. A dual-link architecture ensures the real-time performance and integrity of data transmission. The first transmission link uses low-latency bus technology to transmit optimal path parameters and core operational data of the physical robot, including position coordinates, joint torques, and movement speed, ensuring rapid response to command issuance and status feedback. The second transmission link uses high-speed network technology to transmit real-time environmental parameters and high-definition images of the physical scene. Environmental parameters include dynamic changes in temperature, humidity, ground friction, and wind speed, while high-definition images are used to intuitively present the real-time physical scene. The operator connects to the digital twin platform through an industrial-grade visual interactive device, overlaying the optimal path onto the physical scene in an intuitive way, such as highlighting lines, to achieve a visual fusion of the virtual path and the physical environment. During robot operation, the operator can mark abnormal locations such as stuttering points and offset points in real time, forming problem marker information and associating it with corresponding timestamps. All collected actual operating parameters of the physical robot, real-time environmental parameters of the physical scene, and problem marker information are synchronously fed back to the digital twin scene through the dual transmission links, updating the virtual data in real time and ensuring a high degree of synchronization between the digital twin and the physical entity.

[0029] Step S105: Compare the optimal path parameters in the digital twin scene with the actual operating parameters of the physical robot, identify parameter deviations and analyze the root causes of deviations, dynamically adjust the digital twin scene parameters and the motion parameters of the physical robot based on the root causes of deviations, realize bidirectional mapping optimization between the digital twin and the physical entity, and complete the debugging.

[0030] In some embodiments, bidirectional mapping optimization forms a closed loop through deviation analysis and dynamic adjustment to ensure debugging accuracy. The system compares the optimal path parameters in the digital twin scenario with the actual operating parameters of the physical robot at fixed time intervals, calculating parameter deviations in dimensions such as position, speed, and acceleration. When the deviation exceeds a preset threshold, the deviation root cause analysis process is automatically triggered, retrieving multi-source data corresponding to the deviation node, including the physical robot's joint torque and current changes, environmental data of the physical scenario such as temperature and humidity and ground friction, and corresponding parameters of the digital twin scenario. A fusion model is used to perform in-depth analysis of this data, first coarsely classifying the deviation as belonging to environmental changes, equipment status, or path planning issues, and then combining historical debugging data for fine classification to accurately locate the specific root cause of the deviation, such as decreased ground friction, joint torque exceeding the threshold, or cargo center of gravity shift. Targeted adjustment measures are taken based on the root cause of the deviation: if it stems from excessive equipment parameters, parameter adjustment commands are sent to the physical robot via the transmission link to correct parameters such as motion speed and acceleration; if it stems from environmental changes, the corresponding environmental sub-parameters of the digital twin scenario are updated, and subsequent path planning is optimized synchronously. Continuously monitor the deviation of the adjusted parameters until the deviation returns to the allowable range. At the same time, establish a complete multi-source data log to record the parameters, deviations, analysis results and adjustment records of the entire debugging process, so as to provide a reference for subsequent debugging and optimization and complete the entire debugging process.

[0031] Please refer to Figure 2 In the aforementioned bidirectional mapping debugging method between the digital twin and the physical entity of the handling robot, as a preferred embodiment, step S101, which involves constructing the digital twin scene of the handling robot, generating digital twin scene parameters, and establishing the initial mapping relationship between the digital twin and the physical entity, includes the following steps: Step S201: Collect real-time environmental parameters of the physical scene and import them into the digital twin scene. Dynamically adjust the virtual environment parameters to form the environmental sub-parameters in the digital twin scene parameters.

[0032] In some embodiments, environmental parameter acquisition needs to achieve full-dimensional coverage and real-time synchronization to ensure dynamic matching between the virtual environment and the physical scene. Sensors are uniformly deployed in key areas of the physical scene. Temperature and humidity sensors collect environmental temperature and humidity data at a fixed frequency to accurately capture temperature and humidity fluctuations in scenarios such as warehouses and workshops. Ground friction sensors are installed on the robot's main travel paths and turning areas to monitor changes in the ground friction coefficient at different locations in real time, paying particular attention to friction reduction caused by factors such as water stains and oil stains. Wind speed acquisition devices are deployed in open areas and near ventilation openings to record airflow speed and direction. This real-time environmental data is automatically imported into the digital twin scene through a high-speed transmission link. The system dynamically adjusts the virtual environment parameters based on preset mapping rules: the friction coefficient of the virtual ground is corrected according to temperature and humidity data, and the friction coefficient is appropriately reduced in high humidity environments to simulate a slippery ground condition; the air resistance parameters of the virtual environment are adjusted in combination with wind speed data, increasing the resistance value in the corresponding direction along the wind direction to ensure that the virtual environment can truly reflect the environmental mechanical characteristics of the physical scene. The final environmental sub-parameters cover key indicators such as friction coefficient, air resistance, temperature, and humidity, providing realistic environmental constraints for subsequent path generation.

[0033] Step S202: Import the cargo model and calculate the cargo center of gravity parameters. Simulate the impact of cargo state changes on the center of gravity parameters to form cargo sub-parameters in the digital twin scene parameters.

[0034] In some embodiments, the generation of cargo sub-parameters needs to balance accuracy and dynamism, fully adapting to state changes during cargo handling. It supports importing 3D cargo models in various formats, using an automatic center of gravity calculation algorithm to divide the model into volumetric meshes, analyzing the cargo's mass distribution with preset precision, and accurately calculating the initial center of gravity coordinates with minimal calculation error. Based on this, the system simulates various state changes that cargo may experience during handling: for stacked cargo, it simulates the impact of different stacking layers and stacking offsets on the overall center of gravity, generating a dynamic curve of the center of gravity changing with the stacking state; for irregularly shaped cargo, it simulates possible posture adjustments during handling, calculating the center of gravity offset under different postures; simultaneously, it considers factors such as slight shaking and local load changes during cargo loading, dynamically correcting the center of gravity parameters. The initial center of gravity coordinates, dynamic change curves, and correction rules are integrated to form complete cargo sub-parameters, providing accurate data support for cargo protection constraints in path generation.

[0035] Step S203: Import the hardware parameters of the handling robot as constraint sub-parameters for path generation in the digital twin scenario, forming the device sub-parameters in the digital twin scenario parameters.

[0036] In some embodiments, the device sub-parameters must comprehensively cover the robot's hardware characteristics, providing a strict constraint basis for path generation. The core hardware parameters of the handling robot are imported, including mechanical structure parameters and performance parameters: mechanical structure parameters cover the number of joints, the range of motion of each joint, link lengths, etc., used to construct the robot's kinematic model; performance parameters include the rated torque of each joint, maximum speed, upper limit of acceleration, motor power, endurance, etc., clearly defining the robot's operational capability boundaries. Simultaneously, the robot's perception and execution-related parameters are imported, such as the detection range and accuracy of obstacle avoidance sensors, and the grip characteristics of the drive wheels. These parameters are categorized and organized, and key constraints for path generation, such as rated joint torque, maximum acceleration, and maximum speed, are selected to form the device sub-parameters. These sub-parameters will serve as hard constraints in the path generation process, ensuring that the planned path does not exceed the robot's hardware performance limits, avoiding equipment damage or operational malfunctions due to parameter exceeding limits.

[0037] Step S204: Integrate the environmental sub-parameters, cargo sub-parameters, and equipment sub-parameters into digital twin scene parameters. Periodically compare the digital twin scene parameters with the real-time environmental parameters of the physical scene and the hardware parameters of the physical robot to correct the deviation of the digital twin scene parameters in order to maintain dynamic consistency.

[0038] In some embodiments, the integration and maintenance of digital twin scene parameters must ensure integrity and dynamic consistency. A structured approach is adopted to integrate environmental sub-parameters, cargo sub-parameters, and equipment sub-parameters into unified digital twin scene parameters. A parameter association mapping table is established to clarify the interrelationships between different sub-parameters, such as the impact of cargo center of gravity changes on robot joint torque requirements and the constraint of environmental friction coefficient on movement speed. A periodic comparison cycle is set, and the system automatically extracts real-time environmental parameters of the physical scene and hardware operating parameters of the physical robot periodically, comparing them one by one with the digital twin scene parameters. For environmental sub-parameters, the focus is on comparing easily changeable indicators such as friction coefficient and wind speed; for cargo sub-parameters, the consistency of the center of gravity parameter with the actual loading state is verified; for equipment sub-parameters, the attenuation of performance parameters such as joint torque and motor power is monitored. When parameter deviations exceed preset thresholds, an automatic correction mechanism is activated: environmental sub-parameters are directly updated based on measured data; cargo sub-parameters adjust the center of gravity coordinates based on the actual loading situation; and equipment sub-parameters correct their rated value boundaries based on performance degradation. Through continuous comparison and correction, we ensure that the parameters of the digital twin scene are always highly consistent with the physical scene and the robot's state, providing a reliable parameter basis for subsequent path generation, virtual pre-simulation, and virtual-real linkage.

[0039] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of the handling robot, as a preferred embodiment, the step of collecting real-time environmental parameters of the physical scene and importing them into the digital twin scene, and dynamically adjusting the virtual environmental parameters to form environmental sub-parameters of the digital twin scene, includes: collecting temperature and humidity data and ground friction data of the physical scene through temperature and humidity sensors and ground friction sensors deployed in the physical scene, obtaining wind speed data of the physical scene through wind speed acquisition devices, and integrating them to form real-time environmental parameters of the physical scene; automatically synchronizing the real-time environmental parameters of the physical scene to the digital twin scene, adjusting the friction coefficient parameter of the virtual environment based on the temperature and humidity data, and adjusting the air resistance parameter of the virtual environment based on the wind speed data to form environmental sub-parameters of the digital twin scene; periodically comparing the environmental sub-parameters of the digital twin scene with the real-time environmental parameters of the physical scene, setting a deviation threshold, and adjusting the environmental sub-parameters of the digital twin scene when the deviation between the environmental sub-parameters and the real-time environmental parameters exceeds the threshold, so that the deviation between the environmental sub-parameters and the real-time environmental parameters is controlled within a preset range.

[0040] Please refer to Figure 3 In the above-mentioned bidirectional mapping debugging method between the digital twin and the physical entity of the handling robot, as a preferred solution, step S102, which generates at least one set of candidate paths that satisfy multiple constraints based on the digital twin scene parameters and outputs the candidate path parameters, includes the following steps: Step S301: Extract the device sub-parameters and cargo sub-parameters from the digital twin scene parameters as constraints; where the device sub-parameters are robot hardware parameters and the cargo sub-parameters are cargo center of gravity parameters, and construct a multi-objective optimization function.

[0041] In some embodiments, the extraction of constraints needs to accurately cover the robot's operational limits and cargo characteristic requirements to ensure path feasibility. Core indicators of the equipment sub-parameters are extracted from the digital twin scene parameters, including robot hardware parameters such as the rated torque of each joint, maximum motion speed, acceleration upper limit, and motor power threshold, to clarify the mechanical performance constraint boundaries for path generation. Key data such as the cargo's center of gravity coordinates and dynamic change range of the center of gravity are extracted from the cargo sub-parameters as the core basis for cargo protection-related constraints. Based on the extracted constraints, a multi-objective optimization function is constructed. This function needs to integrate multiple core requirements of path planning, ensuring that the path is executable in physical space while also considering multiple objectives such as operational efficiency, energy consumption control, cargo safety, and equipment wear. Through mathematical modeling, the dispersed requirements are transformed into a solvable unified optimization objective, providing a quantitative computational foundation for subsequent path generation.

[0042] Step S302: The multi-objective optimization function takes path length, robot energy consumption, cargo protection requirements and equipment life constraints as the core optimization dimensions. The four core optimization dimensions correspond to the basic requirements of path planning, energy consumption control requirements, cargo safety requirements and equipment loss control requirements, respectively.

[0043] In some embodiments, the four core optimization dimensions of the multi-objective optimization function each have a clear functional positioning and computational logic, comprehensively covering the core needs of the debugging scenario. The path length dimension, as a basic requirement, is achieved by calculating the spatial distance between the robot's starting and ending points, aiming to shorten the transportation distance and improve handling efficiency while meeting constraints. The robot energy consumption dimension, based on the robot's kinematic and dynamic models, combines speed changes, acceleration fluctuations, and load distribution along the path to calculate the total energy consumption, with the core objective of reducing energy loss and extending the robot's endurance. The cargo protection requirement dimension focuses on cargo safety, using parameters such as acceleration, jerk, and turning impact along the path for quantitative evaluation to prevent cargo tipping, collisions, or damage due to sudden changes in motion. The equipment lifespan constraint dimension revolves around robot hardware wear and tear, controlling the frequency and magnitude of parameter exceedances by monitoring the ratio of parameters such as joint torque and motor current along the path to their rated values, reducing equipment wear, and extending service life. These four dimensions work together to form a comprehensive path optimization evaluation system.

[0044] Step S303: Based on the cargo weight and fragility characteristics in the cargo sub-parameters, dynamically allocate the weight ratio of each dimension of the multi-objective optimization function to achieve the adaptation of the optimization function to the cargo attributes.

[0045] In some embodiments, the weight allocation adopts a scenario-based dynamic adjustment strategy to ensure that the optimization function is accurately adapted to the cargo attributes. The system pre-establishes mapping rules between cargo attributes and weight allocation, classifying cargo into heavy, medium, and light categories based on cargo weight in the cargo sub-parameters, and labeling them as fragile, ordinary, and wear-resistant based on fragility characteristics. For heavy cargo, the weight ratio of the robot's energy consumption dimension is increased, prioritizing the control of energy consumption and joint torque load to avoid equipment overload; for fragile cargo, the weight of the cargo protection requirement dimension is increased, strictly constraining acceleration and impact parameters in the path to reduce the risk of cargo damage; for light and ordinary cargo, the dominant weight of the path length dimension is maintained, with improving handling efficiency as the core objective; for complex cargo that takes into account multiple characteristics, the weights of each dimension are dynamically balanced through weighted calculation to ensure that the path achieves optimal equilibrium among multiple objectives.

[0046] Step S304: Use a multi-objective optimization algorithm to solve the function and generate multiple sets of initial path parameters. Combine the constraint sub-parameters in the digital twin scenario parameters, i.e., the hardware rating values, to filter out the initial path parameters that do not meet the constraints, and finally output at least one set of candidate path parameters.

[0047] In some embodiments, path solving and selection must balance efficiency and reliability to ensure high-quality candidate paths are output. A mature multi-objective optimization algorithm is used to solve the function. This algorithm has efficient global search capabilities and can generate multiple sets of non-dominated solutions in complex constraint spaces, i.e., multiple sets of initial path parameters, each set of path parameters corresponding to different optimization objective equilibrium results. After the solution is completed, the initial path parameters are rigorously selected by combining the constraint sub-parameters in the digital twin scene parameters, i.e., the robot hardware specifications: the joint torques, motion speeds, accelerations, and other parameters of each node in the path are checked one by one to see if they exceed the specified range, and paths with parameters exceeding the specifications, inability to avoid obstacles, or conflicts with the physical scene space are eliminated; the overall feasibility and practicality of the path are further evaluated, and path parameters that meet all constraints and have balanced optimization objective performance are retained, and finally at least one set of candidate path parameters is output to provide high-quality samples for subsequent virtual simulation and selection.

[0048] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of the handling robot, as a preferred embodiment, the construction of the multi-objective optimization function includes: The expression for the multi-objective optimization function is: F=ω1×L+ω2×E+ω3×a+ω4×T; Where F is the comprehensive value of the optimization function, L is the path length in the candidate path parameters, and ω1 is the weight of the path length; E is the path energy consumption calculated based on the robot kinematic model and candidate path parameters, and ω2 is the weight of the path energy consumption; a is the maximum path acceleration in the candidate path parameters, and ω3 is the weight of the maximum path acceleration; T is the ratio of the maximum joint torque in the candidate path parameters to the sub-parameters of the digital twin scene equipment, where the digital twin scene equipment sub-parameters are the rated values ​​of the joint torques, and ω4 is the weight of the ratio T; the weights are adjusted according to the cargo type in the cargo sub-parameters: heavy cargo. When handling goods, if the weight parameter of the goods is greater than or equal to the set value, increase the weight of ω2 to prioritize energy consumption control; for fragile goods, increase the weight of ω3 to prioritize acceleration control; for light goods, maintain ω1 as the dominant weight to prioritize path length control; set constraint thresholds for each optimization dimension: L meets the path planning range in the digital twin scenario parameters, E is lower than the upper limit of robot single-run energy consumption in the device sub-parameters of the digital twin scenario, a does not exceed the maximum allowable acceleration for goods protection in the goods sub-parameters, and T does not exceed the set ratio to ensure that the joint torque does not exceed the rated value in the device sub-parameters.

[0049] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of the handling robot, as a preferred solution, the virtual pre-run and feasibility verification of candidate path parameters in the digital twin scenario, and the selection of the optimal path parameters, includes: loading candidate path parameters in the digital twin scenario and simulating typical abnormal scenarios; typical abnormal scenarios include sudden personnel intrusion, cargo center of gravity parameter deviation, and temporary equipment sub-parameter failure, and running the candidate path parameters; selecting paths through two core indicators: pass rate and average deviation; the pass rate is the percentage of times the candidate path parameter successfully avoids obstacles and has no parameters exceeding the limit in all abnormal scenarios, and the average deviation is the average deviation between the planned position and the actual running position of the candidate path parameter, prioritizing candidate path parameters with high pass rate and small average deviation; extracting the motion parameters of key nodes in the candidate path parameters, where key nodes are turning points and shelf docking points, and the motion parameters are joint angles, speeds, and torques, verifying whether the motion parameters conform to the hardware rated range in the equipment sub-parameters of the digital twin scenario, and if the motion parameters of a certain node exceed the rated range, automatically adjusting the motion parameters of that node and re-verifying the candidate path parameters until all node motion parameters conform to the rated range, thus obtaining the optimal path parameters.

[0050] In the aforementioned bidirectional mapping and debugging method between the digital twin and the physical entity of the handling robot, as a preferred embodiment, the establishment of a real-time data transmission link between the digital twin and the physical entity, sending the optimal path parameters to the physical robot, and simultaneously collecting the actual operating parameters of the physical robot and the real-time environmental parameters of the physical scene, and feeding them back to the digital twin scene, includes: deploying dual transmission links on the physical robot end; the first transmission link is used to transmit the optimal path parameters and the actual operating parameters of the physical robot, the actual operating parameters being position coordinates and joint torques; the second transmission link is used to transmit the real-time environmental parameters of the physical scene and high-definition images of the physical scene, the real-time environmental parameters being temperature, humidity, friction, and wind speed; the operator connects to the digital twin platform through a visual interactive device, superimposing the optimal path parameters from the digital twin scene onto the physical scene in a preset form, marking the stuttering points and offset points in the operation of the physical robot in real time, forming problem marking information, and associating it with the corresponding timestamps of the actual operating parameters; the actual operating parameters of the physical robot, the real-time environmental parameters of the physical scene, and the problem marking information are synchronized to the digital twin scene through the dual transmission links, updating the real-time data in the digital twin scene, and ensuring real-time alignment of data between the virtual end and the physical end.

[0051] In the aforementioned bidirectional mapping debugging method between the digital twin and the physical entity of a handling robot, as a preferred embodiment, the step of comparing the optimal path parameters in the digital twin scene with the actual operating parameters of the physical robot, identifying parameter deviations and analyzing their root causes, and dynamically adjusting the digital twin scene parameters and the motion parameters of the physical robot based on the root causes of the deviations to achieve bidirectional mapping optimization and complete the debugging includes: comparing the optimal path parameters in the digital twin scene with the actual operating parameters of the physical robot at set time intervals; the optimal path parameters include planned position, planned speed, and planned acceleration; the actual operating parameters are actual position, actual speed, and actual acceleration, and calculating the parameter deviation value; when the parameter deviation value exceeds a set threshold, triggering deviation analysis and retrieving the deviation node pair. The system analyzes the actual operating parameters, real-time environmental parameters of the physical scene, and parameters of the digital twin scene. Actual operating parameters include joint torque and motion speed. Real-time environmental parameters of the physical scene include temperature, humidity, and friction. A fusion model is used to identify the root causes of deviations. Parameters are adjusted based on these root causes: if the deviation originates from torque exceeding a threshold (i.e., the joint torque in the actual operating parameters exceeds the rated value of the digital twin scene device sub-parameters), a motion speed adjustment command is sent via the transmission bus to correct the motion parameters of the physical robot; if the deviation originates from environmental changes (i.e., the deviation between the real-time environmental parameters of the physical scene and the environmental sub-parameters of the digital twin scene exceeds the limit), the environmental sub-parameters of the digital twin scene are updated. The adjusted parameter deviation values ​​are continuously monitored until the deviation returns to the set range, completing the debugging process.

[0052] In the aforementioned bidirectional mapping and debugging method for the digital twin and physical entity of a handling robot, as a preferred embodiment, the step of retrieving the actual operating parameters, real-time environmental parameters of the physical scene, and digital twin scene parameters corresponding to the deviation node, and using a fusion model to identify the root cause of the deviation, includes: recording multi-source data of the deviation node with a set precision: actual operating parameters, real-time environmental parameters of the physical scene, and digital twin scene parameters, while associating the timestamp of the problem marker information; wherein, the actual operating parameters include the actual value of joint torque, the actual value of movement speed, and the parameter deviation value; the real-time environmental parameters of the physical scene include the temperature and humidity, ground friction, and wind speed at the time of the deviation; the digital twin scene parameters include the environmental sub-parameters and equipment sub-parameters at the corresponding time; in the fusion model, the data is first based on the digital twin scene parameters and the real-time data of the physical scene. The input data is coarsely classified based on the comparison results of environmental parameters, the deviation values ​​between actual operating parameters and optimal path parameters, and is initially determined to include environmental deviation, equipment parameter, or path planning categories. Then, based on similar deviation cases in historical debugging data, and by associating these cases with the correspondence between actual operating parameters, real-time environmental parameters, and the root causes of the deviations, the coarse classification results are further refined, outputting specific root causes of deviations and their confidence levels. Environmental deviations caused by changes in ground friction can be considered as one specific root cause, with the confidence level determined based on the calculation results. A multi-source data log is established to store digital twin scene parameters, real-time environmental parameters of the physical scene, actual operating parameters of the physical robot, parameter deviation values, deviation analysis results, and parameter adjustment records. The log includes timestamps and data hash values ​​for subsequent deviation backtracking and debugging optimization reference.

[0053] Extending from the above embodiments, this application provides a more specific embodiment applied to a shelf handling scenario in e-commerce logistics warehousing. The core hardware includes an AGV handling robot, temperature and humidity sensors, ground friction sensors, wind speed data acquisition equipment, industrial-grade AR glasses, a CXL switch, and a GPU+FPGA heterogeneous computing platform. The AGV handling robot has a built-in EtherCAT bus and 5GSA transmission module, supporting real-time acquisition of data such as position coordinates and joint torques. The digital twin platform is based on a GPU+FPGA heterogeneous computing architecture, with the GPU responsible for scene rendering and path visualization, and the FPGA responsible for parallel algorithm computation. The sensor sampling frequency meets the real-time data acquisition requirements, and the AR glasses support accurate overlay of virtual paths and physical scenes. The specific implementation steps are as follows: Step 1: Constructing a digital twin scenario Parameter Acquisition and Import: Temperature and humidity sensors, ground friction sensors, and wind speed acquisition devices collect temperature and humidity data, friction data, and wind speed data from the physical scene, respectively. This data is synchronized to the digital twin platform via the network. The platform automatically adjusts the friction coefficient and air resistance parameters of the virtual environment to form environmental sub-parameters. CAD models of three types of goods—cardboard boxes, metal parts, and glass products—are imported. The center of gravity coordinates of each goods are obtained through a center of gravity calculation algorithm, simulating changes in the center of gravity under different stacking layers to form goods sub-parameters. Hardware parameters such as the joint torque rating, maximum speed, and maximum acceleration of the AGV robot are imported to form equipment sub-parameters.

[0054] Initial mapping and consistency maintenance: Integrate three types of sub-parameters to form digital twin scene parameters, and establish an initial mapping relationship between the digital twin and the physical robot. The virtual scene parameters are compared with the measured data of the physical scene every hour. When the deviation of the environmental sub-parameters exceeds 5%, adjustments are automatically triggered to ensure dynamic consistency between the virtual and physical scenes.

[0055] Step 2: Generate candidate path parameters Multi-objective optimization function construction: Path length, energy consumption, acceleration, and joint torque are used as optimization dimensions, with basic weights set as ω1=0.2, ω2=0.3, ω3=0.3, and ω4=0.2. For glass products (fragile goods), ω3 is adjusted to 0.5; for metal parts (heavy goods), ω2 is adjusted to 0.4; and for cardboard boxes (light goods), ω1 is maintained as the dominant weight. Constraint thresholds are set for each dimension: path length does not exceed the planned range between warehouse shelves, energy consumption is lower than the robot's single-cycle energy consumption limit, acceleration does not exceed the allowable value for fragile goods protection, and the joint torque ratio does not exceed 1.0.

[0056] Path generation and filtering: The NSGA-II algorithm is used to solve the function and generate 10 sets of initial path parameters. Combined with constraints such as the rated value of robot joint torque, 3 sets of parameters that do not meet the requirements are eliminated, and finally 7 sets of candidate path parameters are output.

[0057] Step 3: Virtual simulation to select the optimal path Abnormal scenario simulation: Load 7 sets of candidate path parameters into the digital twin scenario to simulate abnormal scenarios such as sudden personnel intrusion, cargo center of gravity shift of 100mm, and joint speed decrease of 20%, and run each candidate path.

[0058] Path selection and verification: Paths were selected based on pass rate and average deviation. Three paths achieved a 100% pass rate, with average deviations all less than 2mm. Motion parameters for turning points and shelf docking points were extracted from these three paths. Verification revealed that the joint torque at a certain turning point in one path exceeded the rated value. After automatically adjusting the motion time of that node, the path passed re-verification, ultimately determining the optimal path parameters.

[0059] Step 4: Real-time data transmission and feedback Link Deployment: Deploy dual transmission links of EtherCAT bus and 5GSA at the AGV robot end. EtherCAT bus transmits real-time control data such as optimal path parameters, position coordinates, and joint torque, while 5GSA transmits high-definition images of the physical scene and environmental data such as temperature, humidity, and friction.

[0060] Visual interaction and data synchronization: Operators wear AR glasses to overlay the optimal path as highlighted lines onto the physical scene, observe the robot's operating status in real time, and mark bottlenecks at shelf corners. Actual operating parameters, environmental parameters, and bottleneck mark information are synchronized to the digital twin scene via dual links, updating the virtual data.

[0061] Step 5: Deviation Analysis and Two-Way Optimization Deviation identification: The system compares the virtual and real path parameters every 10ms. If the actual position of the bottleneck point deviates from the planned position by 3mm, which exceeds the set threshold of 2mm, deviation analysis is triggered.

[0062] Root cause analysis: By retrieving the actual joint torque value, ground friction data, and corresponding parameters of the virtual scene for this node, and through fusion model analysis, it was determined that the root cause of the deviation was that the joint torque exceeded the threshold.

[0063] Parameter Adjustment and Verification: Commands were sent to the robot via the EtherCAT bus to reduce the speed at the corner from 0.5 m / s to 0.3 m / s, while simultaneously updating the corresponding parameters in the digital twin scene. Continuous monitoring showed that the deviation was reduced to within 1 mm after adjustment, meeting the requirements. A multi-source data log was established to store relevant data from the entire debugging process, completing this debugging operation.

[0064] Please refer to Figure 4 The embodiments of this application also provide a bidirectional mapping and debugging system for a handling robot's digital twin and its physical entity, the system comprising: A digital twin scene construction module is used to collect real-time environmental parameters of the physical scene, cargo center of gravity parameters and robot hardware parameters, generate digital twin scene parameters, construct a digital twin scene of the handling robot, and establish an initial mapping relationship between the digital twin and the physical entity. The candidate path generation module is used to construct a multi-objective optimization function based on the equipment sub-parameters and cargo sub-parameters in the digital twin scenario parameters, solve to generate multiple sets of initial path parameters, and filter and output at least one set of candidate path parameters. The virtual pre-simulation and screening module is used to perform virtual pre-simulation and feasibility verification on candidate path parameters in a digital twin scenario, and to screen the optimal path parameters by comparing the verification results. The real-time data transmission module is used to establish a real-time data transmission link between the digital twin and the physical entity, issue optimal path parameters, and synchronously collect and feedback the actual operating parameters of the physical robot and the real-time environmental parameters of the physical scene. The bidirectional mapping optimization module is used to compare the optimal path parameters with the actual operating parameters, identify parameter deviations and analyze their root causes, and adjust the environmental sub-parameters, equipment sub-parameters and physical robot motion parameters in the digital twin scene parameters based on the root causes to achieve bidirectional mapping optimization and complete the debugging.

[0065] Those skilled in the art will understand that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the program can also be stored in a server, another computer, disk, optical disk, flash drive, or external hard drive, etc., and can be downloaded or copied to the memory of a local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be achieved.

[0066] The above examples illustrate this application only to aid understanding and are not intended to limit its scope. Those skilled in the art to which this application pertains can make various simple deductions, modifications, or substitutions based on the ideas presented.

Claims

1. A bidirectional mapping and debugging method for a handling robot between a digital twin and its physical entity, characterized in that, The method comprises the following steps: A digital twin of the handling robot is constructed, digital twin scene parameters are generated, and an initial mapping relationship between the digital twin and the physical entity is established; At least one set of candidate paths that meet multiple constraint conditions is generated based on the digital twin scene parameters, and candidate path parameters are output; The candidate path parameters are virtually preformed and feasibility verified in the digital twin scene, and optimal path parameters are selected; A real-time data transmission link between the digital twin and the physical entity is established, the optimal path parameters are sent to the physical robot, actual operation parameters of the physical robot and real-time environmental parameters of the physical scene are synchronously collected, and the collected parameters are fed back to the digital twin scene; The optimal path parameters in the digital twin scene are compared with the actual operation parameters of the physical robot, parameter deviations are identified and the root causes of the deviations are analyzed, the digital twin scene parameters and the motion parameters of the physical robot are dynamically adjusted based on the root causes of the deviations, the bidirectional mapping optimization between the digital twin and the physical entity is realized, and the debugging is completed.

2. The bi-directional mapping debugging method of a digital twin of a transport robot and a physical entity according to claim 1, wherein, The method comprises the following steps: Real-time environmental parameters of the physical scene are collected and imported into the digital twin scene, virtual environmental parameters are dynamically adjusted, and environmental sub-parameters in the digital twin scene parameters are formed; A cargo model is imported and cargo gravity center parameters are calculated, the influence of cargo state changes on the gravity center parameters is simulated, and cargo sub-parameters in the digital twin scene parameters are formed; Hardware parameters of the handling robot are imported and used as constraint sub-parameters for path generation in the digital twin scene, and device sub-parameters in the digital twin scene parameters are formed; The environmental sub-parameters, the cargo sub-parameters, and the device sub-parameters are integrated into the digital twin scene parameters, the digital twin scene parameters are compared with the real-time environmental parameters of the physical scene and the hardware parameters of the physical robot at regular intervals, and the digital twin scene parameter deviations are corrected to maintain dynamic consistency. 3.The method of claim 2, wherein, The method comprises the following steps: Temperature and humidity data and ground friction data of the physical scene are collected through temperature and humidity sensors and ground friction sensors deployed in the physical scene, wind speed data of the physical scene are collected through a wind speed collection device, and real-time environmental parameters of the physical scene are integrated and formed; The real-time environmental parameters of the physical scene are automatically synchronized to the digital twin scene, the friction coefficient parameters of the virtual environment are adjusted based on the temperature and humidity data, the air resistance parameters of the virtual environment are adjusted based on the wind speed data, and environmental sub-parameters of the digital twin scene are formed; The environmental sub-parameters of the digital twin scene are compared with the real-time environmental parameters of the physical scene at regular intervals, a deviation threshold is set, and when the deviation between the environmental sub-parameters and the real-time environmental parameters exceeds the threshold, the environmental sub-parameters of the digital twin scene are adjusted so that the deviation between the environmental sub-parameters and the real-time environmental parameters is controlled within a preset range.

4. The bi-directional mapping debugging method of a digital twin of a transport robot and a physical entity according to claim 2, wherein, The method comprises the following steps: Extract the equipment sub-parameters and cargo sub-parameters in the digital twin scene parameters as constraint conditions; wherein the equipment sub-parameters are robot hardware parameters, and the cargo sub-parameters are cargo gravity center parameters, and a multi-objective optimization function is constructed; The multi-objective optimization function takes path length, robot energy consumption, cargo protection requirements, and equipment life constraints as core optimization dimensions, and the four core optimization dimensions correspond to the basic requirements of path planning, energy consumption control requirements, cargo safety requirements, and equipment wear control requirements, respectively; According to the cargo weight and fragility characteristics in the cargo sub-parameters, the weight proportion of each dimension of the multi-objective optimization function is dynamically allocated to realize the adaptation of the optimization function to the cargo attributes; A multi-objective optimization algorithm is used to solve the function to generate multiple sets of initial path parameters, and the constraint sub-parameters in the digital twin scene parameters, i.e., the hardware rating, are combined for screening to eliminate initial path parameters that do not meet the constraint conditions, and finally at least one set of candidate path parameters is output. 5.The method of claim 4, wherein, The multi-objective optimization function is constructed, including: The multi-objective optimization function expression is: F = ω1 × L + ω2 × E + ω3 × a + ω4 × T; Wherein F is the comprehensive value of the optimization function, L is the path length in the candidate path parameters, ω1 is the weight of the path length; E is the path energy consumption calculated based on the robot kinematics model and the candidate path parameters, ω2 is the weight of the path energy consumption; a is the maximum acceleration of the path in the candidate path parameters, ω3 is the weight of the maximum acceleration of the path; T is the ratio of the maximum joint torque of the path in the candidate path parameters to the digital twin scene equipment sub-parameters, and the digital twin scene equipment sub-parameters are the joint torque rating, ω4 is the weight of the ratio T; According to the cargo type in the cargo sub-parameters, the weight is adjusted: when heavy cargo is transported, the cargo weight parameter is greater than or equal to a set value, the weight proportion of ω2 is increased to prioritize energy consumption control; when fragile goods are transported, the weight proportion of ω3 is increased to prioritize acceleration control; when light cargo is transported, ω1 is maintained as the dominant weight to prioritize path length control; The constraint thresholds of each optimization dimension are set: L satisfies the path planning range in the digital twin scene parameters, E is lower than the upper limit of the single endurance energy consumption of the robot in the digital twin scene equipment sub-parameters, a does not exceed the maximum acceleration allowed by the cargo protection in the cargo sub-parameters, and T does not exceed a set value to ensure that the joint torque does not exceed the rating in the equipment sub-parameters. 6.The method of claim 2, wherein, The virtual rehearsal and feasibility verification of the candidate path parameters in the digital twin scene are performed to obtain the optimal path parameters, including: Load the candidate path parameters in the digital twin scene, simulate typical abnormal scenes, run the candidate path parameters, and the typical abnormal scenes are sudden intrusions, cargo gravity center parameter deviations, and temporary equipment sub-parameter failures; Select the path by the pass rate and the average deviation: the pass rate is the proportion of the number of times that the candidate path parameters successfully avoid obstacles and do not exceed the parameters in all abnormal scenes, and the average deviation is the average deviation between the planned position and the actual running position of the candidate path parameters, and the candidate path parameters with high pass rate and small average deviation are preferentially selected; The key node motion parameters in the candidate path parameters are extracted, the key nodes being turning points and shelf docking points; the motion parameters including joint angles, speeds and torques; whether the motion parameters conform to the hardware rated ranges in the digital twin scene device sub-parameters is verified; if the motion parameters of a node exceed the rated ranges, the motion parameters of the node are automatically adjusted and the candidate path parameters are re-verified until the motion parameters of all nodes conform to the rated ranges, and the optimal path parameters are obtained.

7. The bi-directional mapping debugging method of a digital twin of a transport robot and a physical entity according to claim 2, wherein, The real-time data transmission link between the digital twin and the physical entity is established, the optimal path parameters are sent to the physical robot, the actual operation parameters of the physical robot and the real-time environment parameters of the physical scene are synchronously collected and fed back to the digital twin scene, including: A double transmission link is deployed at the end of the physical robot, the first transmission link being used for transmitting the optimal path parameters and the actual operation parameters of the physical robot, the actual operation parameters being position coordinates and joint torques; the second transmission link being used for transmitting the real-time environment parameters of the physical scene and high-definition images of the physical scene, the real-time environment parameters being temperature and humidity, friction and wind speed; An operator connects the digital twin platform through a visual interactive device, superimposes the optimal path parameters in the digital twin scene onto the physical scene in a preset form, real-time marks the lag points and deviation points in the operation of the physical robot, forms problem marking information and associates corresponding actual operation parameter time stamps; The actual operation parameters of the physical robot, the real-time environment parameters of the physical scene and the problem marking information are synchronously fed back to the digital twin scene through the double transmission link, the real-time data in the digital twin scene is updated, and the data of the virtual end and the physical end are real-time aligned. 8.The method of claim 2, wherein, The optimal path parameters in the digital twin scene are compared with the actual operation parameters of the physical robot, parameter deviations are identified and deviation sources are analyzed, the digital twin scene parameters and the motion parameters of the physical robot are dynamically adjusted based on the deviation sources, bidirectional mapping optimization is realized, and debugging is completed, including: The optimal path parameters in the digital twin scene are compared with the actual operation parameters of the physical robot at a set time interval; the optimal path parameters including planned positions, planned speeds and planned accelerations; the actual operation parameters being actual positions, actual speeds and actual accelerations, and parameter deviation values are calculated; When the parameter deviation values exceed a set threshold, deviation analysis is triggered, the actual operation parameters of the deviation nodes, the real-time environment parameters of the physical scene and the digital twin scene parameters are called; the actual operation parameters including joint torques and motion speeds; the real-time environment parameters of the physical scene including temperature and humidity and friction; a fusion model is used to identify the deviation sources; Parameters are adjusted according to the deviation sources: if the deviation is caused by a torque exceeding a threshold, i.e. the joint torque in the actual operation parameters exceeds the rated value of the digital twin scene device sub-parameters, a motion speed adjustment instruction is sent through a transmission bus to correct the motion parameters of the physical robot; if the deviation is caused by environmental changes, i.e. the real-time environment parameters of the physical scene deviate from the environment sub-parameters of the digital twin scene beyond the limit, the environment sub-parameters of the digital twin scene are updated; the adjusted parameter deviation values are continuously monitored until the deviation returns to the set range, and the debugging is completed. 9.The method of claim 8, wherein, The real-time environment parameters of the actual operation parameters and the physical scene corresponding to the deviation node and the digital twin scene parameters adopt a fusion model to identify the deviation root, including: Record the multi-source data of the deviation node according to the set accuracy: actual operation parameters, real-time environment parameters of the physical scene, digital twin scene parameters, and time stamp of associated problem mark information; wherein the actual operation parameters include joint torque actual value, motion speed actual value, and parameter deviation value; the real-time environment parameters of the physical scene include temperature and humidity, ground friction, and wind speed at the time of deviation occurrence; the digital twin scene parameters include environment sub-parameters and equipment sub-parameters at the corresponding time; In the fusion model, the input data is first roughly classified based on the comparison results of the digital twin scene parameters and the real-time environment parameters of the physical scene, and the deviation value of the actual operation parameters and the optimal path parameters, and the rough classification result is preliminarily determined, including environment deviation class, equipment parameter class or path planning class; Then, based on the same type of deviation cases in the historical debugging data, the same type of deviation cases are associated with the corresponding relationship between the actual operation parameters, the real-time environment parameters and the deviation root, the rough classification result is finely classified, the specific deviation root and the confidence are output, the environment deviation class caused by the change of ground friction can be used as one of the specific deviation roots, and the confidence is determined according to the calculation result; A multi-source data log is established to store digital twin scene parameters, real-time environment parameters of the physical scene, actual operation parameters of the physical robot, parameter deviation value, deviation analysis result and parameter adjustment record, the log contains time stamp and data hash value, which is used for subsequent deviation backtracking and debugging optimization reference. 10.A two-way mapping commissioning system of a digital twin of a material handling robot and a physical entity, the system comprising: The system comprises: A digital twin scene construction module, which is used to collect real-time environment parameters of a physical scene, a center of gravity parameter of goods and a hardware parameter of a robot, generate digital twin scene parameters, construct a digital twin scene of a carrying robot, and establish an initial mapping relationship between a digital twin and a physical entity; A candidate path generation module, which is used to construct a multi-objective optimization function based on equipment sub-parameters and goods sub-parameters in the digital twin scene parameters, solve and generate a plurality of initial path parameters, and screen and output at least one set of candidate path parameters; A virtual preview screening module, which is used to perform virtual preview and feasibility verification on the candidate path parameters in the digital twin scene, compare and verify the results to screen and obtain optimal path parameters; A real-time data transmission module, which is used to establish a real-time data transmission link between the digital twin and the physical entity, issue the optimal path parameters, and synchronously collect and feedback actual operation parameters of the physical robot and real-time environment parameters of the physical scene; A bidirectional mapping optimization module, which is used to compare the optimal path parameters with the actual operation parameters, identify parameter deviation and analyze the root, adjust environment sub-parameters, equipment sub-parameters in the digital twin scene parameters and motion parameters of the physical robot based on the root, realize bidirectional mapping optimization, and complete debugging.

Images