A data cockpit system based on virtual-real fusion

Abstract

The present invention relates to the field of visualization display technology, and discloses a data cockpit system based on virtual-real fusion, including an information acquisition module, a model construction module, a data warehouse, an indicator setting module, a data request module, a deep learning module, a simulation analysis module and a visualization module. The digital twin technology is used to measure, simulate and analyze the real scene, perceive, diagnose and predict the state of the entity object in real time, and set up indicator parameters of different dimensions; a mathematical model between the entity object parameters and the monitoring perception data is constructed to realize the information interaction between the entity object and the virtual simulation model, fit the polynomial expression function, and obtain a digital twin data cockpit synchronized with reality. The present invention solves the problems that the existing data cockpit has poor visual effects, the information on the large screen cannot be intuitively understood, the data lacks visualization and intuition, is not closely connected with reality, and the accuracy of each indicator in the data cockpit needs to be improved.

Description

Technical Field

[0001] The present invention relates to the field of visualization display technology, and in particular to a data cockpit system based on virtual-reality fusion. Background Art

[0002] With the continuous development of technology, the amount of information generated within various organizations is increasing. Ordinary internal systems are increasingly unable to meet the needs of their own business, and it is becoming increasingly difficult to obtain relevant information dynamically and intuitively. The visual data cockpit can visualize, concretize, and intuitively display the data of various organizations. By using common charts to display the data information of interest, you can clearly understand the current changes in various information, thereby providing support for decision-making.

[0003] However, the existing visual data cockpit has a lot of information to display and the data analysis results are complex and difficult to understand, which makes it difficult for users to obtain effective information intuitively through the screen. It is time-consuming and laborious to find the required information in a large number of reports on the complicated screen, and the understanding of the reports is not high.

[0004] my country's patent application number: CN202210203821.X, discloses an operating system-type big data cockpit system, including: a server; the server has an AI data analysis system, a visualization display system and multiple pre-processors; when data needs to be displayed, it has two modes: the first mode, based on the retrieval control command to obtain the first display data range contained in the retrieval control command, select the corresponding database sub-table according to the first display data range; filter the corresponding data card according to the database sub-table, and display the data card visually on the display; the second mode, select the database sub-table to be displayed, drag the corresponding data card under the database sub-table to the pre-display unit, the controller controls the pre-display unit to form the data card into a pre-display style, and transmits the pre-display style to the display for visual display after selecting the pre-display style for confirmation, wherein the drag recorder is used to record the dragged data card.

[0005] However, in the process of implementing the technical solutions of the above-mentioned application embodiments, it was found that the above-mentioned technology has at least the following technical problems: the visual effect of the data cockpit is not good, the information on the large screen cannot be intuitively understood, there is a lack of data visualization and intuition, there is no close connection with reality, and the accuracy of various indicators in the data cockpit needs to be improved. Summary of the invention

[0006] The present invention provides a data cockpit system based on virtual-reality fusion, which solves the problems of poor visual effects of existing data cockpits, inability to intuitively understand the information on the large screen, lack of data visualization and intuition, lack of close connection with reality, and the accuracy of various indicators in the data cockpit needs to be improved.

[0007] The technical solution of the present invention is as follows:

[0008] A data cockpit system based on virtual-reality fusion includes the following parts:

[0009] Information acquisition module, model building module, data warehouse, indicator setting module, data request module, deep learning module, simulation analysis module and visualization module;

[0010] The model construction module is used to build a digital twin model through an interactive window editing mode or a structured data instantiation mode, and establish a mapping relationship between model elements and physical object elements. The model construction module sends the constructed digital twin model to the visualization module through data transmission;

[0011] The indicator setting module is used to set indicator parameters of different dimensions that need to be analyzed by data visualization. The indicator setting module sends the data required for the indicator to the data request module by means of data transmission;

[0012] The deep learning module is used to build a deep fitting neural network model based on training and learning of a multi-layer neural network, obtain a mathematical model between physical object parameters and monitoring perception data, realize instant information interaction between physical objects and virtual simulation models, and fit a polynomial expression function. The deep learning module sends the calculation result of the fitting function to the model building module via data transmission.

[0013] A method for implementing a data cockpit system based on virtual-reality fusion comprises the following steps:

[0014] S1. Build a digital twin model, use digital twin technology to measure, simulate, and analyze real scenes, perceive, diagnose, and predict the state of physical objects in real time, and set up indicator parameters of different dimensions that require data visualization analysis;

[0015] S2. Based on the training and learning of multi-layer neural networks, a mathematical model between the parameters of physical objects and monitoring perception data is constructed to achieve instant information interaction between physical objects and virtual simulation models, and a polynomial expression function is fitted to obtain a digital twin data cockpit synchronized with reality, and the indicator analysis results are displayed.

[0016] Further, the step S1 specifically includes:

[0017] Digital twin technology is used to measure, simulate, and analyze real scenes to sense, diagnose, and predict the status of physical objects in real time.

[0018] Further, the step S1 specifically includes:

[0019] The raw data is obtained, preprocessed, and the preprocessed data is edited to form data sets with different attributes, where the attributes include classification fields of the data; the data sets are structured according to the classification fields and stored in the data warehouse of the data cockpit system.

[0020] Further, the step S2 specifically includes:

[0021] Based on the training and learning of multi-layer neural networks, a mathematical model between the parameters of physical objects and monitoring perception data is constructed to achieve instant information interaction between physical objects and virtual simulation models, and a polynomial expression function is fitted.

[0022] Further, the step S2 specifically includes:

[0023] A deep fitting neural network model is constructed, and N groups of M-dimensional monitoring perception data and their corresponding entity object parameters are selected as training samples. The N groups of M-dimensional detection perception data are input into the deep fitting neural network model in sequence. The model output is obtained through neural network learning and training, and error analysis is performed on the entity object parameters corresponding to the input data. According to the error analysis, the parameters in the neural network model are corrected so that the accuracy of the deep fitting neural network reaches the preset target.

[0024] Further, the step S2 specifically includes:

[0025] The deep fitting neural network includes input layer, feature division layer, filtering layer, fitting layer, self-learning layer and output layer.

[0026] Further, the step S2 specifically includes:

[0027] The input layer uniformly assigns values ​​to the training samples to facilitate subsequent calculations. The feature partitioning layer divides all data into multiple blocks through a sliding window. The filtering layer selects the data in each window that meets the preset threshold as the data feature. The fitting layer performs difference fitting based on the relationship between the previous and subsequent data. The self-learning layer performs multiple self-verification and correction on the fitting results. The output layer outputs the final polynomial expression function.

[0028] The present invention has at least the following technical effects or advantages:

[0029] 1. Use digital twin technology to measure, simulate, and analyze data of real scenes, project the physical world with digital world twins, connect virtual and real, interpret the target scene of virtual and real integration, reflect the operating status of the project in the target scene in real time, establish indicator parameters of different dimensions that require data visualization analysis, and solve the problem that users have difficulty understanding the report information on the large screen in a targeted manner, making the data cockpit easier to understand and improving the visual effect.

[0030] 2. Based on the training and learning of multi-layer neural networks, a mathematical model between the parameters of physical objects and monitoring perception data is constructed to realize instant information interaction between physical objects and virtual simulation models, so as to establish a mirror connection between physical objects and virtual simulation models, realize the integration of virtual and real, and ensure the accuracy of various indicators in the data cockpit; integrating physical information into the neural network can better capture data fluctuations, and at the same time, based on the simulation results, the neural network self-learning evolution is carried out, the neural network model is dynamically updated, and the fitting accuracy is improved. The use of the data cockpit system will make decisions based on data more precise and convenient.

[0031] 3. The technical solution of the present invention can effectively solve the problems that the existing data cockpit has poor visual effects, the information on the large screen cannot be intuitively understood, there is a lack of data visualization and intuition, there is no close connection with reality, and the accuracy of various indicators in the data cockpit needs to be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a structural diagram of a data cockpit system based on virtual-real fusion according to the present invention;

[0033] Figure 2 The present invention is a flowchart of a method for implementing a data cockpit system based on virtual-reality fusion. DETAILED DESCRIPTION

[0034] The embodiment of the present application provides a data cockpit system based on the fusion of virtual and real, which solves the problems of poor visual effects of the existing data cockpit, the inability to intuitively understand the information on the large screen, the lack of data visualization and intuition, the lack of close connection with reality, and the need to improve the accuracy of various indicators in the data cockpit.

[0035] The technical solution in the embodiment of the present application is to solve the above problems, and the overall idea is as follows:

[0036] Digital twin technology is used to measure, simulate, and analyze real scenes, project the physical world with digital world twins, connect virtual and reality, interpret the target scene of virtual-real integration, reflect the operating status of the project in the target scene in real time, establish indicator parameters of different dimensions that require data visualization analysis, and solve the problem that users find it difficult to understand the report information on the large screen, making the data cockpit easier to understand and improving the visual effect; based on the training and learning of multi-layer neural networks, a mathematical model between the parameters of physical objects and monitoring perception data is constructed to achieve instant information interaction between physical objects and virtual simulation models, so as to establish a mirror connection between physical objects and virtual simulation models, achieve virtual-real integration, and ensure the accuracy of various indicators in the data cockpit; integrating physical information into neural networks can better capture data fluctuations, and at the same time, carry out self-learning evolution of neural networks based on simulation results, dynamically update neural network models, and improve fitting accuracy. Using the data cockpit system will make decisions that rely on data more precise and convenient.

[0037] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0038] Refer to the attached Figure 1 The data cockpit system based on virtual-real fusion described in the present invention includes the following parts:

[0039] Information acquisition module 10, model building module 20, data warehouse 30, indicator setting module 40, data request module 50, deep learning module 60, simulation analysis module 70 and visualization module 80.

[0040] The information acquisition module 10 is used to obtain the original data required for building the digital twin model and preprocess the original data. The information acquisition module 10 sends the preprocessed data to the model building module 20 through data transmission and stores it in the data warehouse 30;

[0041] The model building module 20 is used to build a digital twin model through an interactive window editing mode or a structured data instantiation mode, and establish a mapping relationship between model elements and physical object elements. The model building module 20 sends the constructed digital twin model to the visualization module 80 through data transmission;

[0042] The data warehouse 30 is used to store the preprocessed raw data. The data warehouse 30 sends the indicator parameters to the data request module 50 through data transmission, and sends the training samples to the deep learning module 60;

[0043] The indicator setting module 40 is used to set indicator parameters of different dimensions that need to be analyzed by data visualization. The indicator setting module 40 sends the data required for the indicator to the data request module 50 by means of data transmission;

[0044] The data request module 50 is used to obtain data supporting data visualization analysis from the data warehouse 30 according to the indicator parameters. The data request module 50 sends the data request instruction to the data warehouse 30 by means of data transmission, and sends the obtained indicator parameters to the simulation analysis module 70 by means of data transmission;

[0045] The deep learning module 60 is used to build a deep fitting neural network model based on the training and learning of the multi-layer neural network, obtain a mathematical model between the physical object parameters and the monitoring perception data, realize the instant information interaction between the physical object and the virtual simulation model, and fit a polynomial expression function. The deep learning module 60 sends the calculation result of the fitting function to the model building module 20 by means of data transmission;

[0046] The simulation analysis module 70 is used to obtain the analysis result of each indicator according to the calculation process of each indicator in the indicator parameters, and send the indicator analysis result to the visualization module 80 by means of data transmission;

[0047] The visualization module 80 is used to display the digital twin model and indicator analysis results on the data cockpit large screen.

[0048] The method for implementing a data cockpit system based on virtual-reality fusion described in the present invention comprises the following steps:

[0049] S1. Build a digital twin model, use digital twin technology to measure, simulate, and analyze real-world scenarios, perceive, diagnose, and predict the status of physical objects in real time, and establish indicator parameters of different dimensions that require data visualization analysis.

[0050] The data cockpit system can visualize, intuit and concretely display the collected data through a detailed indicator system, project the physical world with the digital world twin, connect virtual and reality, interpret the destination scene of virtual and real integration, reflect the operation status of the project in the destination scene in real time, and provide support for relevant decision-making of the project. The data cockpit system needs to comprehensively analyze the original data, select the specific data to be presented according to different project modules, and uniformly display and manage the data analysis results.

[0051] In order to achieve virtual-real integration, digital twin technology is used to measure, simulate, and analyze data in real scenes, so as to perceive, diagnose, and predict the state of physical objects in real time, help data cockpit users make decisions, and optimize and adjust physical objects through decision-making. First, the information acquisition module 10 obtains the original data required for building the digital twin model, and preprocesses the original data. The preprocessing includes: noise processing, missing processing, isolated point processing, and structured processing. The preprocessing method adopts the existing technology. For example, noise processing can use smoothing filtering or median filtering to filter out useless noise signals received by the sensor during the acquisition process; missing processing requires distinguishing null values ​​in the transmitted data and filling in default values; isolated point processing can use clustering algorithms to distinguish outlier data and remove isolated point data; structured processing can define the structure protocol according to the digital twin, and load the data into the structure according to the protocol rules. The information acquisition module 10 edits the preprocessed data into data sets with different attributes, and the attributes include the classification fields of the data; the data sets are structured according to the classification fields and stored in the data warehouse 30 of the data cockpit system.

[0052] Furthermore, in the simulation environment, the model building module 20 builds simulation models of static scenes and various states of dynamic objects, builds a digital twin model through an interactive window editing mode or a structured data instantiation mode, and establishes a mapping relationship between model elements and physical object elements for simulation analysis and data visualization.

[0053] After the digital twin model is constructed, the user sets up indicator parameters of different dimensions required for data visualization analysis in the indicator setting module 40. The indicator parameters include the indicator name, function and required original data. The data request module 50 obtains data supporting data visualization analysis from the data warehouse 30 based on the indicator parameters.

[0054] As a specific embodiment, the data cockpit can set at least one indicator, and each indicator parameter can be obtained from the data stored in the data warehouse 30; for example: a user sets up a user information dimension indicator, and the indicator parameter of the user information dimension indicator refers to: the indicator name is the user information dimension indicator, and the function of the indicator is to set up a user label based on the information obtained from the user, and the original data required for the indicator includes user name, age, gender, etc.

[0055] The beneficial effects of step S1 are: using digital twin technology to measure, simulate, and analyze data on real scenes, projecting the physical world with digital world twins, connecting virtual and real, and interpreting the target scene of virtual-real integration, reflecting the operating status of the project in the target scene in real time, and establishing indicator parameters of different dimensions that require data visualization analysis, so as to solve the problem that users have difficulty understanding the report information on the large screen in a targeted manner, making the data cockpit easier to understand and improving the visual effect.

[0056] S2. Based on the training and learning of multi-layer neural networks, a mathematical model between the parameters of physical objects and monitoring perception data is constructed to achieve instant information interaction between physical objects and virtual simulation models, and a polynomial expression function is fitted to obtain a digital twin data cockpit synchronized with reality, and the indicator analysis results are displayed.

[0057] To achieve real-time updating of the digital twin model, the key is to determine the correspondence between the monitoring perception data and the physical objects, so as to establish a mirror connection between the physical objects and the virtual simulation model, realize the integration of virtual and real, and ensure the accuracy of each indicator in the data cockpit.

[0058] Specifically, the deep learning module 60 constructs a mathematical model between physical object parameters and monitoring perception data based on the training and learning of multi-layer neural networks, realizes instant information interaction between physical objects and virtual simulation models, and fits a polynomial expression function.

[0059] As a specific embodiment, a deep fitting neural network model is constructed, N groups of M-dimensional monitoring perception data and their corresponding entity object parameters are selected as training samples, and the N groups of M-dimensional detection perception data are sequentially input into the deep fitting neural network model. The model output is obtained through neural network learning and training, and an error analysis is performed on the entity object parameters corresponding to the input data. According to the error analysis, the parameters in the neural network model are corrected so that the accuracy of the deep fitting neural network reaches the preset target.

[0060] The deep fitting neural network includes input layer, feature division layer, filtering layer, fitting layer, self-learning layer and output layer.

[0061] The deep learning module 60 selects N groups of detection perception data from the training samples to form N×M vectors D={D 1 , D 2 , ..., D N}, for any set of M-dimensional vectors

[0062] For any N vectors of one dimension

[0063] The N×M vectors D are sent to the deep fitting neural network model for training. The specific processing process of the deep fitting neural network model is as follows:

[0064] The N×M vectors D are sent to the input layer, which uniformly assigns values ​​to the training samples to facilitate subsequent calculations. The calculation formula of the input layer is:

[0065]

[0066] Among them, D j represents the dataset of the jth dimension after assignment, Represents the i+1th data under the jth dimension; the input layer transmits the processed data to the feature partitioning layer;

[0067] The feature partitioning layer divides all data into multiple blocks through a sliding window. The calculation formula of the feature partitioning layer is:

[0068]

[0069]

[0070] Among them, f 0 represents the input of the feature extraction layer, ω 12 represents the connection weight between the input layer and the feature extraction layer, b 2 represents the bias of the feature extraction layer, f 1 represents the output of the feature extraction layer, l min represents the minimum sliding window step size, l max Indicates the maximum sliding window step size; local fluctuation information is captured through the hyperbolic tangent function, and data features of different precisions are obtained by adjusting the window step size. The feature division layer transmits the divided data to the filtering layer;

[0071] The filtering layer selects the data that meets the preset threshold in each window as data features and transmits the data features to the fitting layer;

[0072] The fitting layer performs difference fitting based on the relationship between the previous and next data. The calculation formula of the fitting layer is:

[0073] fit 0 =ω 34 fil 1 +b 4

[0074]

[0075]

[0076] Among them, fit 0 Represents the input of the fitting layer, fil 1 represents the output of the filter layer, ω 34 represents the connection weight between the fitting layer and the filtering layer, b 4 Represents the bias of the fitting layer, fit 1 represents the output of the fitting layer, Represents the mean of the input data of the fitting layer. The fitting layer will fit 1 Send to the self-learning layer;

[0077] The self-learning layer performs multiple self-verification and correction on the fitting results. The calculation formula of the self-learning layer is:

[0078] sl 0 =ω 45 fit 1 +b 5

[0079] sl 1 =θ×sl 0

[0080] Among them, sl 0 represents the input of the self-learning layer, sl 1 represents the output of the self-learning layer, ω 45 represents the connection weight between the fitting layer and the self-learning layer, b 5 represents the bias of the self-learning layer, and θ represents the learning factor. The self-learning layer transmits the final result to the output layer;

[0081] The output layer outputs the final polynomial expression function, the monitoring perception data is input into the polynomial expression function to obtain the corresponding entity object parameters, the calculated entity object parameters and the actual entity object parameters are subjected to error analysis, and the parameters in the deep fitting neural network are corrected according to the error analysis until the calculation accuracy of the deep fitting neural network reaches the expected level, thereby obtaining a trained deep fitting neural network; the error analysis and parameter correction method adopts the existing technology.

[0082] The monitoring perception data is updated in real time, thereby obtaining a digital twin data cockpit that is synchronized with reality. The simulation analysis module 70 obtains the analysis result of each indicator according to the calculation process of each indicator in the indicator parameter, and sends the indicator analysis result to the visualization module 80, which displays it on the large screen of the data cockpit, providing users with a high degree of realism and comprehensive data analysis results to assist users in making relevant decisions.

[0083] The beneficial effects of step S2 are: based on the training and learning of multi-layer neural networks, a mathematical model between physical object parameters and monitoring perception data is constructed to achieve instant information interaction between physical objects and virtual simulation models, so as to establish a mirror relationship between physical objects and virtual simulation models, achieve virtual-real fusion, and ensure the accuracy of various indicators in the data cockpit; integrating physical information into the neural network can better capture data fluctuations, and at the same time, based on the simulation results, the neural network self-learning evolution is carried out, the neural network model is dynamically updated, and the fitting accuracy is improved. The use of the data cockpit system will make decisions based on data more precise and convenient.

[0084] In summary, the data cockpit system based on virtual-reality fusion described in the present invention is completed.

[0085] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0086] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0087] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0088] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for implementing a data cockpit system based on virtual-reality fusion, characterized in that: The following steps are involved: S1. Build a digital twin model, use digital twin technology to measure, simulate, and analyze real-world scenarios, perceive, diagnose, and predict the state of physical objects in real time, and establish indicator parameters of different dimensions that require data visualization analysis; S2. Based on the training and learning of multi-layer neural networks, a mathematical model between physical object parameters and monitoring perception data is constructed to realize instant information interaction between physical objects and virtual simulation models, and a polynomial expression function is fitted; a deep fitting neural network model is constructed, N groups of M-dimensional monitoring perception data and their corresponding physical object parameters are selected as training samples, and N groups of M-dimensional detection perception data are sequentially input into the deep fitting neural network model. Through the learning and training of the neural network, the model output is obtained, and an error analysis is performed on the physical object parameters corresponding to the input data. According to the error analysis, the parameters in the neural network model are corrected, so that the accuracy of the deep fitting neural network reaches the preset target; The deep fitting neural network includes an input layer, a feature division layer, a filtering layer, a fitting layer, a self-learning layer and an output layer; N×M vectors D={D 1 , D 2 , …D N } is sent to the deep fitting neural network model for training. For any set of M-dimensional vectors For any N vectors of any dimension The specific processing process of the deep fitting neural network model is as follows: The N×M vectors D are sent to the input layer, and the input layer uniformly assigns values ​​to the training samples. The calculation formula of the input layer is: Among them, D j represents the dataset of the jth dimension after assignment, Represents the i+1th data under the jth dimension; the input layer transmits the processed data to the feature partitioning layer; The feature partitioning layer divides all data into multiple blocks through a sliding window. The calculation formula of the feature partitioning layer is: Among them, f 0 represents the input of the feature extraction layer, ω 12 represents the connection weight between the input layer and the feature extraction layer, b2 represents the bias of the feature extraction layer, and f 1 represents the output of the feature extraction layer, l min represents the minimum sliding window step size, l max Indicates the maximum sliding window step size; local fluctuation information is captured through the hyperbolic tangent function, and data features of different precisions are obtained by adjusting the window step size. The feature division layer transmits the divided data to the filtering layer; The filtering layer selects the data that meets the preset threshold in each window as data features and transmits the data features to the fitting layer; The fitting layer performs difference fitting based on the relationship between the previous and next data. The calculation formula of the fitting layer is: fit 0 =ω 34 file 1 +b4 Among them, fit 0 Represents the input of the fitting layer, fil 1 represents the output of the filter layer, ω 34 represents the connection weight between the fitting layer and the filtering layer, b4 represents the bias of the fitting layer, and fit 1 represents the output of the fitting layer, Represents the mean of the input data of the fitting layer; the fitting layer will fit 1 Send to the self-learning layer; The self-learning layer performs multiple self-verification and correction on the fitting results. The calculation formula of the self-learning layer is: sl 0 =ω 45 fit 1 +b5 sl 1 =θ×sl 0 Among them, sl 0 represents the input of the self-learning layer, sl 1 represents the output of the self-learning layer, ω 45 represents the connection weight between the fitting layer and the self-learning layer, b5 represents the bias of the self-learning layer; the self-learning layer transmits the final result to the output layer; The output layer outputs the final polynomial expression function, inputs the monitoring perception data into the polynomial expression function, obtains the corresponding entity object parameters, performs error analysis between the calculated entity object parameters and the actual entity object parameters, and corrects the parameters in the deep fitting neural network according to the error analysis until the calculation accuracy of the deep fitting neural network reaches the expected level, thereby obtaining a trained deep fitting neural network; Update detection perception data in real time to obtain a digital twin data cockpit synchronized with reality and display indicator analysis results.

2. The method for implementing a data cockpit system based on virtual-reality fusion as claimed in claim 1, characterized in that: The step S1 specifically includes: The raw data is obtained, preprocessed, and the preprocessed data is edited to form data sets with different attributes, where the attributes include classification fields of the data; the data sets are structured according to the classification fields and stored in the data warehouse of the data cockpit system.

3. A data cockpit system based on virtual-reality fusion, applied to the implementation method of a data cockpit system based on virtual-reality fusion as claimed in claim 1, characterized in that: Includes the following sections: Information acquisition module, model building module, data warehouse, indicator setting module, data request module, deep learning module, simulation analysis module and visualization module; The model construction module is used to build a digital twin model through an interactive window editing mode or a structured data instantiation mode, and establish a mapping relationship between model elements and physical object elements. The model construction module sends the constructed digital twin model to the visualization module through data transmission; The indicator setting module is used to set indicator parameters of different dimensions that need to be analyzed by data visualization. The indicator setting module sends the data required for the indicator to the data request module by means of data transmission; The deep learning module is used to build a deep fitting neural network model based on training and learning of a multi-layer neural network, obtain a mathematical model between physical object parameters and monitoring perception data, realize instant information interaction between physical objects and virtual simulation models, and fit a polynomial expression function. The deep learning module sends the calculation result of the fitting function to the model building module via data transmission.

Images