Virtual reality digital twin real-time mapping method based on global internet of things sensors

Abstract

The application relates to the technical field of digital twinning, and discloses a virtual reality digital twinning real-time mapping method based on a global Internet of Things sensor, which comprises the following steps: acquiring global Internet of Things sensor data, constructing data labels, calculating based on the global Internet of Things sensor data, forming a multi-dimensional feature vector set, carrying out numerical prediction based on the multi-dimensional feature vector set, calculating a measured data set residual error based on the predicted data, comparing the measured data set residual error with a threshold, carrying out sensor determination, if the sensor determination is normal, determining that the collected data can be directly used for real-time mapping of a digital twinning model, if the sensor determination is abnormal, positioning an abnormal type according to the abnormal sensor determination. The system effectively avoids the accumulation of small errors of the sensor, guarantees the reliability of input data of the digital twinning model, and finally improves the accuracy of physical entity state judgment and decision making.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology, specifically to a real-time mapping method for virtual reality digital twins based on global Internet of Things (IoT) sensors. Background Technology

[0002] Real-time mapping of virtual reality digital twins is an advanced technological concept that combines virtual reality (VR), digital twins, and real-time data processing. Its aim is to create a virtual environment that closely resembles the physical world and allows for real-time interaction, accurately reflecting various objects, systems, or processes from the real world within this virtual space. By constructing digital replicas within this virtual environment, people can use virtual reality devices to observe, analyze, and manipulate these virtual models in an immersive way, just as in the real world. This provides new ways of understanding, decision support, and optimization solutions for many fields, helping to improve efficiency, reduce costs, and enhance overall performance.

[0003] In real-time mapping of virtual reality digital twins, IoT sensors are mainly relied upon as key devices for data acquisition. However, the accuracy of IoT sensors directly affects the reliability of the acquired data. Even tiny measurement errors can accumulate over time during the aggregation of large amounts of data and long-term operation, leading to deviations in the final digital twin model and consequently affecting the accuracy of judgments and decisions regarding the physical entity's state. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] To address the shortcomings of existing technologies, this invention provides a real-time mapping method for virtual reality digital twins based on omni-domain IoT sensors. This method systematically solves the core problem of digital twin model deviation caused by the accumulation of sensor measurement errors through the collaborative construction of a data acquisition module, a linkage processing module, a virtual-real verification module, a sensor judgment module, and a sensor anomaly identification module. It effectively avoids the accumulation of minute sensor errors, ensures the reliability of input data for the digital twin model, and ultimately improves the accuracy of judging and making decisions about the state of physical entities.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the present invention provides the following technical solution: a real-time mapping method for virtual reality digital twins based on global IoT sensors, comprising the following steps; Step 1: Build a data acquisition module to obtain IoT sensor data across the entire domain and construct data tags; Step 2: Build a linkage processing module to perform calculations based on the data from IoT sensors across the entire domain, forming a set of multi-dimensional feature vectors; Step 3: Build a virtual-real verification module, which is used to perform numerical prediction based on a multi-dimensional feature vector set, and calculate the residual of the actual dataset based on the predicted data; Step 4: Build a sensor judgment module. Based on the comparison between the residual of the measured dataset and the threshold, perform sensor judgment. If the sensor judgment is normal, the judgment data can be directly used for real-time mapping of the digital twin model. If the sensor judgment is abnormal, proceed to the next step. Step 5: Construct a sensor anomaly identification module to determine abnormal situations and locate the anomaly type for the sensor.

[0008] Preferably, in step one, the global IoT sensor data includes real-time monitoring data from multiple IoT sensors deployed across the entire domain. These real-time data are used to build a unified data acquisition gateway based on the Industrial Internet Protocol and are numbered to form a global IoT sensor data set.

[0009] Preferably, the global IoT sensor data set is numbered. The expression is: ; in, Representing the Real-time monitoring data from an IoT sensor; Represents the number of IoT sensors. This represents the data tag of the corresponding IoT sensor; the data tag includes the sensor's unique identifier, deployment location, measurement parameter type, measurement range, accuracy level, and data acquisition timestamp.

[0010] Preferably, in step two, the multi-dimensional feature vector set includes the real-time dynamic characteristics of a single sensor and the correlation characteristics of multiple sensors; The real-time dynamic characteristics of the single sensor The calculation formula is: ; in, Representing the The real-time dynamic characteristics of a single sensor. Representing the One sensor in Real-time monitoring data at all times; The number of time-series data samples representing a single sensor; Representing the The average data from each sensor, Representing the Data deviation values ​​of each sensor , Represents weight, .

[0011] Preferably, the multi-sensor correlation characteristic The calculation formula is: ; in, Represents the spatial correlation of multiple sensors. Represents the temporal coordination of multiple sensors. This indicates consistency in deviations.

[0012] Preferably, the , , The calculation formula is: ; in, Representing the , Pearson correlation coefficient of single-point dynamic characteristic values ​​of a sensor Representing the , Spatial correlation coefficient of each sensor; ; in, Representing the Single-point dynamic characteristic value of a sensor The slope of the time series trend, The mean of the trend slope of the entire sensor area. Represents the maximum value of the trend slope; ; in, A benchmark value representing the single-point dynamic characteristics of a global sensor. This represents the maximum standard deviation under normal historical operating conditions.

[0013] Preferably, step three includes: S1, Geometric parameters based on physical entities Material properties Operating rules Constructing a high-precision virtual twin model that satisfies physical constraints : ; in, Modeling functions driven by physical mechanisms; S2. Based on the simulation output of the virtual twin model under normal operating conditions, determine the first... Reference range of each sensor : ; in, The first, representing the output of the virtual twin model Each sensor during the simulation period The reference data sequence within; Representing the first Minimum and maximum reference values ​​for each sensor during normal operation; This represents the simulation duration under normal operating conditions; S3. Train a time-series prediction model using a multi-dimensional dynamic feature vector set as input and sensor measured data sequences as output. : ; in, Represents the length of the training dataset. Representative time series prediction algorithms; Representing the The first IoT sensor in the Real-time measured data values ​​collected at all times; Representing the The first sensor Multidimensional dynamic feature vector at time step As input, through a time series prediction model The calculated predicted data values; S4, Real-time input of the first The sensor at the first Dynamic feature vector at time step The prediction model, once trained, outputs predicted data values. ; ; in, Representing the One sensor in Predicted data values ​​at any given time; S5. Compare the measured data from the sensor with the predicted data values, and calculate the residual of the measured data. ; ; Preferably, in step four: The first One sensor Residual data of measured time Compare with the residual threshold of the measured data, when If the residual is less than or equal to the measured data residual threshold, then the first... One sensor It should always be running normally; otherwise, it is considered abnormal.

[0014] Preferably, step five includes: B1, retrieve the first Historical time-series data of each abnormal sensor, single-point dynamic characteristic sequence, and its data tag. ; B2, Extraction and the first For each sensor, retrieve the real-time measured data residuals of all sensors in the deployment area. If ≥70% of sensors in the same area simultaneously show measured data residuals greater than the measured data residual threshold, it is determined to be a systematic error. If only the first If the residual of the measured data of a single sensor is greater than the residual threshold of the measured data, it is determined that the sensor is an individual anomaly.

[0015] Compared with existing technologies, this invention provides a real-time mapping method for virtual reality digital twins based on global IoT sensors, which has the following advantages: This invention systematically solves the core problem of digital twin model deviation caused by the accumulation of sensor measurement errors by establishing a collaborative data acquisition module, a linkage processing module, a virtual-real verification module, a sensor judgment module, and a sensor anomaly identification module. First, in step one, a data acquisition module is established to create a unified data acquisition gateway, tagging and numbering all sensor data across the entire domain to form a traceable, standardized, and comprehensive IoT sensor data set, thus avoiding additional errors introduced by data chaos at the source. Then, in step two, a linkage processing module is established to extract the real-time dynamic characteristics of individual sensors and the correlation characteristics of multiple sensors, constructing a multi-dimensional feature vector set. This can capture both small drift errors at single points and identify latent anomalies in multi-sensor collaboration. Finally, in step three, a virtual-real verification module is established… A virtual twin model is constructed based on physical entity parameters. A time-series prediction model is trained to output predicted data values. The residual of the measured data is quantified by comparing the measured values ​​with the predicted values, making previously imperceptible minute errors explicit. Then, in step four, a sensor judgment module is built. Based on the comparison results of the measured data residuals and thresholds, the sensor status is quickly determined, ensuring that normal data is directly input into the twin model and abnormal data triggers subsequent identification processes, blocking the error accumulation path from the data access stage. Finally, in step five, a sensor anomaly identification module is built to identify abnormal situations of the sensors, locate the anomaly type, and complete the accurate source tracing and closed-loop handling of errors. This effectively avoids the accumulation of minute sensor errors, ensures the reliability of the input data of the digital twin model, and ultimately improves the accuracy of judgment and decision-making on the physical entity status. Attached Figure Description

[0016] Figure 1 This is a diagram illustrating the steps of the method of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. It is worth noting that this application also relates to prior art. Since prior art is well known to those skilled in the art, it will not be described in detail in this application.

[0018] Please see Figure 1 A real-time mapping method for virtual reality digital twins based on global IoT sensors includes: Step 1: Build a data acquisition module to obtain IoT sensor data across the entire domain and construct data tags; A unified data acquisition gateway is built based on the Industrial Internet protocol to connect to the globally deployed network. Real-time monitoring data from individual IoT sensors is used to bind data tags to each data point, including the sensor's unique identifier, deployment location, measurement parameter type, measurement range, accuracy class, and data acquisition timestamp. This process forms a standardized set of sensor data across the entire domain, completing the standardized collection of raw data. It solves the problems of inconsistent data formats and difficulty in tracing the source of data from multiple types and manufacturers of sensors, providing a traceable and standardized data source for subsequent error analysis. This avoids additional errors introduced by data chaos from the source and ensures the quality of basic data for error analysis. Step 2: Build a linkage processing module to perform calculations based on the data from IoT sensors across the entire domain, forming a set of multi-dimensional feature vectors; For the first in the dataset Real-time monitoring data from each sensor The mean and deviation values ​​are calculated and weighted to obtain the real-time dynamic characteristics of a single sensor. Then, using indicators such as Pearson correlation coefficient, time series trend slope, and standard deviation of deviation, the spatial correlation of multiple sensors is calculated respectively. Temporal synergy Deviation consistency Weighted integration into multi-sensor correlation characteristics Ultimately, this results in a multi-dimensional dynamic feature vector set that considers both single-point and correlation characteristics. It breaks through the limitations of single sensor data, quantifies sensor data characteristics from two dimensions: single-point data fluctuation and multi-sensor collaborative consistency. It can capture the small drift error of a single sensor and identify hidden anomalies through the correlation of multiple sensors, providing accurate feature input for subsequent error prediction and judgment, and avoiding the problem of isolated data errors being ignored and accumulating. Step 3: Build a virtual-real verification module, which is used to perform numerical prediction based on a multi-dimensional feature vector set, and calculate the residual of the actual dataset based on the predicted data; Geometric parameters based on physical entities Material properties Operating rules Constructing a high-precision virtual twin model that satisfies physical constraints Preset the reference range for normal sensor operation Then, using multi-dimensional dynamic feature vectors as input, a time-series prediction model is trained. Real-time output of sensor predicted data values Finally, the measured data from the sensor is compared with the predicted data values, and the residual of the measured data is calculated. ; By using the theoretical benchmark value provided by the virtual twin model and the predicted value output by the prediction model, a dual data reference system is constructed. The actual sensor data is compared with the theoretical and predicted values, and the originally imperceptible small measurement error is quantified into the residual of the actual measurement data. This realizes the explicit expression of the error and provides a quantifiable core indicator for subsequent error judgment, thus solving the problem of accurate identification before error accumulation at the root. Step 4: Build a sensor judgment module. Based on the comparison between the residual of the measured dataset and the threshold, perform sensor judgment. If the sensor judgment is normal, the judgment data can be directly used for real-time mapping of the digital twin model. If the sensor judgment is abnormal, proceed to the next step. Through measured data residuals Compare with the residual threshold of the measured data, when If the residual is less than or equal to the measured data residual threshold, then the first... One sensor It operates normally at all times; otherwise, it is judged as abnormal, ensuring that the data from normal sensors are directly used for digital twin mapping and avoiding the misjudgment of valid data; at the same time, it accurately locates abnormal sensors to prevent data containing errors from entering the twin model, blocking the accumulation path of errors from the data access stage and ensuring the reliability of the input data of the twin model. Step 5: Construct a sensor anomaly identification module to identify sensor anomalies, locate the anomaly type, and complete the precise source tracing and closed-loop handling of errors. This completely solves the problem of sensor error accumulation, ensures the accuracy of the data input to the digital twin model, and ultimately improves the accuracy of physical entity state judgment and decision-making.

[0019] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A real-time mapping method for virtual reality digital twins based on global IoT sensors, characterized in that, Includes the following steps; Step 1: Build a data acquisition module to obtain IoT sensor data across the entire domain and construct data tags; Step 2: Build a linkage processing module to perform calculations based on the data from IoT sensors across the entire domain, forming a set of multi-dimensional feature vectors; Step 3: Build a virtual-real verification module, which is used to perform numerical prediction based on a multi-dimensional feature vector set, and calculate the residual of the actual dataset based on the predicted data; Step 4: Build a sensor judgment module. Based on the comparison between the residual of the measured dataset and the threshold, perform sensor judgment. If the sensor judgment is normal, the judgment data can be directly used for real-time mapping of the digital twin model. If the sensor judgment is abnormal, proceed to the next step. Step 5: Construct a sensor anomaly identification module to determine abnormal situations and locate the anomaly type for the sensor.

2. The real-time mapping method for virtual reality digital twins based on global IoT sensors according to claim 1, characterized in that, In step one, the global IoT sensor data includes real-time monitoring data from multiple IoT sensors deployed across the entire domain. These real-time data are used to build a unified data acquisition gateway based on the Industrial Internet Protocol and are numbered to form a global IoT sensor data set.

3. The real-time mapping method for virtual reality digital twins based on global IoT sensors according to claim 2, characterized in that, The global IoT sensor data set number The expression is: ； in, Representing the Real-time monitoring data from an IoT sensor; Represents the number of IoT sensors. This represents the data tag of the corresponding IoT sensor; the data tag includes the sensor's unique identifier, deployment location, measurement parameter type, measurement range, accuracy level, and data acquisition timestamp.

4. The real-time mapping method for virtual reality digital twins based on global IoT sensors according to claim 3, characterized in that, In step two, the multi-dimensional feature vector set includes the real-time dynamic characteristics of a single sensor and the correlation characteristics of multiple sensors; The real-time dynamic characteristics of the single sensor The calculation formula is: ； in, Representing the The real-time dynamic characteristics of a single sensor. Representing the One sensor in Real-time monitoring data at all times; The number of time-series data samples representing a single sensor; Representing the The average data from each sensor, Representing the Data deviation values ​​of each sensor , Represents weight, .

5. The real-time mapping method for virtual reality digital twins based on global IoT sensors according to claim 4, characterized in that, The multi-sensor correlation characteristics The calculation formula is: ； in, Represents the spatial correlation of multiple sensors. Represents the temporal coordination of multiple sensors. This indicates consistency in deviations.

6. The real-time mapping method for virtual reality digital twins based on global IoT sensors according to claim 5, characterized in that, The , , The calculation formula is: ； in, Representing the , Pearson correlation coefficient of single-point dynamic characteristic values ​​of a sensor Representing the , Spatial correlation coefficient of each sensor; ； in, Representing the Single-point dynamic characteristic value of a sensor The slope of the time series trend, The mean of the trend slope of the entire sensor area. Represents the maximum value of the trend slope; ； in, A benchmark value representing the single-point dynamic characteristics of a global sensor. This represents the maximum standard deviation under normal historical operating conditions.

7. The real-time mapping method for virtual reality digital twins based on global IoT sensors according to claim 1, characterized in that, Step three includes: S1, Geometric parameters based on physical entities Material properties Operating rules Constructing a high-precision virtual twin model that satisfies physical constraints : ； in, Modeling functions driven by physical mechanisms; S2. Based on the simulation output of the virtual twin model under normal operating conditions, determine the first... Reference range of each sensor : ； in, The first, representing the output of the virtual twin model Each sensor during the simulation period The reference data sequence within; Representing the first Minimum and maximum reference values ​​for each sensor during normal operation; This represents the simulation duration under normal operating conditions; S3. Train a time-series prediction model using a multi-dimensional dynamic feature vector set as input and sensor measured data sequences as output. : ； in, Represents the length of the training dataset. Representative time series prediction algorithms; Representing the The first IoT sensor in the Real-time measured data values ​​collected at all times; Representing the The first sensor Multidimensional dynamic feature vector at time step As input, through a time series prediction model The calculated predicted data values; S4, Real-time input of the first The sensor at the first Dynamic feature vector at time step The prediction model, once trained, outputs predicted data values. ; ； in, Representing the One sensor in Predicted data values ​​at any given time; S5. Compare the sensor's measured data with the predicted data values, and calculate the residual of the measured data. ; 。 8. The real-time mapping method for virtual reality digital twins based on global IoT sensors according to claim 7, characterized in that, In step four: The first One sensor Residual data of measured time Compare with the residual threshold of the measured data, when If the residual is less than or equal to the measured data residual threshold, then the first... One sensor It should always be running normally; otherwise, it is considered abnormal.

9. The real-time mapping method for virtual reality digital twins based on global IoT sensors according to claim 8, characterized in that, Step five includes: B1, retrieve the first Historical time-series data of each abnormal sensor, single-point dynamic characteristic sequence, and its data tag. ; B2, Extraction and the first For each sensor, retrieve the real-time measured data residuals of all sensors in the deployment area. If ≥70% of sensors in the same area simultaneously show measured data residuals greater than the measured data residual threshold, it is determined to be a systematic error. If only the first If the residual of the measured data of a single sensor is greater than the residual threshold of the measured data, it is determined that the sensor is an individual anomaly.

Images