A mine internet of things terminal anomaly detection method based on digital twinning

Abstract

The application belongs to the technical field of mine Internet of Things, and particularly relates to a mine Internet of Things terminal abnormality detection method based on digital twinning, which comprises the following steps: S1, constructing a mine Internet of Things digital twinning mapping with topological perception, setting a time sliding window, and collecting communication data of a target terminal and a neighbor terminal set within a physical radius of the target terminal; S2, calculating a neighborhood spatial isolation index according to the fluctuation correlation of the target terminal and the neighbor terminal set in received signal strength; and S3, calculating a hardware timing deviation degree according to the confusion degree of the data packet arrival time interval distribution of the target terminal and the received signal strength. The application utilizes spatial coordination logic and physical deviation logic to accurately identify hidden hardware faults such as loose antennas and chip aging while effectively eliminating environmental false alarms such as mine car shielding.

Description

Technical Field

[0001] This invention belongs to the field of mining Internet of Things (IoT) technology, specifically relating to a method for detecting anomalies in mining IoT terminals based on digital twins. Background Technology

[0002] In the construction of modern smart mines, a massive number of IoT terminals are deployed in underground roadways to monitor gas concentration, roof pressure, and personnel location in real time. These terminals typically transmit data back to the ground control center via wireless sensor networks or 5G private networks. Because mine roadways are typically narrow and confined spaces with rough rock walls and high-power electromechanical equipment such as coal mining machines and tunneling machines, the communication environment faces extremely strong multipath fading and sudden electromagnetic interference. This complex electromagnetic environment makes the terminal communication links highly unstable, posing a significant challenge to the accurate monitoring of equipment status.

[0003] Current technical methods for detecting terminal anomalies mainly include heartbeat detection mechanisms and signal strength threshold methods. Heartbeat detection mechanisms set timeout thresholds in the cloud; if no data packets are received from the terminal within a specified time, it is considered a disconnection. Signal strength threshold methods monitor the received signal strength and trigger an alarm when the value falls below a certain fixed threshold. These methods are logically simple and computationally inefficient, playing a role in early mine information system development, primarily for determining whether equipment has completely failed.

[0004] However, existing technologies are mainly based on single-point, single-dimensional threshold judgments, which cannot distinguish between external environmental interference and internal hardware failures, leading to the following prominent problems: First, false alarms caused by environmental interference. When a mine car passes by or a damper closes, causing signal blockage, the received signal strength will drop sharply. At this time, the terminal hardware is intact, but existing technologies will misjudge it as a failure, increasing unnecessary maintenance costs. Second, missed detection of hidden hardware failures. Some terminals may experience data transmission timing disorder due to antenna impedance mismatch, crystal oscillator aging, or voltage regulator circuit failure, but their transmission power still remains at a normal level. Existing technologies cannot identify this sub-healthy state of strong signal but poor quality, resulting in the long-term existence of potential faults, which may lead to data loss at critical moments. Summary of the Invention

[0005] This invention provides a digital twin-based method for detecting anomalies in mining IoT terminals, which solves the technical problems in the prior art that cannot distinguish between environmental interference and terminal malfunctions, resulting in a high false alarm rate in complex mining environments, and cannot identify hidden faults with normal signal strength but chaotic internal logic.

[0006] This invention provides a method for anomaly detection of IoT terminals in mines based on digital twins, comprising the following steps: S1, construct a topology-aware digital twin mapping of the mine Internet of Things, set a time sliding window, and collect communication data of the target terminal and its neighboring terminal set within its physical radius; S2, calculate the neighborhood spatial isolation index based on the fluctuation correlation of the target terminal and the set of neighboring terminals in the received signal strength; S3, calculate the hardware timing deviation based on the disorder of the data packet arrival time interval distribution of the target terminal and the received signal strength; S4 calculates a comprehensive fault confidence score based on the neighborhood spatial isolation index and hardware timing deviation, and compares the comprehensive fault confidence score with a preset threshold to achieve anomaly detection of mining IoT terminals.

[0007] Its effects are as follows: In the complex electromagnetic environment of mines, this invention integrates spatial coordination logic and physical divergence logic. By establishing a digital twin model to comprehensively evaluate neighborhood isolation and hardware timing divergence, it can intelligently and accurately distinguish between general signal fluctuations caused by environmental interference such as mine car obstruction and hidden hardware faults caused by loose terminal antennas, crystal oscillators, etc. in practical applications. It solves the pain points of high false alarm rate and missed deep faults in traditional methods in complex scenarios, and greatly improves the maintenance efficiency and data monitoring security of the mine Internet of Things system.

[0008] Furthermore, in the process of calculating the neighborhood spatial isolation index, the Pearson correlation coefficient between the received signal strength sequence of the target terminal and the mean received signal strength sequence of the neighboring terminal set is first calculated. If the Pearson correlation coefficient approaches 1, it indicates that the target terminal fluctuates synchronously with the environment and the neighborhood spatial isolation index is low. If the Pearson correlation coefficient approaches 0 and the target terminal itself fluctuates violently, the neighborhood spatial isolation index increases significantly.

[0009] Its effect is that, compared with traditional isolated monitoring that ignores the spatial correlation between nodes, this invention introduces the Pearson correlation coefficient to accurately quantify the synchronization of the target terminal and its neighboring nodes in signal fluctuations. In mining applications, it can quickly isolate regional environmental interference and provide a reliable reference mechanism for judging whether the equipment has experienced independent individual anomalies that are detached from the environment.

[0010] Furthermore, the neighborhood spatial isolation index The calculation formula is as follows:

[0011] In the formula, The standard deviation of the received signal strength sequence of the target terminal within the current time sliding window; It is the arithmetic mean of the standard deviations of the received signal strength of all terminals in the neighboring terminal set; It is a smoothing constant; The Pearson correlation coefficient is the ratio between the received signal strength sequence of the target terminal and the mean received signal strength sequence of the neighboring terminal set. For the natural constant An exponential function with base 0.

[0012] Its effect is that the isolation index formula constructed in this invention integrates the fluctuation range and correlation, and cleverly introduces the smoothing constant and the mean of the standard deviation of the neighbors as denominators, so that the system can automatically reduce the detection sensitivity when facing the generally harsh mine environment and large fluctuations in the overall signal, effectively suppressing the batch false alarms caused by the deterioration of environmental consistency, and enhancing the robustness of the algorithm under extreme working conditions.

[0013] Furthermore, in the process of calculating the hardware timing deviation, the arrival time interval sequence of adjacent data packets is first calculated, and the arrival time interval sequence is discretized and distributed to calculate its Shannon entropy to characterize the transmission timing entropy; the larger the transmission timing entropy, the more chaotic the packet sending rhythm.

[0014] Its effect is that, compared with the traditional heartbeat timeout detection that only focuses on whether the data packets arrive on time, this invention transforms the arrival time interval into Shannon entropy, which can extremely sensitively capture the micro-level disorder of the device's packet sending rhythm in actual communication scenarios. This enables the operation and maintenance system to perceive the sub-health state of the device and to identify deep hardware logic defects that cannot be found by the traditional threshold method in advance.

[0015] Furthermore, hardware timing deviation The calculation formula is as follows:

[0016] In the formula, The transmission time sequence entropy within the current time sliding window; The relative gain value is the average received signal strength of the target terminal; The minimum communicable signal strength of the target terminal; The effective dynamic range of the signal strength; This is the penalty gain coefficient; This is a constraint function used to ensure that the value within the parentheses is not less than 0.

[0017] Its effect is as follows: Based on the physical constraint that the stronger the signal, the smaller the data jitter should be, the present invention constructs a nonlinear penalty gain model. In the strong signal area of ​​the mining area, once the terminal has a phenomenon that violates the physical law of extremely good signal but extremely disordered packet transmission timing, the model can amplify the abnormal features by a factor of penalty and accurately locate the hidden hardware faults that are easily ignored by traditional methods.

[0018] Furthermore, the calculation of the comprehensive fault confidence score includes: normalizing the product of the neighborhood spatial isolation index and the hardware timing deviation using a calibration constant; Comprehensive Fault Confidence Score The calculation formula is as follows:

[0019] In the formula, It is the spatial isolation index of the neighborhood; Hardware timing deviation; For digital twin calibration constants, the digital twin calibration constants are based on the target terminal's historical health status. The average value is determined.

[0020] Furthermore, constructing a topology-aware digital twin mapping for the Internet of Things in a mine includes: establishing a physical coordinate mapping table of the terminal on a cloud server by parsing drawings or geographic information system data of the mine roadways, and dynamically maintaining the logical adjacency relationship between the target terminal and the set of neighboring terminals based on a wireless signal propagation model.

[0021] Furthermore, the calculation process for the standard deviation of the received signal strength sequence of the target terminal and the arithmetic mean of the standard deviations of the received signal strength of all terminals in the neighboring terminal set includes: acquiring multiple received signal strength sample values ​​within a sliding window of acquisition time; calculating the arithmetic mean of the sample values; calculating the sum of squares of the differences between each sample value and the arithmetic mean; dividing the sum of squares by the total number of samples and then taking the square root to obtain the standard deviation, which characterizes the dispersion of the signal.

[0022] Furthermore, the relative gain value of the average received signal strength of the target terminal is obtained as follows: the average value of the received signal strength of the target terminal within the current time sliding window is obtained, and the receiving sensitivity noise floor value of the target terminal is subtracted from the average value to obtain the relative gain value relative to the noise floor; if the relative gain value is less than zero, the relative gain value is set to zero.

[0023] Furthermore, transmission time entropy The calculation formula is as follows:

[0024] in, This represents the interval index after discretizing the data packet arrival time interval. This represents the total number of intervals after discretizing the data packet arrival time interval. For the first The probability of a data packet arrival time interval occurring within a discrete interval. Represents the natural logarithm operation.

[0025] The beneficial effects are: This invention proposes a detection concept that combines spatial coordination logic and physical divergence logic. In terms of spatial coordination, neighboring nodes are used as a dynamic reference system. If the signals of all nodes in the region fluctuate synchronously, it is determined to be environmental interference. If only the target node fluctuates, it is determined to be an individual anomaly. This mechanism greatly improves the system's anti-interference capability in dynamic and complex environments.

[0026] Regarding physical deviations, this invention utilizes the physical constraints of signal-to-noise ratio and timing jitter. Based on the normal principle that stronger signals result in less jitter, a deviation model is constructed. If a deviation phenomenon of high signal strength and high timing jitter occurs, it is accurately determined to be a hardware logic fault. This mechanism enables the system to perceive sub-health states and identify deep-seated hardware problems that traditional threshold methods cannot detect, significantly improving the maintenance efficiency and security of the mining IoT system. Attached Figure Description

[0027] Figure 1 This is a flowchart of the anomaly detection method for mining IoT terminals based on digital twins according to the present invention.

[0028] Figure 2 This is a thermal topology diagram of the mine space according to the present invention.

[0029] Figure 3 This is a scatter plot of the signal-to-noise divergence feature of the present invention.

[0030] Figure 4 This is a spatial comparison diagram of the detection effect of the present invention. Detailed Implementation

[0031] 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, not all, of the embodiments of the present invention. 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.

[0032] An embodiment of the anomaly detection method for mining IoT terminals based on digital twins provided by this invention: like Figure 1 As shown, the anomaly detection method for IoT terminals in mines based on digital twins includes the following steps: S1. Construct a topology-aware digital twin mapping for the mine's Internet of Things (IoT), set a time sliding window, and collect communication data from the target terminal and its neighboring terminals within its physical radius.

[0033] Specifically, this step aims to establish a mapping relationship between the physical world and the digital world, providing a data foundation for subsequent spatial analysis.

[0034] First, a digital twin of the mine's Internet of Things (IoT) is built on a cloud server. This process involves analyzing CAD drawings or geographic information system data from the mine tunnels to determine the physical coordinates of all IoT terminals. It maps to a virtual space and maintains the physical location and logical topology of the device in real time.

[0035] Secondly, set a time sliding window. For example, take Seconds, with each terminal to be detected as the target terminal. Searching for its physical radius based on Euclidean distance All other terminals within the range constitute the neighbor terminal set. .

[0036] Finally, communication data for each device in the target terminal and neighboring terminal sets is collected in real time within a time window. The collected metrics include two core dimensions: time Received signal strength The unit is dBm, which reflects the signal quality of the link.

[0037] timestamps of data packets arriving in the cloud The unit is milliseconds (ms), used to analyze transmission timing characteristics.

[0038] For example, in a fully mechanized mining face scenario, the hydraulic support pressure sensor numbered ID-001 is selected as the target terminal, and it is automatically associated with six other sensors within a 50-meter radius around it as a neighbor set, continuously collecting the received signal strength value and arrival time of all reported data packets within 30 seconds.

[0039] By constructing a topology-aware digital twin map and acquiring multidimensional data, a foundational data support with spatiotemporal correlation is provided for subsequent anomaly analysis. Traditional single-point monitoring methods ignore the spatial relationships between devices, while this step, by maintaining the neighbor set in real time, enables the system to perceive environmental changes within a local area. This ensures that subsequent steps can distinguish between general signal fluctuations caused by regional environmental deterioration and abnormal behavior specific to a single device, laying the foundation for spatial collaborative analysis.

[0040] S2, calculate the neighborhood spatial isolation index based on the fluctuation correlation of the target terminal and the neighboring terminal set in the received signal strength.

[0041] Specifically, this step aims to quantify the degree to which the target terminal is independent of its surrounding environment in order to isolate it from environmental interference. The physical basis for this is that in mine tunnels, environmental interference has regional consistency and will affect all equipment in the entire area simultaneously.

[0042] First, calculate the Pearson correlation coefficient. Obtain the received signal strength sequence of the target terminal within the window. and the sequence of average received signal strength of all neighboring terminals Calculate the correlation coefficient between the two:

[0043] In the formula, For the target terminal within the time sliding window The received signal strength sample value at each moment. For the neighbor terminal set in the th The average received signal strength at each time point. These are the sequence mean; To prevent extremely small positive numbers with a denominator of 0, the value is taken as... , The range of values ​​is A value close to 1 indicates a positive correlation, while a value close to 0 indicates no correlation.

[0044] Next, the neighborhood isolation index is calculated. This index combines the volatility of the target terminal with the correlation of its neighborhood, and the calculation formula is as follows:

[0045] In the formula: This represents the standard deviation of the signal strength sequence received by the target terminal within the current time window, reflecting the degree of fluctuation of the target itself; the calculation formula is... ; It represents the arithmetic mean of the standard deviations of the received signal strength of all terminals in the neighbor set, reflecting the general fluctuation level of the current environment; This represents the smoothing constant, with a value of 1.0, used to prevent... A value of 0 causes division to fail, and a baseline threshold for fluctuation is set. For the natural constant An exponential function with base 0.

[0046] Calculation example: Scenario 1, Environmental Interference: The passage of a mining truck causes widespread and significant signal fluctuations in the area; assume the standard deviation of the signal strength sequence received by the target terminal. dB; dB; due to the influence of the same interference source, the two are highly correlated, and the calculated values ​​are... ;but ;result: The value is low, and the system determines that the fluctuation is due to environmental following rather than an individual fault.

[0047] Scenario 2, Individual Fault: The target terminal's antenna is loose, and the environment is calm; assume the standard deviation of the target terminal's received signal strength sequence. dB; Stable neighboring environment dB; since only the target fluctuates, the two are uncorrelated, and the calculated value is... ;but ;result: If the value increases significantly, the system determines that the target terminal is detached from the environment and is abnormal, suggesting a possible malfunction.

[0048] By calculating the neighborhood spatial isolation index, a neighborhood reference system is introduced, and the following is utilized: As a denominator term, it automatically reduces detection sensitivity when the environment is generally harsh, effectively suppressing false alarms caused by fluctuations in environmental consistency. This step solves the problem of frequent false alarms caused by environmental factors such as the start-up and shutdown of electromechanical equipment and obstruction by mine cars in underground mines. It can intelligently distinguish between environmental problems and its own problems, significantly reducing ineffective maintenance costs.

[0049] S3, calculate the hardware timing deviation based on the disorder of the data packet arrival time interval distribution of the target terminal and the received signal strength.

[0050] Specifically, this step aims to detect hidden hardware faults where the signal strength is good but the transmission timing is disordered. The physical basis for this is that the higher the channel quality, the smaller the jitter in data transmission should be. If the signal is extremely strong but the jitter is still large, it violates the laws of physics and must be due to a fault in the terminal's internal processing logic or clock source.

[0051] First, calculate the transmission time entropy. Extract the arrival time interval sequence of adjacent data packets. Discretize the time interval values, for example, by binning them into 10ms intervals, and calculate the probability of each interval occurring. Calculate its Shannon entropy:

[0052] in, This represents the interval index after discretizing the data packet arrival time interval. This represents the total number of intervals after discretizing the data packet arrival time interval.

[0053] The larger the value, the more irregular the packet sending intervals and the more chaotic the rhythm.

[0054] Then, construct the hardware timing deviation. The calculation formula is as follows:

[0055] In the formula: This represents the relative gain value of the average received signal strength of the target terminal within the current window, assuming the device noise floor. dBm, the current average received signal strength is dBm, then dB; This represents the minimum communicable signal strength of the target terminal, for example, 10dB; The effective dynamic range of signal strength, for example, 60 dB; This represents the penalty gain coefficient, used to adjust the penalty for bad quality in a good signal; in this embodiment, it is set to 3.0. This indicates a function that restricts the value within the parentheses to be no less than 0.

[0056] Calculation example: Suppose that the internal clock crystal of a certain terminal is aging, causing the packet transmission timing to be very disordered, and the calculated timing entropy... .

[0057] Scenario A, Weak Signal Area: The device is located in the edge area, with an average received signal strength of -95dBm; dB, at this time , Analysis: Under weak signal conditions, large jitter is considered a normal manifestation of poor channel performance, and the system does not impose additional penalties.

[0058] Situation B, strong signal area: The device is located near a base station, and the average received signal strength is -55dBm; dB, at this time dB Analysis: Under strong signals, the system identifies a deviation phenomenon where the signal is excellent but the jitter is extremely large. The penalty factor takes effect, leading to hardware timing deviation. It is magnified three times to highlight the fault.

[0059] By calculating hardware timing deviation and utilizing physical layer channel constraints, a nonlinear penalty model is constructed. This model can keenly detect sub-healthy states where signal strength is normal but internal timing is disordered, thereby accurately identifying hidden faults such as antenna impedance mismatch and crystal oscillator aging. This mechanism overcomes the shortcomings of existing technologies that only focus on signal strength while ignoring the physical consistency of signal quality, effectively preventing the underreporting of hidden faults.

[0060] S4 calculates a comprehensive fault confidence score based on the neighborhood spatial isolation index and hardware timing deviation, and compares the comprehensive fault confidence score with a preset threshold to achieve anomaly detection of mining IoT terminals.

[0061] Specifically, this step integrates features from both spatial and physical dimensions to output the final result, which includes a comprehensive fault confidence score. The calculation formula is as follows:

[0062] In the formula, A calibration constant for the digital twin, which is obtained by analyzing the terminal's health status over the past week. The historical average value was used to eliminate the influence of individual differences at different terminals.

[0063] The preset alarm threshold is 1.0. If the system detects an anomaly in the terminal, it will trigger an alarm and push it to the maintenance personnel's terminal.

[0064] The formula uses multiplicative logic, and the score will only exceed the threshold when the terminal is isolated from the environment and the terminal violates the physical law of signal-to-noise ratio.

[0065] like Figure 2 As shown, this map uses the actual geographical layout of mine roadways as a base, accurately marking the physical coordinates of all IoT terminals underground. It also employs a heatmap color gradient to visually represent the real-time received signal strength of each terminal; different color depths correspond to different received signal strength values, clearly distinguishing areas of strong and weak signals. The map directly reflects the regional consistency characteristics of mine environmental interference; that is, environmental factors such as mine car obstruction and electromechanical equipment interference can cause synchronous fluctuations in the signals of all terminals within a certain local area. This provides intuitive visualization support for using the neighborhood spatial isolation index to isolate environmental interference and distinguish between environmental fluctuations and individual terminal anomalies, providing a clear visual representation of spatial topology and signal distribution.

[0066] like Figure 3 As shown in the figure, the horizontal axis represents the relative gain of the received signal strength, and the vertical axis represents the transmission time entropy. Each scatter point represents the signal and timing characteristics of the terminal within a sliding time window. The figure clearly presents two types of data distributions: normal terminals conform to the physical law that the higher the signal strength, the lower the transmission time entropy, while terminals with hidden hardware faults exhibit the opposite characteristics of high signal strength and high transmission time entropy. This figure intuitively depicts the sample distribution patterns of hidden faults such as loose antennas and aging crystal oscillators, providing a visual basis for calculating hardware timing deviation and identifying hidden faults.

[0067] like Figure 4 As shown, the left figure illustrates the detection results of the existing threshold method, which shows a large number of false alarms in the area where the mining truck passes, and fails to detect the hidden fault points on the right. The right figure illustrates the detection results of the method of the present invention, which eliminates all false alarms in the environmental interference area and accurately marks the hidden fault points that were missed by the existing technology.

[0068] By comprehensively considering fault confidence levels, multiplicative logic is employed to deeply integrate the characteristics of both spatial isolation and physical deviation, fundamentally eliminating the one-sidedness of single-dimensional judgments. This step ensures that alarms are triggered only when the terminal exhibits unique and physically incompatible dual anomalies, significantly improving the system's robustness in the complex electromagnetic environment of mines and achieving high-precision anomaly detection.

[0069] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for anomaly detection in mine IoT terminals based on digital twins, characterized in that, Includes the following steps: S1, construct a topology-aware digital twin mapping of the mine Internet of Things, set a time sliding window, and collect communication data of the target terminal and its neighboring terminal set within its physical radius; S2, calculate the neighborhood spatial isolation index based on the fluctuation correlation of the target terminal and the set of neighboring terminals in the received signal strength; S3, calculate the hardware timing deviation based on the disorder of the data packet arrival time interval distribution of the target terminal and the received signal strength; S4 calculates a comprehensive fault confidence score based on the neighborhood spatial isolation index and hardware timing deviation, and compares the comprehensive fault confidence score with a preset threshold to achieve anomaly detection of mining IoT terminals.

2. The anomaly detection method for mine IoT terminals based on digital twins according to claim 1, characterized in that, In calculating the neighborhood spatial isolation index, the Pearson correlation coefficient between the received signal strength sequence of the target terminal and the mean received signal strength sequence of the neighboring terminal set is first calculated. If the Pearson correlation coefficient approaches 1, it indicates that the target terminal fluctuates synchronously with the environment and the neighborhood spatial isolation index is low; if the Pearson correlation coefficient approaches 0 and the target terminal itself fluctuates violently, the neighborhood spatial isolation index increases significantly.

3. The anomaly detection method for mine IoT terminals based on digital twins according to claim 1, characterized in that, Neighborhood Spatial Isolation Index The calculation formula is as follows: In the formula, The standard deviation of the received signal strength sequence of the target terminal within the current time sliding window; It is the arithmetic mean of the standard deviations of the received signal strength of all terminals in the neighboring terminal set; It is a smoothing constant; The Pearson correlation coefficient is the ratio between the received signal strength sequence of the target terminal and the mean received signal strength sequence of the neighboring terminal set. For the natural constant An exponential function with base 0.

4. The anomaly detection method for mine IoT terminals based on digital twins according to claim 1, characterized in that, In calculating the hardware timing deviation, the arrival time interval sequence of adjacent data packets is first calculated, and the arrival time interval sequence is discretized and distributed. The Shannon entropy is then calculated to characterize the transmission timing entropy. The larger the transmission timing entropy, the more chaotic the packet sending rhythm.

5. The anomaly detection method for mine IoT terminals based on digital twins according to claim 4, characterized in that, Hardware timing deviation The calculation formula is as follows: In the formula, The transmission time sequence entropy within the current time sliding window; The relative gain value is the average received signal strength of the target terminal; The minimum communicable signal strength of the target terminal; The effective dynamic range of the signal strength; This is the penalty gain coefficient; This is a constraint function used to ensure that the value within the parentheses is not less than 0.

6. The anomaly detection method for mine IoT terminals based on digital twins according to claim 5, characterized in that, The calculation of the comprehensive fault confidence score includes: normalizing the product of the neighborhood spatial isolation index and the hardware timing deviation using a calibration constant; Comprehensive Fault Confidence Score The calculation formula is as follows: In the formula, It is the spatial isolation index of the neighborhood; Hardware timing deviation; For digital twin calibration constants, the digital twin calibration constants are based on the target terminal's historical health status. The average value is determined.

7. The anomaly detection method for mine IoT terminals based on digital twins according to claim 1, characterized in that, Constructing a topology-aware digital twin mapping for the Internet of Things in a mine includes: establishing a physical coordinate mapping table of terminals on a cloud server by parsing drawings or geographic information system data of mine roadways, and dynamically maintaining the logical adjacency relationship between the target terminal and the set of neighboring terminals based on a wireless signal propagation model.

8. The anomaly detection method for mine IoT terminals based on digital twins according to claim 3, characterized in that, The calculation process for the standard deviation of the received signal strength sequence of the target terminal and the arithmetic mean of the standard deviations of the received signal strength of all terminals in the neighboring terminal set includes: collecting multiple received signal strength sample values ​​within a sliding window of acquisition time; calculating the arithmetic mean of the sample values; calculating the sum of squares of the differences between each sample value and the arithmetic mean; dividing the sum of squares by the total number of samples and then taking the square root to obtain the standard deviation, which characterizes the dispersion of the signal.

9. The anomaly detection method for mine IoT terminals based on digital twins according to claim 5, characterized in that, The relative gain value of the average received signal strength of the target terminal is obtained by: obtaining the average value of the received signal strength of the target terminal within the current time sliding window, and subtracting the receiving sensitivity noise floor value of the target terminal from the average value to obtain the relative gain value relative to the noise floor. If the relative gain value is less than zero, then the relative gain value is set to zero.

10. The anomaly detection method for mine IoT terminals based on digital twins according to claim 5, characterized in that, Transmission time entropy The calculation formula is as follows: in, This represents the interval index after discretizing the data packet arrival time interval. This represents the total number of intervals after discretizing the data packet arrival time interval. For the first The probability of a data packet arrival time interval occurring within a discrete interval. Represents the natural logarithm operation.

Images