Digital twinning-based intelligent system with body and construction method thereof

Abstract

The invention relates to the technical field of digital twinning and body intelligent simulation, in particular to a digital twinning-based body intelligent system and a construction method thereof. The method comprises the following steps: acquiring industrial manufacturing original data and a manufacturing process monitoring original image by using an intelligent sensor; performing data standardization processing on the industrial manufacturing original data to obtain industrial manufacturing data; performing image enhancement processing on the manufacturing process monitoring original image to obtain a manufacturing monitoring image; performing manufacturing parameter extraction on the industrial manufacturing data to generate industrial manufacturing parameters; and performing manufacturing scene mapping on the manufacturing monitoring image based on the industrial manufacturing parameters to generate industrial manufacturing scene data. Through the data twinning technology, the image processing technology and the deep learning technology, digital twinning fitting is carried out according to the industrial manufacturing feature data, scene simulation operation of the intelligent system with the body is carried out, and therefore the applicability and the automation degree of intelligent operation with the body are improved.

Description

Technical Field

[0001] The present invention relates to the field of digital twin and embodied intelligent simulation technology, and in particular to an embodied intelligent system based on digital twin and a construction method thereof. Background Art

[0002] Embodied intelligence refers to the ability of intelligent systems to interact with the environment through physical entities to achieve the integration of perception, decision-making and action; however, the development of this field faces two major technical challenges: insufficient data and high difficulty in data acquisition. In embodied intelligent systems, in order to achieve efficient interaction and decision-making, a large amount of data is required to train and optimize algorithms; however, since embodied intelligent systems usually need to operate in complex and changeable actual environments, collecting sufficient and high-quality data has become a problem; in addition, these data often need to be obtained from multiple sensors and devices, involving data synchronization, fusion and preprocessing, which further increases the difficulty of data acquisition; the generalization ability of models trained by traditional data collection methods is limited. Digital twin technology is to create a virtual model that completely corresponds to the physical entity and simulate the behavior of the entity and the interaction of the environment in a virtual environment, thereby generating a large amount of simulated data. However, the existing digital twin embodied intelligence technology, in the field of industrial manufacturing, fails to fit the digital twin according to the industrial manufacturing characteristic data, and it is difficult to perform scene simulation operations of the embodied intelligent system, resulting in low applicability and automation of embodied intelligent operations. Summary of the invention

[0003] Based on this, it is necessary to provide an embodied intelligent system based on digital twins and a construction method thereof to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above purpose, a method for constructing embodied intelligence based on digital twins is provided, the method comprising the following steps: Step S1: using intelligent sensors to obtain industrial manufacturing raw data and manufacturing process monitoring raw images; performing data standardization processing on the industrial manufacturing raw data to obtain industrial manufacturing data; performing image enhancement processing on the manufacturing process monitoring raw images to obtain manufacturing monitoring images; Step S2: extract manufacturing parameters from industrial manufacturing data to generate industrial manufacturing parameters; perform manufacturing scene mapping on manufacturing monitoring images based on industrial manufacturing parameters to generate industrial manufacturing scene data; perform digital twin fitting based on industrial manufacturing parameters, manufacturing monitoring images and industrial manufacturing scene data to construct an industrial manufacturing digital twin framework; Step S3: Based on the industrial manufacturing digital twin framework, the kinematics of the robot arm is solved to obtain the motion parameters of the robot arm; an embodied intelligent model is constructed for the motion parameters of the robot arm to obtain the embodied intelligent model; the industrial manufacturing operation of the robot arm is simulated according to the embodied intelligent model to generate the simulation operation data of the robot arm; the industrial simulation operation index detection of the simulation operation data of the robot arm is performed using the industrial manufacturing digital twin framework to obtain the industrial simulation index of the robot arm; Step S4: Verify the simulation operation performance of the robotic arm industrial simulation indicators and generate a robotic arm simulation performance report; dynamically simulate and adjust the embodied intelligent model based on the robotic arm simulation performance report to complete the embodied intelligence construction task of the digital twin.

[0005] The present invention uses the original industrial manufacturing data and the original images of manufacturing process monitoring obtained by intelligent sensors, and after data standardization and image enhancement processing, ensures the consistency of data and the clarity of images, thereby improving the accuracy of manufacturing process monitoring. Parameters are extracted from industrial manufacturing data to generate industrial manufacturing parameters, and scene mapping is performed on manufacturing monitoring images based on these parameters; digital twin fitting is performed in combination with industrial manufacturing parameters, manufacturing monitoring images and scene data, and an industrial manufacturing digital twin framework is constructed. The digital mapping of the manufacturing process is realized, and an accurate digital model is provided for the subsequent kinematic solution and simulation operation of the robot arm, which enhances the visualization and control capabilities of the manufacturing process. The kinematic solution of the robot arm is performed based on the industrial manufacturing digital twin framework, the motion parameters of the robot arm are obtained, and an embodied intelligent model is constructed. The industrial manufacturing operation simulation of the robot arm is performed through the embodied intelligent model, the simulation operation data is generated, and the industrial simulation operation index detection is performed to obtain the industrial simulation index of the robot arm. The accurate simulation and performance evaluation of the robot arm operation are realized, and reliable simulation data and performance indexes are provided for the practical application of the robot arm, and the operation efficiency and accuracy of the robot arm are optimized. The performance of the industrial simulation indicators of the robotic arm is verified, a robotic arm simulation performance report is generated, and the embodied intelligent model is dynamically simulated and adjusted based on this report. The accuracy and practicality of the digital twin model are ensured. Through continuous performance verification and model adjustment, the embodied intelligent construction operation of the digital twin is completed, and the adaptability and reliability of the robotic arm operation are improved. Therefore, the present invention uses data twin technology, image processing technology and deep learning technology to perform digital twin fitting according to industrial manufacturing feature data, and performs scene simulation operations of the embodied intelligent system, thereby improving the applicability and automation of embodied intelligent operations.

[0006] Preferably, step S2 comprises the following steps: Step S21: extracting manufacturing parameters from industrial manufacturing data to generate industrial manufacturing parameters, wherein the industrial manufacturing parameters include material parameters, equipment parameters, and process parameters; Step S22: classify the material parameters by characteristics to obtain the material type identification; match the equipment parameters with the equipment operation mode to obtain the equipment category label; determine the production process quantity of the process parameters to obtain the process flow type; Step S23: classifying the industrial manufacturing data into industrial manufacturing types according to the material type identification, equipment category label and process flow type to obtain the industrial manufacturing type; Step S24: performing image parameterization marking on the manufacturing monitoring image based on the industrial manufacturing parameters to generate parameterized image features; performing manufacturing scene mapping on the parameterized image features according to the industrial manufacturing type to generate industrial manufacturing scene data; Step S25: Perform digital twin fitting based on industrial manufacturing parameters, manufacturing monitoring images, and industrial manufacturing scene data to build an industrial manufacturing digital twin framework.

[0007] The present invention conducts in-depth analysis on industrial manufacturing data and extracts industrial manufacturing parameters including material parameters, equipment parameters and process parameters, so that the key characteristics of the manufacturing process can be quantified, thereby improving the controllability and predictability of the manufacturing process; the extracted industrial manufacturing parameters are further classified and matched to obtain material type identification, equipment category label and process flow type. The classification of manufacturing parameters has been refined, so that each link of the manufacturing process can be identified and distinguished more accurately, providing conditions for the optimization and automation of the manufacturing process; industrial manufacturing data is classified according to material type identification, equipment category label and process flow type to obtain industrial manufacturing types, realizing the systematic classification of manufacturing data, and providing a basis for the standardization and modularization of the manufacturing process; manufacturing monitoring images are parameterized and marked based on industrial manufacturing parameters to generate parametric image features, and these features are mapped to manufacturing scenes according to industrial manufacturing types, and image data is combined with manufacturing parameters to improve the recognition accuracy and visualization level of manufacturing scenes, providing rich data support for the construction of the digital twin framework; digital twin fitting is performed based on industrial manufacturing parameters, manufacturing monitoring images and industrial manufacturing scene data, and an industrial manufacturing digital twin framework is constructed to realize the digital mapping and virtualization of the manufacturing process.

[0008] Preferably, step S25 includes the following steps: Step S251: performing feature synchronization on industrial manufacturing parameters to obtain synchronized manufacturing parameters; performing scene matching on manufacturing monitoring images to obtain image scene matching data; Step S252: aligning manufacturing scene features with the synchronized manufacturing parameters and the image scene matching data to generate scene feature alignment data; fusing feature parameters of the industrial manufacturing scene data according to the scene feature alignment data to obtain fused scene data; Step S253: performing industrial manufacturing digital fitting on the fused scene data to obtain an industrial manufacturing digital fitting set; Step S254: constructing a digital twin framework for the industrial manufacturing digital fitting set to obtain a preliminary digital twin framework; Step S255: Perform industrial manufacturing feature mapping on the preliminary digital twin framework to obtain a mapped digital twin framework; perform industrial manufacturing digital integration on the mapped digital twin framework to generate an industrial manufacturing digital twin framework.

[0009] The present invention performs feature synchronization of industrial manufacturing parameters and scene matching of manufacturing monitoring images, ensuring the consistency of parameters and image data in time and space, providing a basis for accurate alignment of manufacturing scene features, and improving the synchronization and reliability of data; aligning the manufacturing scene features of synchronized manufacturing parameters with image scene matching data, and fusion of feature parameters of industrial manufacturing scene data based on these data, realizing the deep integration of manufacturing parameters and image data, enhancing the integrity and accuracy of scene data, and laying a solid foundation for building an accurate digital twin framework; performing industrial manufacturing digital fitting on the fused scene data, simulating the actual manufacturing process by digital means, generating a set of digital sets that can reflect the actual manufacturing scene, and providing an accurate digital model for the subsequent construction of the digital twin framework; constructing a digital twin framework for the industrial manufacturing digital fitting set, and obtaining a preliminary digital twin framework. The digital fitting set is converted into a preliminary digital framework, providing a basic architecture for the virtualization and visualization of the industrial manufacturing process; industrial manufacturing feature mapping is performed on the preliminary digital twin framework, obtaining a mapping digital twin framework, and industrial manufacturing digital integration is performed on the mapping digital twin framework to generate an industrial manufacturing digital twin framework; the transformation from the preliminary framework to the complete digital twin framework is completed.

[0010] Preferably, step S3 comprises the following steps: Step S31: performing inverse kinematics solution on the main arm based on the industrial manufacturing digital twin framework to obtain the main arm motion parameters; Step S32: performing forward kinematics solution on the slave arm according to the motion parameters of the master arm to obtain the motion parameters of the slave arm; Step S33: integrating the master arm motion parameters and the slave arm motion parameters to generate the robot arm motion parameters; Step S34: constructing an embodied intelligent model for the robot arm motion parameters to obtain an embodied intelligent model; Step S35: simulating the industrial manufacturing operation of the robot arm according to the embodied intelligent model to generate robot arm simulation operation data; Step S36: using the industrial manufacturing digital twin framework to identify the robot arm behavior results on the robot arm simulation operation data to obtain the robot arm behavior results; Step S37: Determine the simulation task operation amount of the robot arm behavior result to generate the total simulation task operation amount; record the task time of the simulation task operation amount to obtain the total simulation task time; perform simulation operation index detection based on the total simulation task operation amount and the total simulation task time to obtain the robot arm industrial simulation index.

[0011] The present invention solves the inverse kinematics of the main arm based on the industrial manufacturing digital twin framework to obtain the motion parameters of the main arm. The accurate calculation of the motion trajectory of the main arm is achieved, which provides key motion parameters for the subsequent coordinated actions of the robot arm and ensures the accuracy and reliability of the robot arm operation; the forward kinematics of the slave arm is solved according to the motion parameters of the master arm to ensure that the slave arm can make corresponding motion adjustments according to the motion state of the master arm, achieve coordination between the various parts of the robot arm, and improve the overall efficiency of the robot arm operation; the motion parameters of the various parts of the robot arm are unified and integrated to provide complete data support for the overall motion control of the robot arm, and enhance the coordination and consistency of the robot arm operation; the embodied intelligent model is constructed for the motion parameters of the robot arm to obtain the embodied intelligent model. The motion parameters of the robot arm are converted into an embodied intelligent model, which realizes the intelligence of the robot arm operation and improves the adaptability and flexibility of the robot arm to complex operating environments; the industrial manufacturing operation of the robot arm is simulated according to the embodied intelligent model to generate robot arm simulation operation data. The feasibility of the robot operation was verified through simulation, which provided a basis for preview and optimization of the actual robot operation and reduced the operation risk. The industrial manufacturing digital twin framework was used to identify the robot behavior results of the robot simulation operation data and obtain the robot behavior results. The accurate identification of the robot simulation operation results was achieved, which provided direct feedback for the evaluation and optimization of the robot operation and improved the accuracy of the simulation.

[0012] Preferably, step S31 includes the following steps: Step S311: extracting manufacturing scene twin information from the industrial manufacturing digital twin framework to obtain manufacturing scene twin information; perceiving manufacturing operation features from the manufacturing scene twin information to generate industrial manufacturing operation features; Step S312: performing operation posture recognition on the industrial manufacturing operation feature to obtain operation posture data; determining the position coordinates of the operation posture data to generate posture position coordinates; Step S313: dividing the posture position coordinates into a posture space range to obtain a posture space structure range; performing posture pattern recognition on the operation posture data based on the posture space structure range to generate an operation posture pattern; Step S314: matching the active joint type of the main arm according to the operation posture mode to obtain the active joint type; measuring the joint degree of freedom of the active joint type to generate the joint degree of freedom value; determining the joint rotation angle of the joint degree of freedom value to obtain the joint rotation angle degree; Step S315: identifying the main arm connecting rod feature of the main arm according to the operation posture mode to obtain the main arm connecting rod feature; mapping the main arm connecting rod feature to the connecting rod parameter according to the joint degree of freedom value to obtain the main arm connecting rod parameter value; determining the connecting rod telescopic length of the main arm connecting rod parameter value to obtain the connecting rod telescopic length; Step S316: Integrate the joint rotation angle and the connecting rod extension length into the main arm motion parameters to obtain the main arm motion parameters.

[0013] The present invention extracts manufacturing scene twin information from an industrial manufacturing digital twin framework, and perceives manufacturing operation features of the manufacturing scene twin information, thereby realizing the digital description of the manufacturing scene and the precise capture of operation features, providing detailed operation data for subsequent posture recognition and attitude pattern recognition, and enhancing the predictability and controllability of manufacturing operations; performing operation posture recognition on industrial manufacturing operation features, and determining position coordinates of the operation posture data, thereby ensuring the precise recognition and positioning of the operation posture of the robot arm, providing accurate posture information for the precise control of the robot arm, and improving the accuracy and reliability of the operation; dividing the posture position coordinates into posture space ranges; and performing attitude pattern recognition on the operation posture data based on the posture space structure range. The spatial range division and posture pattern recognition of the posture data realize the systematic classification of the robot arm's operating posture, providing structured data support for the robot arm's motion planning and execution; according to the operating posture pattern, the main arm's active joint type is matched, the joint degree of freedom of the active joint type is calculated, and the joint degree of freedom value is generated; the joint rotation angle is determined for the joint degree of freedom value, and the joint motion parameters are accurately determined by the matching of the joint type and the measurement of the degree of freedom, providing the necessary parameter information for the precise motion control of the robot arm; according to the operating posture pattern, the main arm's main arm connecting rod feature is identified to obtain the main arm connecting rod feature; according to the joint degree of freedom value, the main arm connecting rod feature is mapped to the connecting rod parameter value to obtain the main arm connecting rod parameter value; the connecting rod telescopic length is determined for the main arm connecting rod parameter value to obtain the connecting rod telescopic length. Through the identification of the connecting rod feature and parameter mapping, the connecting rod motion parameters are accurately determined, providing detailed connecting rod telescopic information for the precise motion control of the robot arm. The joint rotation angle and the connecting rod telescopic length are integrated into the main arm motion parameters to obtain the main arm motion parameters. The motion parameters of the joints and links are integrated to form a complete set of main arm motion parameters, providing comprehensive data support for the precise motion control and operation of the robot arm, and improving the flexibility and accuracy of the robot arm operation.

[0014] Preferably, step S32 includes the following steps: Step S321: Encoding the main arm motion parameters into operation instructions to generate joint rotation angle instructions and connecting rod extension length instructions; Step S322: determining the slave arm joint rotation angle of the slave arm according to the joint rotation angle instruction, generating the slave arm joint rotation angle; performing slave arm joint rotation execution based on the slave arm joint rotation angle, and obtaining slave arm joint rotation execution data; Step S323: determining the telescopic length of the slave arm link of the slave arm according to the link telescopic length instruction, generating the telescopic length of the slave arm link; performing telescopic execution of the slave arm link based on the telescopic length of the slave arm link, and obtaining telescopic execution data of the slave arm link; Step S324: performing slave arm posture feature recognition based on the slave arm joint rotation execution data and the slave arm link extension and retraction execution data to obtain slave arm posture feature data; recording the posture coordinates of the slave arm posture feature data to generate slave arm posture coordinates; tracking the slave arm execution path according to the slave arm posture coordinates to generate slave arm execution path data; Step S325: Integrate the slave arm motion parameters using the slave arm joint rotation execution data, the slave arm link extension and retraction execution data, and the slave arm execution path data to obtain the slave arm motion parameters.

[0015] The present invention encodes the motion parameters of the master arm into operation instructions, converts the complex motion parameters into executable operation instructions, provides clear instruction information for the precise control of the slave arm, improves the executability of the instructions and the accuracy of the robot arm operation; determines the slave arm joint rotation angle of the slave arm according to the joint rotation angle instruction, generates the slave arm joint rotation angle; executes the slave arm joint rotation based on the slave arm joint rotation angle, and obtains the slave arm joint rotation execution data. The precise control of the slave arm joint rotation is achieved, the accuracy of the slave arm action is ensured, and the stable operation of the slave arm is guaranteed; determines the slave arm connecting rod telescopic length of the slave arm according to the connecting rod telescopic length instruction, generates the slave arm connecting rod telescopic length; executes the slave arm connecting rod telescopic execution based on the slave arm connecting rod telescopic length, and obtains the slave arm connecting rod telescopic execution data. The precise control of the extension and retraction of the slave arm link is achieved, which ensures the coordination of the slave arm movement and improves the flexibility of the robot arm operation; the slave arm posture feature recognition is performed based on the slave arm joint rotation execution data and the slave arm link extension and retraction execution data to obtain the slave arm posture feature data; the posture coordinates of the slave arm posture feature data are recorded to generate the slave arm posture coordinates; the slave arm execution path is tracked according to the slave arm posture coordinates to generate the slave arm execution path data. Through posture feature recognition and posture coordinate recording, the precise tracking of the slave arm execution path is achieved, which improves the traceability and accuracy of the robot arm operation; the slave arm motion parameters are integrated with the slave arm joint rotation execution data, the slave arm link extension and retraction execution data and the slave arm execution path data to obtain the slave arm motion parameters. The key motion data of the slave arm is integrated to form a complete slave arm motion parameter set, which provides comprehensive data support for the precise control and operation of the slave arm and enhances the overall coordination and efficiency of the robot arm operation.

[0016] Preferably, step S34 includes the following steps: Step S341: Calculate the motion parameters of the robot arm to obtain the motion parameters of the robot arm; Step S342: dividing the robot arm motion parameter quantity according to 80% parameter quantity to obtain a robot arm motion training set; dividing the robot arm motion parameter quantity according to 20% parameter quantity to obtain a robot arm motion test set; Step S343: constructing an embodied intelligence model according to the robot arm motion parameters to obtain an embodied intelligence pre-model; and training the embodied intelligence pre-model using the robot arm motion training set to obtain an embodied intelligence training model; Step S344: Use the robotic arm motion test set to perform a model cross-validation test on the embodied intelligence training model to obtain embodied intelligence model test data; adjust the model parameters of the embodied intelligence training model using the embodied intelligence model test data to obtain the embodied intelligence model.

[0017] The present invention measures the mechanical arm motion parameters to obtain the mechanical arm motion parameters; the quantitative analysis of the motion parameters provides an accurate data basis for subsequent data set division and model training, ensuring the accuracy of data analysis and the effectiveness of model training; the mechanical arm motion parameters are divided according to 80% parameters to obtain a mechanical arm motion training set; the mechanical arm motion parameters are divided according to 20% parameters to obtain a mechanical arm motion test set. Effective data distribution is achieved, a structured data set is provided for model training and testing, and the representativeness of model training and the comprehensiveness of testing are enhanced; an embodied intelligence model is constructed according to the mechanical arm motion parameters to obtain an embodied intelligence pre-model; the embodied intelligence pre-model is model trained using the mechanical arm motion training set to obtain an embodied intelligence training model. The pre-model is trained with the training set to form a training model that can reflect the mechanical arm motion characteristics, thereby improving the adaptability and prediction ability of the model; the embodied intelligence training model is model cross-validated using the mechanical arm motion test set to obtain embodied intelligence model test data; the model parameters of the embodied intelligence training model are adjusted using the embodied intelligence model test data to obtain an embodied intelligence model. The training model was verified and adjusted through the test set, the model parameters were optimized, the stability and accuracy of the model were ensured, and the performance of the model in practical applications was improved.

[0018] Preferably, step S35 includes the following steps: Step S351: Performing environmental perception simulation on the main arm based on the embodied intelligent model to obtain environmental perception data of the main arm; performing industrial environmental feature recognition on the environmental perception data of the main arm to generate industrial environmental feature data; Step S352: generating main arm simulation parameters according to the industrial environment characteristic data to obtain main arm motion simulation parameters; encoding the main arm motion simulation parameters into simulation parameter instructions to generate main arm simulation parameter instructions; Step S353: controlling the motion of the slave arm through the master arm simulation parameter instruction and based on the embodied intelligent model to obtain slave arm simulation control data; performing slave arm path planning simulation according to the slave arm simulation control data to generate slave arm path simulation data; Step S354: Based on the embodied intelligent model, the task operation execution simulation is performed on the master arm simulation parameter instructions and the slave arm path simulation data to generate the robot arm simulation operation data.

[0019] The present invention simulates the environment perception of the main arm based on the embodied intelligent model to obtain the environment perception data of the main arm; identifies the industrial environment characteristics of the main arm environment perception data to generate industrial environment characteristic data. The main arm simulates the perception and characteristic identification of the industrial environment, provides accurate environmental information for the subsequent main arm motion simulation, and enhances the adaptability and accuracy of the robot arm operation; generates the main arm simulation parameters according to the industrial environment characteristic data to obtain the main arm motion simulation parameters; encodes the main arm motion simulation parameters by simulation parameter instructions to generate the main arm simulation parameter instructions. The environmental characteristic data is converted into specific motion simulation parameters and instructions, providing clear operation instructions for the precise motion control of the robot arm, improving the accuracy and efficiency of the operation; the slave arm is motion controlled by the master arm simulation parameter instructions and based on the embodied intelligent model to obtain the slave arm simulation control data; the slave arm path planning simulation is performed according to the slave arm simulation control data to generate the slave arm path simulation data. The precise motion control and path planning of the slave arm are realized, which provides detailed path information for the operation of the slave arm and improves the coordination and efficiency of the robot arm operation. Based on the embodied intelligent model, the task operation execution simulation is performed on the master arm simulation parameter instructions and the slave arm path simulation data to generate the robot arm simulation operation data. The simulation data of the master arm and the slave arm are integrated to realize the simulation of the overall operation of the robot arm, providing comprehensive simulation data for the practical application of the robot arm and enhancing the predictability and reliability of the operation.

[0020] Preferably, step S4 comprises the following steps: Step S41: extracting the simulation task operation success amount from the robot arm industrial simulation index to obtain the simulation task operation success amount; calculating the operation task success rate by combining the simulation task operation success amount with the total simulation task operation amount to generate the simulation task success rate; Step S42: recording the simulation success task time of the robot arm industrial simulation index based on the simulation task operation success amount to obtain the simulation success task time; calculating the task time efficiency of the simulation success task time and the simulation task total time to generate the simulation task time efficiency; Step S43: quantifying the simulation operation performance by using the simulation task success rate and the simulation task time efficiency, and generating a robot arm simulation performance report; Step S44: adjusting the main arm perception instruction parameters of the embodied intelligent model according to the robot arm simulation performance report to obtain the main arm perception parameter adjustment measures; Step S45: adjusting the slave arm path planning parameters of the embodied intelligent model according to the robot arm simulation performance report to obtain slave arm path planning adjustment measures; Step S46: Based on the master arm perception parameter adjustment measures and the slave arm execution path adjustment measures, the embodied intelligent model is dynamically simulated and adjusted to complete the embodied intelligent construction operation of the digital twin.

[0021] The present invention extracts the amount of successful operation of the simulation task from the industrial simulation index of the robot arm to obtain the amount of successful operation of the simulation task; calculates the success rate of the operation task by combining the successful operation amount of the simulation task with the total amount of simulation task operation, and generates the success rate of the simulation task. By counting the amount of successful operation and calculating the success rate, a quantitative evaluation of the execution effect of the simulation task is provided, and a clear indicator is provided for performance improvement; based on the amount of successful operation of the simulation task, the time of the successful simulation task is recorded for the industrial simulation index of the robot arm to obtain the time of the successful simulation task; the time efficiency of the successful simulation task time and the total time of the simulation task are calculated to generate the time efficiency of the simulation task. By recording the time required for the successful task and comparing it with the total time, the time efficiency of the task is calculated, which provides data support for evaluating and optimizing the execution speed of the simulation task; the simulation task success rate and the simulation task time efficiency are used to quantify the simulation operation performance, and a robot arm simulation performance report is generated. The two key indicators of success rate and time efficiency are combined to form a comprehensive simulation performance report, which provides a quantitative basis for the performance evaluation and optimization of the robot arm; the main arm perception instruction parameters of the embodied intelligent model are adjusted according to the robot arm simulation performance report to obtain the main arm perception parameter adjustment measures. By analyzing the simulation performance report, the main arm perception instruction parameters are adjusted to optimize the main arm's environmental perception ability and improve the response speed and accuracy of the robot arm; according to the robot arm simulation performance report, the slave arm path planning parameters of the embodied intelligent model are adjusted to obtain the slave arm path planning adjustment measures. By analyzing the simulation performance report, the slave arm path planning parameters are adjusted to optimize the movement path of the slave arm, improve the operation efficiency of the robot arm and the rationality of path planning; based on the main arm perception parameter adjustment measures and the slave arm execution path adjustment measures, the embodied intelligent model is dynamically simulated and adjusted to complete the embodied intelligent construction of the digital twin. By implementing the perception and path planning adjustment measures, the embodied intelligent model is dynamically simulated and adjusted, realizing the optimized construction of the embodied intelligent model in the digital twin environment, and improving the overall performance and intelligence level of the robot arm.

[0022] In this specification, a digital twin-based embodied intelligence system is provided, which is used to execute the above-mentioned digital twin-based embodied intelligence construction method. The digital twin-based embodied intelligence system includes: The industrial manufacturing data acquisition module is used to obtain the original industrial manufacturing data and the original images for monitoring the manufacturing process using intelligent sensors; perform data standardization processing on the original industrial manufacturing data to obtain industrial manufacturing data; perform image enhancement processing on the original images for monitoring the manufacturing process to obtain manufacturing monitoring images; The industrial manufacturing digital twin framework fitting module is used to extract manufacturing parameters from industrial manufacturing data and generate industrial manufacturing parameters; perform manufacturing scene mapping on manufacturing monitoring images based on industrial manufacturing parameters to generate industrial manufacturing scene data; perform digital twin fitting based on industrial manufacturing parameters, manufacturing monitoring images and industrial manufacturing scene data to build an industrial manufacturing digital twin framework; The embodied intelligent model construction module is used to solve the kinematics of the robot arm based on the industrial manufacturing digital twin framework to obtain the robot arm motion parameters; to construct an embodied intelligent model for the robot arm motion parameters to obtain an embodied intelligent model; to simulate the industrial manufacturing operation of the robot arm according to the embodied intelligent model to generate the robot arm simulation operation data; to detect the industrial simulation operation index of the robot arm simulation operation data using the industrial manufacturing digital twin framework to obtain the robot arm industrial simulation index; The embodied intelligent model simulation adjustment module is used to verify the simulation operation performance of the robotic arm industrial simulation indicators and generate a robotic arm simulation performance report; based on the robotic arm simulation performance report, the embodied intelligent model is dynamically simulated and adjusted to complete the embodied intelligence construction task of the digital twin.

[0023] The present invention uses an industrial manufacturing data acquisition module to obtain industrial manufacturing raw data and manufacturing process monitoring raw images obtained by intelligent sensors. After data standardization and image enhancement processing, the consistency of data and the clarity of images are ensured, and high-quality input data is provided for subsequent parameter extraction and scene mapping, thereby improving the accuracy of manufacturing process monitoring. Through the industrial manufacturing digital twin framework fitting module, parameter extraction is performed on industrial manufacturing data to generate industrial manufacturing parameters, and scene mapping is performed on manufacturing monitoring images based on these parameters to generate industrial manufacturing scene data. Digital twin fitting is performed in combination with industrial manufacturing parameters, manufacturing monitoring images and scene data to construct an industrial manufacturing digital twin framework. Digital mapping of the manufacturing process is realized, and an accurate digital model is provided for subsequent kinematic solution and simulation operation of the manipulator, which enhances the visualization and control capabilities of the manufacturing process. Through the embodied intelligent model construction module, the kinematic solution of the manipulator is performed based on the industrial manufacturing digital twin framework to obtain the motion parameters of the manipulator, and an embodied intelligent model is constructed. The industrial manufacturing operation simulation of the manipulator is performed through the embodied intelligent model to generate simulation operation data, and the industrial simulation operation index detection is performed to obtain the industrial simulation index of the manipulator. The precise simulation and performance evaluation of the robot arm operation is realized, which provides reliable simulation data and performance indicators for the practical application of the robot arm, and optimizes the operating efficiency and accuracy of the robot arm. Through the embodied intelligent model simulation adjustment module, the performance of the robot arm industrial simulation indicators is verified, a robot arm simulation performance report is generated, and the embodied intelligent model is dynamically simulated and adjusted based on this report. The accuracy and practicality of the digital twin model are ensured, and the embodied intelligent construction operation of the digital twin is completed through continuous performance verification and model adjustment, which improves the adaptability and reliability of the robot arm operation. Therefore, the present invention uses data twin technology, image processing technology and deep learning technology to perform digital twin fitting according to industrial manufacturing feature data, and performs scene simulation operations of the embodied intelligent system, thereby improving the applicability and automation of embodied intelligent operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic diagram of the steps of a method for building embodied intelligence based on digital twins; Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart; Figure 3 for Figure 1 Detailed implementation steps of step S4 in FIG. The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0025] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0026] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities are implemented in software form, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0027] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiment, the first unit is referred to as the second unit, and similarly the second unit is referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0028] To achieve this, please refer to Figures 1 to 3 , a method for constructing embodied intelligence based on digital twins, the method comprising the following steps: Step S1: using intelligent sensors to obtain industrial manufacturing raw data and manufacturing process monitoring raw images; performing data standardization processing on the industrial manufacturing raw data to obtain industrial manufacturing data; performing image enhancement processing on the manufacturing process monitoring raw images to obtain manufacturing monitoring images; Step S2: extract manufacturing parameters from industrial manufacturing data to generate industrial manufacturing parameters; perform manufacturing scene mapping on manufacturing monitoring images based on industrial manufacturing parameters to generate industrial manufacturing scene data; perform digital twin fitting based on industrial manufacturing parameters, manufacturing monitoring images and industrial manufacturing scene data to construct an industrial manufacturing digital twin framework; Step S3: Based on the industrial manufacturing digital twin framework, the kinematics of the robot arm is solved to obtain the motion parameters of the robot arm; an embodied intelligent model is constructed for the motion parameters of the robot arm to obtain the embodied intelligent model; the industrial manufacturing operation of the robot arm is simulated according to the embodied intelligent model to generate the simulation operation data of the robot arm; the industrial simulation operation index detection of the simulation operation data of the robot arm is performed using the industrial manufacturing digital twin framework to obtain the industrial simulation index of the robot arm; Step S4: Verify the simulation operation performance of the robotic arm industrial simulation indicators and generate a robotic arm simulation performance report; dynamically simulate and adjust the embodied intelligent model based on the robotic arm simulation performance report to complete the embodied intelligence construction task of the digital twin.

[0029] The present invention uses the original industrial manufacturing data and the original images of manufacturing process monitoring obtained by intelligent sensors, and after data standardization and image enhancement processing, ensures the consistency of data and the clarity of images, provides high-quality input data for subsequent parameter extraction and scene mapping, and thus improves the accuracy of manufacturing process monitoring. Parameters are extracted from industrial manufacturing data to generate industrial manufacturing parameters, and scene mapping is performed on manufacturing monitoring images based on these parameters to generate industrial manufacturing scene data. Digital twin fitting is performed in combination with industrial manufacturing parameters, manufacturing monitoring images and scene data, and an industrial manufacturing digital twin framework is constructed. Digital mapping of the manufacturing process is realized, and an accurate digital model is provided for subsequent kinematic solution and simulation operation of the robot arm, which enhances the visualization and control capabilities of the manufacturing process. Based on the industrial manufacturing digital twin framework, the kinematic solution of the robot arm is performed to obtain the motion parameters of the robot arm, and an embodied intelligent model is constructed. The industrial manufacturing operation simulation of the robot arm is performed through the embodied intelligent model, simulation operation data is generated, and industrial simulation operation index detection is performed to obtain the industrial simulation index of the robot arm. The precise simulation and performance evaluation of the robot arm operation is achieved, which provides reliable simulation data and performance indicators for the practical application of the robot arm, and optimizes the operating efficiency and accuracy of the robot arm. The performance of the robot arm industrial simulation indicators is verified, a robot arm simulation performance report is generated, and the embodied intelligent model is dynamically simulated and adjusted based on this report. The accuracy and practicality of the digital twin model are ensured. Through continuous performance verification and model adjustment, the embodied intelligent construction operation of the digital twin is completed, and the adaptability and reliability of the robot arm operation are improved. Therefore, the present invention uses data twin technology, image processing technology and deep learning technology to perform digital twin fitting according to industrial manufacturing feature data, and performs scene simulation operations of the embodied intelligent system, thereby improving the applicability and automation of embodied intelligent operations.

[0030] In the embodiment of the present invention, reference Figure 1 As shown, in this example, the method for constructing embodied intelligence based on digital twins includes the following steps: Step S1: using intelligent sensors to obtain industrial manufacturing raw data and manufacturing process monitoring raw images; performing data standardization processing on the industrial manufacturing raw data to obtain industrial manufacturing data; performing image enhancement processing on the manufacturing process monitoring raw images to obtain manufacturing monitoring images; In the embodiment of the present invention, first, the smart sensor is used to obtain the original data of industrial manufacturing and the original image of manufacturing process monitoring; the smart sensor technology, such as MEMS (micro-electromechanical system) sensor, for example, monitors the equipment status and environmental parameters in real time by deploying temperature sensors, pressure sensors and vibration sensors, etc., and obtains the original data; then the industrial manufacturing original data is subjected to data standardization. Data standardization involves converting the collected data into a unified format and unit for easy analysis and comparison. It is achieved by adopting big data analysis technology, which can process and analyze a large amount of data from smart sensors. For example, temperature data is converted from Celsius to Fahrenheit, or pressure readings of different devices are converted into a unified Pascal unit to achieve data standardization. Then, the original image of manufacturing process monitoring is subjected to image enhancement processing. Image enhancement technology, such as Gaussian filtering, smoothes the image through a Gaussian kernel function to effectively reduce the influence of random noise. For example, in the inspection of mechanical parts, Gaussian filtering is used to reduce the noise in the image and improve the image quality so as to more accurately identify and measure the size and shape of the parts. In addition, color space conversion is also an important method of image enhancement. For example, converting from RGB color space to HSV color space enhances the saturation and hue of an image, making image analysis and processing tasks more efficient. In image processing, RGB color space is used to implement image color enhancement, exposure adjustment, color filters, and other functions. In practical applications, RGB values ​​are normalized to the range of [0,1], and then a series of formulas are used to convert RGB to XYZ color space, and then XYZ is converted to Lab color space, where L represents brightness, a represents the red-green axis, and b represents the yellow-blue axis.

[0031] Step S2: extract manufacturing parameters from industrial manufacturing data to generate industrial manufacturing parameters; perform manufacturing scene mapping on manufacturing monitoring images based on industrial manufacturing parameters to generate industrial manufacturing scene data; perform digital twin fitting based on industrial manufacturing parameters, manufacturing monitoring images and industrial manufacturing scene data to construct an industrial manufacturing digital twin framework; In an embodiment of the present invention, manufacturing parameters are first extracted from industrial manufacturing data to generate industrial manufacturing parameters. Data analysis techniques, such as machine learning algorithms, are usually used to identify and extract key manufacturing parameters. By using the support vector machine (SVM) algorithm, the collected sensor data is classified and regressed to extract key parameters that affect product quality, such as temperature, pressure, and vibration frequency. Next, the manufacturing monitoring image is mapped to a manufacturing scene based on the industrial manufacturing parameters to generate industrial manufacturing scene data. In this process, image processing techniques, such as color space conversion technology, are used to convert RGB images into Lab color space to better distinguish and identify different objects and scenes in the manufacturing process. For example, by analyzing the L (brightness), a (range from green to red), and b (range from blue to yellow) channels in the Lab color space, the various components in the manufacturing scene can be more accurately identified and classified. Finally, digital twin fitting is performed based on industrial manufacturing parameters, manufacturing monitoring images, and industrial manufacturing scene data to build an industrial manufacturing digital twin framework. Digital twin technology achieves real-time monitoring and prediction of physical entities by creating virtual copies of physical entities. In the process, the Internet of Things layer technology is used to obtain equipment status data in real time, and combined with the three-dimensional model and GIS data, a twin digital framework is established that corresponds one-to-one with the real space location.

[0032] Step S3: Based on the industrial manufacturing digital twin framework, the kinematics of the robot arm is solved to obtain the motion parameters of the robot arm; an embodied intelligent model is constructed for the motion parameters of the robot arm to obtain the embodied intelligent model; the industrial manufacturing operation of the robot arm is simulated according to the embodied intelligent model to generate the simulation operation data of the robot arm; the industrial simulation operation index detection of the simulation operation data of the robot arm is performed using the industrial manufacturing digital twin framework to obtain the industrial simulation index of the robot arm; In an embodiment of the present invention, first, the kinematics of the robot arm is solved to obtain the motion parameters of the robot arm. Specifically, the forward and inverse kinematics solution technology is usually used to define the DH parameters (Denavit-Hartenberg parameters) of the robot arm to construct a mathematical model of the robot arm; for example, for a six-degree-of-freedom robot arm, DH parameter mapping information is established; using these parameters, the forward kinematic model of the robot arm is constructed, and the position and posture of the end effector are solved by matrix multiplication. Specifically, by continuous transformation matrix product, the homogeneous transformation matrix of the end effector relative to the base coordinate system is obtained, which contains the position and posture information. Then, the embodied intelligent model is constructed for the motion parameters of the robot arm to obtain the embodied intelligent model. The construction of the embodied intelligent model involves machine learning technology, especially reinforcement learning. For example, a policy network (PolicyNetwork) is used to learn the optimal behavior of the robot arm in a specific task. During the training process, the network optimizes its parameters by interacting with the environment, thereby generating a probability distribution that can guide the action of the robot arm. Then, the industrial manufacturing operation simulation of the robot arm is performed according to the embodied intelligent model to generate the simulation operation data of the robot arm. This is achieved through simulation software, such as MATLAB's Simulink toolbox, which contains a complete control system model of sensors, controllers, actuators, and the robot body. By applying different input signals, the effect of the controller on the dynamic response of the robot is observed, and simulation operation data is generated. Finally, the industrial simulation operation index detection of the robot simulation operation data is performed using the industrial manufacturing digital twin framework to obtain the industrial simulation index of the robot. This includes the evaluation of key performance indicators in the simulation process, such as overshoot, rise time, steady-state error, and response time.

[0033] Step S4: Verify the simulation operation performance of the robotic arm industrial simulation indicators and generate a robotic arm simulation performance report; dynamically simulate and adjust the embodied intelligent model based on the robotic arm simulation performance report to complete the embodied intelligence construction task of the digital twin.

[0034] In the embodiment of the present invention, the industrial simulation index of the robot arm is firstly verified by simulation operation performance, and a simulation performance report of the robot arm is generated; specifically, MATLAB Simulink is used for simulation verification, by building a dynamic model of the robot arm in Simulink, setting initial conditions, such as the initial position and speed of the joint, and then running the simulation and observing whether the behavior of the robot arm is consistent with expectations. The simulation results evaluate the accuracy of the model by comparing the simulation output with the theoretical solution or experimental data. For example, by comparing the robot end position fed back in Simulink with the theoretical calculation results, the correctness of the robot kinematic solution is verified. Then, the embodied intelligent model is dynamically simulated and adjusted based on the robot arm simulation performance report to complete the embodied intelligent construction operation of the digital twin. In, the embodied intelligent model is trained and adjusted using reinforcement learning technology. For example, the optimal behavior of the robot arm in a specific task is learned through a policy network, and the network optimizes its parameters by interacting with the environment. During the training process, simulation software such as MATLAB's Robotics Toolbox is used to simulate the movement of the robot arm, and the parameters of the policy network are adjusted according to the simulation results to improve the operational performance of the robot arm. During the dynamic simulation adjustment process, sensitivity analysis is performed on key parameters (such as mass, moment of inertia, friction coefficient, etc.) to understand the impact of these parameters on system performance. In addition, the simulation results are used to optimize the control strategy of the robot arm to obtain better dynamic performance. Through these technical means, it is ensured that the embodied intelligence model can accurately simulate and predict the actual behavior of the robot arm in the digital twin framework, thereby completing the embodied intelligence construction task of the digital twin.

[0035] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes: Step S21: extracting manufacturing parameters from industrial manufacturing data to generate industrial manufacturing parameters, wherein the industrial manufacturing parameters include material parameters, equipment parameters, and process parameters; Step S22: classify the material parameters by characteristics to obtain the material type identification; match the equipment parameters with the equipment operation mode to obtain the equipment category label; determine the production process quantity of the process parameters to obtain the process flow type; Step S23: classifying the industrial manufacturing data into industrial manufacturing types according to the material type identification, equipment category label and process flow type to obtain the industrial manufacturing type; Step S24: performing image parameterization marking on the manufacturing monitoring image based on the industrial manufacturing parameters to generate parameterized image features; performing manufacturing scene mapping on the parameterized image features according to the industrial manufacturing type to generate industrial manufacturing scene data; Step S25: Perform digital twin fitting based on industrial manufacturing parameters, manufacturing monitoring images, and industrial manufacturing scene data to build an industrial manufacturing digital twin framework.

[0036] In an embodiment of the present invention, manufacturing parameters are extracted from industrial manufacturing data, and data mining technology, in particular, decision tree algorithm, is specifically used to analyze a large amount of industrial data to extract key manufacturing parameters. For example, by constructing a decision tree model, the physical properties of raw materials, the operating status of equipment, and the parameters of the process flow are classified and extracted; the industrial manufacturing data is divided into industrial manufacturing types according to the material type identification, equipment category label, and process flow type, and the type division of industrial manufacturing data is achieved by constructing a classification model, such as a support vector machine (SVM) or a random forest algorithm. These models can output the corresponding industrial manufacturing type based on the input materials, equipment, and process parameters. The manufacturing monitoring image is parameterized and marked based on the industrial manufacturing parameters, and the manufacturing scene mapping of the parameterized image features is performed according to the industrial manufacturing type to generate industrial manufacturing scene data; computer vision technology, in particular image recognition and feature extraction technology, such as convolutional neural network (CNN), is applied to analyze and parameterize the manufacturing monitoring image. Through these technologies, key features in the image are identified and associated with industrial manufacturing parameters to generate parameterized image features. Then, using these features and industrial manufacturing type information, image processing technologies such as image segmentation and target tracking are used to map the manufacturing scene and generate industrial manufacturing scene data; digital twin fitting is performed based on industrial manufacturing parameters, manufacturing monitoring images, and industrial manufacturing scene data to build an industrial manufacturing digital twin framework; digital twin technology is specifically applied, and real-time data is collected using Internet of Things (IoT) technology, combined with 3D modeling technology, such as computer-aided design (CAD) software, to build a virtual model of physical equipment. Through the integration of real-time data and virtual models, a digital twin framework for industrial manufacturing is built.

[0037] Preferably, step S25 includes the following steps: Step S251: performing feature synchronization on industrial manufacturing parameters to obtain synchronized manufacturing parameters; performing scene matching on manufacturing monitoring images to obtain image scene matching data; Step S252: aligning manufacturing scene features with the synchronized manufacturing parameters and the image scene matching data to generate scene feature alignment data; fusing feature parameters of the industrial manufacturing scene data according to the scene feature alignment data to obtain fused scene data; Step S253: performing industrial manufacturing digital fitting on the fused scene data to obtain an industrial manufacturing digital fitting set; Step S254: constructing a digital twin framework for the industrial manufacturing digital fitting set to obtain a preliminary digital twin framework; Step S255: Perform industrial manufacturing feature mapping on the preliminary digital twin framework to obtain a mapped digital twin framework; perform industrial manufacturing digital integration on the mapped digital twin framework to generate an industrial manufacturing digital twin framework.

[0038] In an embodiment of the present invention, feature synchronization is performed on industrial manufacturing parameters to obtain synchronized manufacturing parameters; scene matching is performed on manufacturing monitoring images to obtain image scene matching data. Feature synchronization is achieved through time series analysis technology, which can align data features at different time points to ensure the temporal consistency of data; for image scene matching, feature-based visual synchronization positioning and mapping technology is adopted to achieve scene matching by establishing a corresponding relationship between features in two images; synchronized manufacturing parameters and image scene matching data are aligned for manufacturing scene features to generate scene feature alignment data; feature parameter fusion is performed on industrial manufacturing scene data according to the scene feature alignment data to obtain fused scene data. Manufacturing scene feature alignment is achieved through feature matching technology, which can establish a corresponding relationship between features between different data sources. Feature parameter fusion is achieved through data fusion technology, such as FreqFusion feature fusion calculation, which can fuse feature information from different sources to improve data consistency and availability. Industrial manufacturing digital fitting is performed on the fused scene data to obtain an industrial manufacturing digital fitting set. Industrial manufacturing digital fitting is achieved through simulation technology, which can simulate the actual industrial manufacturing process and fit the data to predict and optimize the production process; the digital twin framework is constructed for the industrial manufacturing digital fitting set to obtain a preliminary digital twin framework. The construction of the digital twin framework is achieved through three-dimensional modeling technology and real-time data integration technology, which can create accurate geometric models of physical entities and integrate real-time data into the model to achieve a high degree of integration between the physical and virtual worlds; the industrial manufacturing feature mapping is performed on the preliminary digital twin framework to obtain a mapped digital twin framework; the industrial manufacturing digital integration of the mapped digital twin framework is performed to generate an industrial manufacturing digital twin framework. Industrial manufacturing feature mapping is achieved through human-computer interaction technology and artificial intelligence and machine learning technology, which can provide visual mapping presentation capabilities and interactive operation capabilities, while analyzing data, predicting trends, and optimizing decisions. Industrial manufacturing digital integration is achieved through cloud computing and edge computing technologies, which provide powerful computing power and data storage capabilities to support large-scale digital twin applications. Through these technologies, a complete industrial manufacturing digital twin framework is built.

[0039] Preferably, step S3 comprises the following steps: Step S31: performing inverse kinematics solution on the main arm based on the industrial manufacturing digital twin framework to obtain the main arm motion parameters; Step S32: performing forward kinematics solution on the slave arm according to the motion parameters of the master arm to obtain the motion parameters of the slave arm; Step S33: integrating the master arm motion parameters and the slave arm motion parameters to generate the robot arm motion parameters; Step S34: constructing an embodied intelligent model for the robot arm motion parameters to obtain an embodied intelligent model; Step S35: simulating the industrial manufacturing operation of the robot arm according to the embodied intelligent model to generate robot arm simulation operation data; Step S36: using the industrial manufacturing digital twin framework to identify the robot arm behavior results on the robot arm simulation operation data to obtain the robot arm behavior results; Step S37: Determine the simulation task operation amount of the robot arm behavior result to generate the total simulation task operation amount; record the task time of the simulation task operation amount to obtain the total simulation task time; perform simulation operation index detection based on the total simulation task operation amount and the total simulation task time to obtain the robot arm industrial simulation index.

[0040] In an embodiment of the present invention, the inverse kinematics of the main arm is solved based on the industrial manufacturing digital twin framework to obtain the main arm motion parameters. The inverse kinematics solution usually involves the problem of converting the position and posture of the end effector into joint angles. This is achieved through numerical methods, such as the Newton-Raphson method. In the specific operation, the forward kinematics model of the robot arm is first established, and then the joint angle is inferred from the desired position and posture of the end effector by iterative approximation; the forward kinematics of the slave arm is solved according to the main arm motion parameters to obtain the slave arm motion parameters. Forward kinematics solution is the process of calculating the position and posture of the end effector of the robot arm from known joint parameters. This is usually achieved by multiplying the DH transformation matrix, where the DH parameters include joint angles, connecting rod lengths and offsets; integrating the main arm motion parameters and the slave arm motion parameters to generate the robot arm motion parameters. It involves merging the motion parameters of the main arm and the slave arm to obtain the motion parameters of the complete robot arm. It is achieved through matrix operations, that is, multiplying the transformation matrices of the master arm and the slave arm to obtain the transformation matrix of the entire robot arm; the robot arm motion parameters are embodied intelligent model construction to obtain the embodied intelligent model. The construction of the embodied intelligent model involves the use of machine learning techniques, such as deep learning, to simulate and predict the behavior of the robot arm. The motion parameters of the robot arm are learned and simulated using a neural network to build an embodied intelligent model. The industrial manufacturing operation of the robot arm is simulated according to the embodied intelligent model to generate the robot arm simulation operation data. The embodied intelligent model is used to simulate the operation of the robot arm in a virtual environment to generate simulation operation data. This is achieved through robot arm simulation software, such as the MATLAB Robot Toolbox; the robot arm simulation operation data is used to identify the robot arm behavior results using the industrial manufacturing digital twin framework to obtain the robot arm behavior results. The behavior result recognition is achieved through pattern recognition technology, such as identifying the behavior results of the robot arm by analyzing specific patterns in the simulation operation data. The simulation task operation amount of the robot arm behavior results is determined to generate the total simulation task operation amount; the task time of the simulation task operation amount is recorded to obtain the total simulation task time; the simulation operation index is detected based on the total simulation task operation amount and the total simulation task time to obtain the industrial simulation index of the robot arm. The determination of the simulation task operation amount and time is achieved through statistical analysis of the simulation operation data, and the simulation operation index detection involves the evaluation and analysis of these statistical results to determine the performance index of the robot arm. Specifically, in implementing the above-mentioned mathematical model of the master arm and the slave arm based on Robotics Toolbox or DH parameter method, the forward kinematics solution of the robot arm is completed, as well as the analytical or numerical solution of the inverse kinematics. First, it is necessary to accurately construct a mathematical model of the robot arm, which can describe in detail the spatial position and posture of each link and joint of the robot arm, providing a solid foundation for subsequent kinematic analysis.In forward kinematics, it is necessary to be able to calculate the position and attitude of the end effector through a series of coordinate transformations based on the given joint variables. This process involves continuous coordinate transformations from the base of the robot to the end effector, each of which is accurately determined by the corresponding joint and link parameters. Such a calculation process ensures that the end effector can operate accurately according to the predetermined trajectory and attitude, which is crucial for performing complex tasks such as precise positioning and assembly. In inverse kinematics, it is necessary to be able to provide an analytical or numerical solution to solve the rotation angle of each link based on the known end effector posture. The analytical solution directly solves the joint variables through algebraic operations, which is suitable for robots with relatively simple geometric structures and can quickly and accurately obtain the precise value of the joint angle. The numerical solution is solved through iterative or optimization algorithms, which is suitable for robots with more complex structures and can handle complex situations that are difficult to solve with analytical solutions. These inverse kinematics solutions enable the robot to reversely calculate the required joint angles based on the target posture of the end effector, thereby achieving accurate path planning and task execution.

[0041] Preferably, step S31 includes the following steps: Step S311: extracting manufacturing scene twin information from the industrial manufacturing digital twin framework to obtain manufacturing scene twin information; perceiving manufacturing operation features from the manufacturing scene twin information to generate industrial manufacturing operation features; Step S312: performing operation posture recognition on the industrial manufacturing operation feature to obtain operation posture data; determining the position coordinates of the operation posture data to generate posture position coordinates; Step S313: dividing the posture position coordinates into a posture space range to obtain a posture space structure range; performing posture pattern recognition on the operation posture data based on the posture space structure range to generate an operation posture pattern; Step S314: matching the active joint type of the main arm according to the operation posture mode to obtain the active joint type; measuring the joint degree of freedom of the active joint type to generate the joint degree of freedom value; determining the joint rotation angle of the joint degree of freedom value to obtain the joint rotation angle degree; Step S315: identifying the main arm connecting rod feature of the main arm according to the operation posture mode to obtain the main arm connecting rod feature; mapping the main arm connecting rod feature to the connecting rod parameter according to the joint degree of freedom value to obtain the main arm connecting rod parameter value; determining the connecting rod telescopic length of the main arm connecting rod parameter value to obtain the connecting rod telescopic length; Step S316: Integrate the joint rotation angle and the connecting rod extension length into the main arm motion parameters to obtain the main arm motion parameters.

[0042] In an embodiment of the present invention, manufacturing scene twin information is extracted from the industrial manufacturing digital twin framework to obtain manufacturing scene twin information; manufacturing operation feature perception is performed on the manufacturing scene twin information to generate industrial manufacturing operation features. Specifically, the digital twin technology is used to extract twin information in the manufacturing scene through the virtual-real interaction of physical space and digital space. Manufacturing operation feature perception is realized by machine vision technology, for example, an image processing module is used to identify the reflective tape in the workpiece image taken by the camera, and the length of the reflective tape of each workpiece is determined; the operation posture recognition is performed on the industrial manufacturing operation feature to obtain operation posture data; the position coordinates of the operation posture data are determined to generate posture position coordinates. Operation posture recognition is realized by visual posture estimation technology, such as using Mask R-CNN to perform 6D object posture estimation. Position coordinate determination is performed by using a posture estimation algorithm in computer vision technology, such as PoseCNN, to determine the position and posture of the workpiece in three-dimensional space; the posture position coordinates are divided into a posture space range to obtain a posture space structure range; posture pattern recognition is performed on the operation posture data based on the posture space structure range to generate an operation posture pattern. The spatial range division of posture is achieved by defining the boundaries and regions of the workspace, while posture pattern recognition is performed on the operation posture data through machine learning algorithms, such as deep learning networks. According to the operation posture pattern, the active joint type of the main arm is matched to obtain the active joint type; the joint degree of freedom is measured for the active joint type to generate the joint degree of freedom value; the joint angle is determined for the joint degree of freedom value to obtain the degree of joint rotation. The active joint type matching is achieved by analyzing the physical characteristics and range of motion of the joint, while the joint degree of freedom measurement is determined by the kinematic model of the robot arm. The joint angle determination is achieved by solving inverse kinematics, such as using the Newton-Raphson method; the main arm connecting rod feature is identified according to the operation posture pattern to obtain the main arm connecting rod feature; the connecting rod parameter mapping is performed on the main arm connecting rod feature according to the joint degree of freedom value to obtain the main arm connecting rod parameter value; the connecting rod telescopic length is determined for the main arm connecting rod parameter value to obtain the connecting rod telescopic length. The main arm link feature recognition is achieved by analyzing the geometric parameters of the robot arm. The link parameter mapping is defined by the DH parameters, and the link telescopic length is determined by calculating the forward kinematics model of the robot arm. The joint rotation angle and the link telescopic length are integrated into the main arm motion parameters to obtain the main arm motion parameters. This involves integrating parameters such as joint angles and link lengths into a unified robot arm motion parameter model, which is achieved through the kinematic equations of the robot arm, such as using the MATLAB Robot Toolbox for simulation motion.

[0043] Preferably, step S32 includes the following steps: Step S321: Encoding the main arm motion parameters into operation instructions to generate joint rotation angle instructions and connecting rod extension length instructions; Step S322: determining the slave arm joint rotation angle of the slave arm according to the joint rotation angle instruction, generating the slave arm joint rotation angle; performing slave arm joint rotation execution based on the slave arm joint rotation angle, and obtaining slave arm joint rotation execution data; Step S323: determining the telescopic length of the slave arm link of the slave arm according to the link telescopic length instruction, generating the telescopic length of the slave arm link; performing telescopic execution of the slave arm link based on the telescopic length of the slave arm link, and obtaining telescopic execution data of the slave arm link; Step S324: performing slave arm posture feature recognition based on the slave arm joint rotation execution data and the slave arm link extension and retraction execution data to obtain slave arm posture feature data; recording the posture coordinates of the slave arm posture feature data to generate slave arm posture coordinates; tracking the slave arm execution path according to the slave arm posture coordinates to generate slave arm execution path data; Step S325: Integrate the slave arm motion parameters using the slave arm joint rotation execution data, the slave arm link extension and retraction execution data, and the slave arm execution path data to obtain the slave arm motion parameters.

[0044] In an embodiment of the present invention, the motion parameters of the main arm are encoded with operation instructions to generate joint rotation angle instructions and connecting rod telescopic length instructions. The operation instruction encoding is realized by parsing the DH parameters of the manipulator, which include joint angle, connecting rod length, connecting rod offset and connecting rod torsion angle. For example, the SerialLink class in the MATLAB robot toolbox is used to define the manipulator model and generate corresponding operation instructions; the slave arm joint rotation angle is determined for the slave arm according to the joint rotation angle instruction to generate the slave arm joint rotation angle; the slave arm joint rotation is performed based on the slave arm joint rotation angle, and the joint rotation angle determination is realized by parsing the specific angle value obtained by parsing the operation instruction encoding. The joint rotation execution is completed by controlling the servo motor of the manipulator, thereby obtaining the joint rotation execution data. The slave arm connecting rod telescopic length is determined for the slave arm according to the connecting rod telescopic length instruction; the slave arm connecting rod telescopic execution is performed based on the slave arm connecting rod telescopic length to obtain the slave arm connecting rod telescopic execution data. The connecting rod telescopic length determination is realized by parsing the specific length value obtained by parsing the operation instruction encoding. The connecting rod extension and retraction is performed by controlling the electric cylinder or air cylinder of the robot arm to obtain the connecting rod extension and retraction execution data; the slave arm posture feature recognition is performed based on the slave arm joint rotation execution data and the slave arm connecting rod extension and retraction execution data; the posture coordinates of the slave arm posture feature data are recorded to generate the slave arm posture coordinates; the slave arm execution path tracking is performed according to the slave arm posture coordinates. Posture feature recognition determines the current posture of the robot arm by analyzing the joint rotation and connecting rod extension and retraction data. Posture coordinate recording is achieved by measuring the position and posture of the end effector of the robot arm. Execution path tracking is achieved by recording the movement trajectory of the end effector of the robot arm; the slave arm motion parameters are integrated with the slave arm joint rotation execution data, the slave arm connecting rod extension and retraction execution data and the slave arm execution path data to obtain the slave arm motion parameters.

[0045] Preferably, step S34 includes the following steps: Step S341: Calculate the motion parameters of the robot arm to obtain the motion parameters of the robot arm; Step S342: dividing the robot arm motion parameter quantity according to 80% parameter quantity to obtain a robot arm motion training set; dividing the robot arm motion parameter quantity according to 20% parameter quantity to obtain a robot arm motion test set; Step S343: constructing an embodied intelligence model according to the robot arm motion parameters to obtain an embodied intelligence pre-model; and training the embodied intelligence pre-model using the robot arm motion training set to obtain an embodied intelligence training model; Step S344: Use the robotic arm motion test set to perform a model cross-validation test on the embodied intelligence training model to obtain embodied intelligence model test data; adjust the model parameters of the embodied intelligence training model using the embodied intelligence model test data to obtain the embodied intelligence model.

[0046] In an embodiment of the present invention, the mechanical arm motion parameters are measured and calculated, which is specifically achieved through physical parameter measurement and experimental identification; by directly measuring the physical parameters of the mechanical arm, such as the length, mass, moment of inertia of the connecting rod, etc., using measuring tools such as calipers, balances, three-coordinate measuring machines, etc. for precise measurement; and summarizing and statistically analyzing the measured parameters to obtain the mechanical arm motion parameters. The mechanical arm motion parameters are divided into 80% parameters to obtain a mechanical arm motion training set; the mechanical arm motion parameters are divided into 20% parameters to obtain a mechanical arm motion test set. It is achieved by a random sampling method, for example, 80% is randomly selected from the total data set as a training set, and the remaining 20% ​​is used as a test set; an embodied intelligence model is constructed according to the mechanical arm motion parameters to obtain an embodied intelligence pre-model; the embodied intelligence pre-model is model trained using the mechanical arm motion training set to obtain an embodied intelligence training model. The embodied intelligence model is constructed using deep learning technology, such as using a convolutional neural network (CNN) or a recurrent neural network (RNN) to construct a pre-model. Model training uses supervised learning to train the model using labeled data in the training set and optimize model parameters; the robot arm motion test set is used to perform model cross-validation testing on the embodied intelligence training model to obtain embodied intelligence model test data; the embodied intelligence training model is adjusted using the embodied intelligence model test data to obtain the embodied intelligence model. The model cross-validation test is performed using the K-fold cross-validation method, which divides the test set into K subsets, and uses one of the subsets as the validation set in turn, and the rest as the training set. The performance indicators of the model on each subset are calculated to finally obtain the model test data.

[0047] Preferably, step S35 includes the following steps: Step S351: Performing environmental perception simulation on the main arm based on the embodied intelligent model to obtain environmental perception data of the main arm; performing industrial environmental feature recognition on the environmental perception data of the main arm to generate industrial environmental feature data; Step S352: generating main arm simulation parameters according to the industrial environment characteristic data to obtain main arm motion simulation parameters; encoding the main arm motion simulation parameters into simulation parameter instructions to generate main arm simulation parameter instructions; Step S353: controlling the motion of the slave arm through the master arm simulation parameter instruction and based on the embodied intelligent model to obtain slave arm simulation control data; performing slave arm path planning simulation according to the slave arm simulation control data to generate slave arm path simulation data; Step S354: Based on the embodied intelligent model, the task operation execution simulation is performed on the master arm simulation parameter instructions and the slave arm path simulation data to generate the robot arm simulation operation data.

[0048] In an embodiment of the present invention, an environmental perception simulation is performed on the main arm based on an embodied intelligent model to obtain environmental perception data of the main arm; industrial environmental feature recognition is performed on the environmental perception data of the main arm to generate industrial environmental feature data. Environmental perception simulation realizes perception of the environment around the robot arm by combining machine vision technology, such as the YOLO algorithm. Industrial environmental feature recognition uses image processing technology, such as edge detection and feature matching, to identify key objects and features in the environment; main arm simulation parameters are generated according to the industrial environmental feature data to obtain main arm motion simulation parameters; the main arm motion simulation parameters are encoded by simulation parameter instructions to generate main arm simulation parameter instructions. Simulation parameter generation is determined by analyzing environmental feature data and the kinematic model of the robot arm. Simulation parameter instruction encoding is realized by converting motion parameters into instruction sequences recognizable by the robot arm controller, for example, using the DH parameter method to define the joint angle and position of the robot arm; the slave arm is motion controlled by the main arm simulation parameter instruction and based on the embodied intelligent model to obtain slave arm simulation control data; slave arm path planning simulation is performed according to the slave arm simulation control data to generate slave arm path simulation data. Motion control uses advanced control algorithms such as PID control, fuzzy control, neural network control or adaptive control to achieve precise control of the slave arm. Path planning simulation is performed through algorithms such as the RRT algorithm to ensure that the slave arm avoids obstacles and reaches the predetermined target during movement; based on the embodied intelligent model, the task operation execution simulation is performed on the master arm simulation parameter instructions and the slave arm path simulation data to generate the robot arm simulation operation data. Task operation execution simulation is implemented through robot arm simulation software, such as the MATLAB Robot Toolbox, which provides the Simulink toolbox to easily build complex control system simulation models and generate simulation operation data. Through these data, the actual operation of the robot arm in an industrial environment is simulated to provide a reference for practical applications.

[0049] As an example of the present invention, refer to Figure 3 As shown, in this example, step S4 includes: Step S41: extracting the simulation task operation success amount from the robot arm industrial simulation index to obtain the simulation task operation success amount; calculating the operation task success rate by combining the simulation task operation success amount with the total simulation task operation amount to generate the simulation task success rate; Step S42: recording the simulation success task time of the robot arm industrial simulation index based on the simulation task operation success amount to obtain the simulation success task time; calculating the task time efficiency of the simulation success task time and the simulation task total time to generate the simulation task time efficiency; Step S43: quantifying the simulation operation performance by using the simulation task success rate and the simulation task time efficiency, and generating a robot arm simulation performance report; Step S44: adjusting the main arm perception instruction parameters of the embodied intelligent model according to the robot arm simulation performance report to obtain the main arm perception parameter adjustment measures; Step S45: adjusting the slave arm path planning parameters of the embodied intelligent model according to the robot arm simulation performance report to obtain slave arm path planning adjustment measures; Step S46: Based on the master arm perception parameter adjustment measures and the slave arm execution path adjustment measures, the embodied intelligent model is dynamically simulated and adjusted to complete the embodied intelligent construction operation of the digital twin.

[0050] In the embodiment of the present invention, the successful amount of simulation task operation is extracted from the industrial simulation index of the robot arm; the successful amount of simulation task operation and the total amount of simulation task operation are used to calculate the success rate of the operation task, and the success rate of the simulation task is generated. The successful amount of simulation task operation is extracted by recording the success or failure of each task completed by the robot arm, for example, by setting a mark of successfully completed task in the simulation software, and then counting the number of these marks; the calculation of the success rate of the operation task is achieved by the following formula: Based on the success amount of simulation task operation, the successful simulation task time of the robot arm industrial simulation index is recorded to obtain the successful simulation task time; the successful simulation task time and the total simulation task time are used to calculate the task time efficiency to generate the simulation task time efficiency. The successful simulation task time is recorded by recording the start and end time of each successful task in the simulation software. The calculation of task time efficiency is achieved by the following formula: The simulation operation performance is quantified by the success rate of the simulation task and the time efficiency of the simulation task, and a robot simulation performance report is generated. The simulation operation performance quantification is achieved by comprehensively considering the two indicators of success rate and time efficiency. For example, a comprehensive performance score is calculated by a weighted average method; the main arm perception instruction parameters of the embodied intelligent model are adjusted according to the robot simulation performance report to obtain the main arm perception parameter adjustment measures. The main arm perception instruction parameter adjustment analyzes the data in the simulation performance report to identify the perception parameters that need to be optimized, for example, adjust the sensitivity of the sensor or adjust the parameters of the perception algorithm according to the error analysis results; the slave arm path planning parameters of the embodied intelligent model are adjusted according to the robot simulation performance report to obtain the slave arm path planning adjustment measures. The slave arm path planning parameter adjustment analyzes the data in the simulation performance report to identify the bottlenecks or deficiencies in the path planning, for example, adjust the parameters of the path planning algorithm according to the statistical data of the path length and planning time, such as adjusting the sampling density in the RRT algorithm and adjusting the heuristic function in the algorithm; the embodied intelligent model is dynamically simulated and adjusted based on the main arm perception parameter adjustment measures and the slave arm execution path adjustment measures to complete the embodied intelligence construction task of the digital twin. Dynamic simulation adjustments are performed by applying these parameter adjustments in real time in a simulation environment, for example, by dynamically adjusting the parameters of a robot model in the MATLAB Robotics Toolbox and observing the impact of these adjustments on the robot's performance.

[0051] In this specification, a digital twin-based embodied intelligence system is provided, which is used to execute the above-mentioned digital twin-based embodied intelligence construction method. The digital twin-based embodied intelligence system includes: The industrial manufacturing data acquisition module is used to obtain the original industrial manufacturing data and the original images for monitoring the manufacturing process using intelligent sensors; perform data standardization processing on the original industrial manufacturing data to obtain industrial manufacturing data; perform image enhancement processing on the original images for monitoring the manufacturing process to obtain manufacturing monitoring images; The industrial manufacturing digital twin framework fitting module is used to extract manufacturing parameters from industrial manufacturing data and generate industrial manufacturing parameters; perform manufacturing scene mapping on manufacturing monitoring images based on industrial manufacturing parameters to generate industrial manufacturing scene data; perform digital twin fitting based on industrial manufacturing parameters, manufacturing monitoring images and industrial manufacturing scene data to build an industrial manufacturing digital twin framework; The embodied intelligent model construction module is used to solve the kinematics of the robot arm based on the industrial manufacturing digital twin framework to obtain the robot arm motion parameters; to construct an embodied intelligent model for the robot arm motion parameters to obtain an embodied intelligent model; to simulate the industrial manufacturing operation of the robot arm according to the embodied intelligent model to generate the robot arm simulation operation data; to detect the industrial simulation operation index of the robot arm simulation operation data using the industrial manufacturing digital twin framework to obtain the robot arm industrial simulation index; The embodied intelligent model simulation adjustment module is used to verify the simulation operation performance of the robotic arm industrial simulation indicators and generate a robotic arm simulation performance report; based on the robotic arm simulation performance report, the embodied intelligent model is dynamically simulated and adjusted to complete the embodied intelligence construction task of the digital twin.

[0052] The present invention uses an industrial manufacturing data acquisition module to obtain industrial manufacturing raw data and manufacturing process monitoring raw images obtained by intelligent sensors. After data standardization and image enhancement processing, the consistency of data and the clarity of images are ensured, and high-quality input data is provided for subsequent parameter extraction and scene mapping, thereby improving the accuracy of manufacturing process monitoring. Through the industrial manufacturing digital twin framework fitting module, parameter extraction is performed on industrial manufacturing data to generate industrial manufacturing parameters, and scene mapping is performed on manufacturing monitoring images based on these parameters to generate industrial manufacturing scene data. Digital twin fitting is performed in combination with industrial manufacturing parameters, manufacturing monitoring images and scene data to construct an industrial manufacturing digital twin framework. Digital mapping of the manufacturing process is realized, and an accurate digital model is provided for subsequent kinematic solution and simulation operation of the manipulator, which enhances the visualization and control capabilities of the manufacturing process. Through the embodied intelligent model construction module, the kinematic solution of the manipulator is performed based on the industrial manufacturing digital twin framework to obtain the motion parameters of the manipulator, and an embodied intelligent model is constructed. The industrial manufacturing operation simulation of the manipulator is performed through the embodied intelligent model to generate simulation operation data, and the industrial simulation operation index detection is performed to obtain the industrial simulation index of the manipulator. The precise simulation and performance evaluation of the robot arm operation is realized, which provides reliable simulation data and performance indicators for the practical application of the robot arm, and optimizes the operating efficiency and accuracy of the robot arm. Through the embodied intelligent model simulation adjustment module, the performance of the robot arm industrial simulation indicators is verified, a robot arm simulation performance report is generated, and the embodied intelligent model is dynamically simulated and adjusted based on this report. The accuracy and practicality of the digital twin model are ensured, and the embodied intelligent construction operation of the digital twin is completed through continuous performance verification and model adjustment, which improves the adaptability and reliability of the robot arm operation. Therefore, the present invention uses data twin technology, image processing technology and deep learning technology to perform digital twin fitting according to industrial manufacturing feature data, and performs scene simulation operations of the embodied intelligent system, thereby improving the applicability and automation of embodied intelligent operations.

[0053] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0054] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for constructing embodied intelligence based on digital twins, applied to an embodied intelligent device, wherein the embodied intelligent device includes a robotic arm and an intelligent sensor, characterized in that: The following steps are involved: Step S1: using intelligent sensors to obtain industrial manufacturing raw data and manufacturing process monitoring raw images; performing data standardization processing on the industrial manufacturing raw data to obtain industrial manufacturing data; Performing image enhancement processing on the original image for monitoring the manufacturing process to obtain a manufacturing monitoring image; Step S2: extract manufacturing parameters from industrial manufacturing data to generate industrial manufacturing parameters; perform manufacturing scene mapping on manufacturing monitoring images based on industrial manufacturing parameters to generate industrial manufacturing scene data; perform digital twin fitting based on industrial manufacturing parameters, manufacturing monitoring images and industrial manufacturing scene data to construct an industrial manufacturing digital twin framework; Step S3: Based on the industrial manufacturing digital twin framework, the kinematics of the robot arm is solved to obtain the motion parameters of the robot arm; an embodied intelligent model is constructed for the motion parameters of the robot arm to obtain the embodied intelligent model; the industrial manufacturing operation of the robot arm is simulated according to the embodied intelligent model to generate the simulation operation data of the robot arm; the industrial simulation operation index detection of the simulation operation data of the robot arm is performed using the industrial manufacturing digital twin framework to obtain the industrial simulation index of the robot arm; Step S4: Verify the simulation operation performance of the robotic arm industrial simulation indicators and generate a robotic arm simulation performance report; dynamically simulate and adjust the embodied intelligent model based on the robotic arm simulation performance report to complete the embodied intelligence construction task of the digital twin.

2. The method for constructing embodied intelligence based on digital twins according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: extracting manufacturing parameters from industrial manufacturing data to generate industrial manufacturing parameters, wherein the industrial manufacturing parameters include material parameters, equipment parameters, and process parameters; Step S22: classify the material parameters by characteristics to obtain the material type identification; match the equipment parameters with the equipment operation mode to obtain the equipment category label; determine the production process quantity of the process parameters to obtain the process flow type; Step S23: classifying the industrial manufacturing data into industrial manufacturing types according to the material type identification, equipment category label and process flow type to obtain the industrial manufacturing type; Step S24: performing image parameterization marking on the manufacturing monitoring image based on the industrial manufacturing parameters to generate parameterized image features; performing manufacturing scene mapping on the parameterized image features according to the industrial manufacturing type to generate industrial manufacturing scene data; Step S25: Perform digital twin fitting based on industrial manufacturing parameters, manufacturing monitoring images, and industrial manufacturing scene data to build an industrial manufacturing digital twin framework.

3. The method for constructing embodied intelligence based on digital twins according to claim 2, characterized in that: Step S25 includes the following steps: Step S251: performing feature synchronization on industrial manufacturing parameters to obtain synchronized manufacturing parameters; performing scene matching on manufacturing monitoring images to obtain image scene matching data; Step S252: aligning manufacturing scene features with the synchronized manufacturing parameters and the image scene matching data to generate scene feature alignment data; fusing feature parameters of the industrial manufacturing scene data according to the scene feature alignment data to obtain fused scene data; Step S253: performing industrial manufacturing digital fitting on the fused scene data to obtain an industrial manufacturing digital fitting set; Step S254: constructing a digital twin framework for the industrial manufacturing digital fitting set to obtain a preliminary digital twin framework; Step S255: Perform industrial manufacturing feature mapping on the preliminary digital twin framework to obtain a mapped digital twin framework; perform industrial manufacturing digital integration on the mapped digital twin framework to generate an industrial manufacturing digital twin framework.

4. The method for constructing embodied intelligence based on digital twins according to claim 1, characterized in that: The mechanical arm includes a master arm and a slave arm, and step S3 includes the following steps: Step S31: performing inverse kinematics solution on the main arm based on the industrial manufacturing digital twin framework to obtain the main arm motion parameters; Step S32: performing forward kinematics solution on the slave arm according to the motion parameters of the master arm to obtain the motion parameters of the slave arm; Step S33: integrating the master arm motion parameters and the slave arm motion parameters to generate the robot arm motion parameters; Step S34: constructing an embodied intelligent model for the robot arm motion parameters to obtain an embodied intelligent model; Step S35: simulating the industrial manufacturing operation of the robot arm according to the embodied intelligent model to generate simulation operation data of the robot arm; Step S36: using the industrial manufacturing digital twin framework to identify the robot arm behavior results on the robot arm simulation operation data to obtain the robot arm behavior results; Step S37: Determine the simulation task operation amount of the robot arm behavior result to generate the total simulation task operation amount; record the task time of the simulation task operation amount to obtain the total simulation task time; perform simulation operation index detection based on the total simulation task operation amount and the total simulation task time to obtain the robot arm industrial simulation index.

5. The method for constructing embodied intelligence based on digital twins according to claim 4, characterized in that: Step S31 includes the following steps: Step S311: extracting manufacturing scene twin information from the industrial manufacturing digital twin framework to obtain manufacturing scene twin information; perceiving manufacturing operation features from the manufacturing scene twin information to generate industrial manufacturing operation features; Step S312: performing operation posture recognition on the industrial manufacturing operation feature to obtain operation posture data; determining the position coordinates of the operation posture data to generate posture position coordinates; Step S313: dividing the posture position coordinates into a posture space range to obtain a posture space structure range; performing posture pattern recognition on the operation posture data based on the posture space structure range to generate an operation posture pattern; Step S314: matching the active joint type of the main arm according to the operation posture mode to obtain the active joint type; measuring the joint degree of freedom of the active joint type to generate the joint degree of freedom value; determining the joint rotation angle of the joint degree of freedom value to obtain the joint rotation angle degree; Step S315: identifying the main arm connecting rod feature of the main arm according to the operation posture mode to obtain the main arm connecting rod feature; mapping the main arm connecting rod feature to the connecting rod parameter according to the joint degree of freedom value to obtain the main arm connecting rod parameter value; determining the connecting rod telescopic length of the main arm connecting rod parameter value to obtain the connecting rod telescopic length; Step S316: Integrate the joint rotation angle and the connecting rod extension length into the main arm motion parameters to obtain the main arm motion parameters.

6. The method for constructing embodied intelligence based on digital twins according to claim 5, characterized in that: Step S32 includes the following steps: Step S321: Encoding the main arm motion parameters into operation instructions to generate joint rotation angle instructions and connecting rod extension length instructions; Step S322: determining the slave arm joint rotation angle of the slave arm according to the joint rotation angle instruction, generating the slave arm joint rotation angle; performing slave arm joint rotation execution based on the slave arm joint rotation angle, and obtaining slave arm joint rotation execution data; Step S323: determining the telescopic length of the slave arm link of the slave arm according to the link telescopic length instruction, generating the telescopic length of the slave arm link; performing telescopic execution of the slave arm link based on the telescopic length of the slave arm link, and obtaining telescopic execution data of the slave arm link; Step S324: performing slave arm posture feature recognition based on the slave arm joint rotation execution data and the slave arm link extension and retraction execution data to obtain slave arm posture feature data; recording the posture coordinates of the slave arm posture feature data to generate slave arm posture coordinates; tracking the slave arm execution path according to the slave arm posture coordinates to generate slave arm execution path data; Step S325: Integrate the slave arm motion parameters using the slave arm joint rotation execution data, the slave arm link extension and retraction execution data, and the slave arm execution path data to obtain the slave arm motion parameters.

7. The method for constructing embodied intelligence based on digital twins according to claim 4, characterized in that: Step S34 includes the following steps: Step S341: Calculate the motion parameters of the robot arm to obtain the motion parameters of the robot arm; Step S342: dividing the robot arm motion parameter quantity according to 80% parameter quantity to obtain a robot arm motion training set; dividing the robot arm motion parameter quantity according to 20% parameter quantity to obtain a robot arm motion test set; Step S343: constructing an embodied intelligence model according to the robot arm motion parameters to obtain an embodied intelligence pre-model; and training the embodied intelligence pre-model using the robot arm motion training set to obtain an embodied intelligence training model; Step S344: Use the robotic arm motion test set to perform a model cross-validation test on the embodied intelligence training model to obtain embodied intelligence model test data; adjust the model parameters of the embodied intelligence training model using the embodied intelligence model test data to obtain the embodied intelligence model.

8. The method for constructing embodied intelligence based on digital twins according to claim 4, characterized in that: Step S35 includes the following steps: Step S351: Performing environmental perception simulation on the main arm based on the embodied intelligent model to obtain environmental perception data of the main arm; performing industrial environmental feature recognition on the environmental perception data of the main arm to generate industrial environmental feature data; Step S352: generating main arm simulation parameters according to the industrial environment characteristic data to obtain main arm motion simulation parameters; encoding the main arm motion simulation parameters into simulation parameter instructions to generate main arm simulation parameter instructions; Step S353: controlling the motion of the slave arm through the master arm simulation parameter instruction and based on the embodied intelligent model to obtain slave arm simulation control data; performing slave arm path planning simulation according to the slave arm simulation control data to generate slave arm path simulation data; Step S354: Based on the embodied intelligent model, the task operation execution simulation is performed on the master arm simulation parameter instructions and the slave arm path simulation data to generate the robot arm simulation operation data.

9. The method for constructing embodied intelligence based on digital twins according to claim 4, characterized in that: Step S4 includes the following steps: Step S41: extracting the simulation task operation success amount from the robot arm industrial simulation index to obtain the simulation task operation success amount; calculating the operation task success rate by combining the simulation task operation success amount with the total simulation task operation amount to generate the simulation task success rate; Step S42: recording the simulation success task time of the robot arm industrial simulation index based on the simulation task operation success amount to obtain the simulation success task time; calculating the task time efficiency of the simulation success task time and the simulation task total time to generate the simulation task time efficiency; Step S43: quantifying the simulation operation performance by using the simulation task success rate and the simulation task time efficiency, and generating a robot arm simulation performance report; Step S44: adjusting the main arm perception instruction parameters of the embodied intelligent model according to the robot arm simulation performance report to obtain the main arm perception parameter adjustment measures; Step S45: adjusting the slave arm path planning parameters of the embodied intelligent model according to the robot arm simulation performance report to obtain slave arm path planning adjustment measures; Step S46: Based on the master arm perception parameter adjustment measures and the slave arm execution path adjustment measures, the embodied intelligent model is dynamically simulated and adjusted to complete the embodied intelligent construction operation of the digital twin.

10. An embodied intelligent system based on digital twins, characterized in that: For executing the embodied intelligence construction method based on digital twins as claimed in claim 1, the embodied intelligence system based on digital twins comprises: The industrial manufacturing data acquisition module is used to obtain the original industrial manufacturing data and the original images for monitoring the manufacturing process using intelligent sensors; perform data standardization processing on the original industrial manufacturing data to obtain industrial manufacturing data; perform image enhancement processing on the original images for monitoring the manufacturing process to obtain manufacturing monitoring images; The industrial manufacturing digital twin framework fitting module is used to extract manufacturing parameters from industrial manufacturing data and generate industrial manufacturing parameters; perform manufacturing scene mapping on manufacturing monitoring images based on industrial manufacturing parameters to generate industrial manufacturing scene data; perform digital twin fitting based on industrial manufacturing parameters, manufacturing monitoring images and industrial manufacturing scene data to build an industrial manufacturing digital twin framework; The embodied intelligent model construction module is used to solve the kinematics of the robot arm based on the industrial manufacturing digital twin framework to obtain the robot arm motion parameters; to construct an embodied intelligent model for the robot arm motion parameters to obtain an embodied intelligent model; to simulate the industrial manufacturing operation of the robot arm according to the embodied intelligent model to generate the robot arm simulation operation data; to detect the industrial simulation operation index of the robot arm simulation operation data using the industrial manufacturing digital twin framework to obtain the robot arm industrial simulation index; The embodied intelligent model simulation adjustment module is used to verify the simulation operation performance of the robotic arm industrial simulation indicators and generate a robotic arm simulation performance report; based on the robotic arm simulation performance report, the embodied intelligent model is dynamically simulated and adjusted to complete the embodied intelligence construction task of the digital twin.

Images