An intelligent tablet anti-intrusion system for field communication

Abstract

This invention relates to the field of smart tablet communication protection technology, specifically a smart tablet anti-intrusion system for field communication. This system uses a smart tablet device to transmit detection signals in a tunnel environment, receives and collects response signals from legitimate communication nodes, and extracts multipath feature parameters. Based on these parameters, it calculates the multipath statistical characteristics of the current channel and compensates for these characteristics using the smart tablet device's real-time movement speed and inertial measurement data. The compensated multipath statistical characteristics are then matched against a tunnel multipath blind zone fingerprint database to obtain a successful match. When a successful match is detected, the smart tablet device is determined to be in a communication blind zone, and a tiered anti-intrusion response is initiated. Once the multipath statistical characteristics are detected to have returned to a non-blind zone channel mode and a secure connection is re-established with a legitimate communication node, the anti-intrusion response is exited, and a blind zone event log is uploaded, effectively solving the communication security and stability problem in tunnel environments.

Description

Technical Field

[0001] This invention relates to the field of smart tablet communication protection technology, specifically a smart tablet anti-intrusion system for outdoor communication. Background Technology

[0002] In field communication scenarios, smart tablets, as important communication terminal devices, are widely used in various work areas requiring mobile communication support. However, in complex environmental conditions, especially in tunnels, the unique spatial structure of tunnels causes signal multipath attenuation.

[0003] In tunnel environments, conventional signal attenuation is usually caused by natural factors such as temporary obstacles, leading to a weakening of signal strength. Because this attenuation is a natural and relatively stable process, communication systems can easily identify abnormal intrusions and take appropriate measures to maintain communication stability, thus reducing the risk of intrusion. Multipath attenuation, however, occurs when a signal travels along different paths to the receiver, causing interference between these paths and resulting in drastic fluctuations in signal strength, creating communication dead zones. If it's impossible to accurately determine whether the communication dead zone is caused by conventional signal attenuation or multipath attenuation, smart tablets may trigger anti-intrusion mode even under conventional signal attenuation conditions, causing unnecessary communication interruptions and reducing communication security and stability.

[0004] Therefore, a smart tablet anti-intrusion system for field communication is proposed, which can accurately identify the causes of communication blind spots and perform anti-intrusion operations, effectively ensuring the communication security and stability of smart tablets for field communication in complex tunnel environments. Summary of the Invention

[0005] (1) Technical problems to be solved

[0006] This invention provides an anti-intrusion system for smart tablets used in outdoor communication, which solves the problem that smart tablets used in outdoor communication cannot accurately identify communication blind spots in complex tunnel environments, resulting in inaccurate anti-intrusion triggering and reduced communication security and stability.

[0007] (2) Technical solution

[0008] To achieve the above objectives, the present invention provides an anti-intrusion system for a smart tablet computer used for field communication. The system includes: a multipath feature parameter extraction module, a multipath statistical feature calculation and compensation module, a blind zone matching and judgment module, and an anti-intrusion response and log management module, with each module being connected to the other in sequence.

[0009] The multipath feature parameter extraction module is used to periodically transmit detection signals in a tunnel environment via a smart tablet device, receive and collect response signals from legitimate communication nodes, and extract multipath feature parameters from the response signals. The multipath feature parameters include signal arrival time difference, signal arrival angle, signal strength distribution, and channel impulse response.

[0010] The multipath statistical feature calculation and compensation module is used to calculate the multipath statistical features of the current channel based on the multipath feature parameters. The multipath statistical features include the number of multipath components, the time delay spread, the angle spread, and the signal strength fluctuation variance. The multipath statistical features are then compensated by performing motion compensation based on the real-time moving speed and inertial measurement data of the smart tablet device to obtain the compensated multipath statistical features.

[0011] The blind zone matching and determination module is used to perform matching analysis between the compensated multipath statistical features and the tunnel multipath blind zone fingerprint database to obtain a comprehensive matching degree. When the comprehensive matching degree exceeds a preset matching threshold, a matching success message is obtained. The tunnel multipath blind zone fingerprint database contains blind zone channel patterns caused by signal multipath effects in the tunnel environment.

[0012] The intrusion prevention response and log management module is used to determine that the smart tablet device is in a communication blind zone and initiate a hierarchical intrusion prevention response when a successful matching information is detected; when the multipath statistical characteristics are detected to recover to the non-blind zone channel mode and a secure connection is re-established with the legitimate communication node, the intrusion prevention response is exited and the blind zone event log is uploaded.

[0013] Furthermore, the multipath statistical feature calculation and compensation module includes:

[0014] An inertial measurement data acquisition module is used to continuously acquire inertial measurement data of the smart tablet device through the inertial measurement unit built into the smart tablet device. The inertial measurement data includes acceleration, angular velocity and azimuth data.

[0015] The motion trajectory and attitude calculation module is used to calculate the three-dimensional motion trajectory and real-time attitude changes of the smart tablet device in the tunnel space by using the inertial measurement data and real-time moving speed of the smart tablet device through a sensor fusion algorithm.

[0016] The multipath compensation module is used to establish a mapping relationship model between the movement state of the smart tablet device and the changes in multipath statistical characteristics based on the three-dimensional motion trajectory and real-time attitude changes; and to perform dynamic inverse compensation calculations on the number of multipath components, time delay spread, angle spread, and signal strength fluctuation variance according to the mapping relationship model to obtain the compensated multipath statistical characteristics.

[0017] Furthermore, the motion trajectory and attitude calculation module includes:

[0018] An inertial measurement data correction module is used to obtain corrected inertial measurement data from the inertial measurement data of the smart tablet device through data preprocessing.

[0019] The Kalman filter pose optimization module is used to input the corrected inertial measurement data and real-time moving speed into the Kalman filter for multi-source data fusion. The Kalman filter calculates the prior estimate of the pose state of the smart tablet device through the prediction stage. In the update stage, the real-time moving speed is used as the observation input and combined with the observation model to correct the prior estimate to obtain the optimized posterior estimate of the pose state.

[0020] The continuous three-dimensional motion trajectory construction module is used to iteratively execute the prediction stage and the update stage, continuously calculate the three-dimensional position coordinates, motion velocity vector and real-time attitude change in Euler angles of the smart tablet device in the tunnel coordinate system; and construct the continuous three-dimensional motion trajectory of the smart tablet device in the tunnel space based on the continuously calculated pose state sequence.

[0021] Furthermore, the inertial measurement data correction module includes:

[0022] The dual-level filtering module is used to perform sliding window mean filtering on the inertial measurement data of the smart tablet device to suppress high-frequency noise components, and then use wavelet transform filtering to separate and eliminate residual random noise to obtain the pre-filtered inertial measurement data.

[0023] The zero-bias error correction module is used to calculate the zero-bias estimate of each inertial measurement channel at the current time by using the zero-bias error compensation model to calculate the inertial measurement data after the initial filtering. The zero-bias error compensation model is established based on the statistical characteristics of inertial measurement data of the smart tablet device operating in the tunnel environment for a long time. The corrected inertial measurement data is obtained by subtracting the zero-bias estimate of the corresponding inertial measurement channel from the initial filtered inertial measurement data.

[0024] Furthermore, the zero-bias error correction module includes the following steps:

[0025] Historical inertial measurement data output by the inertial measurement unit of the intelligent flat panel device under typical tunnel operating conditions are collected. The historical inertial measurement data is preprocessed to remove outliers and aligned. Then, the zero-bias error statistical characteristics of each inertial measurement channel are extracted. The zero-bias error statistical characteristics include mean, variance and drift characteristics over time.

[0026] A mathematical model of the zero bias error is established based on the statistical characteristics of the zero bias error. The mathematical model of the zero bias error includes the reference zero bias value and adaptive adjustment coefficient of each inertial measurement channel. The mathematical model of the zero bias error is trained and optimized by a machine learning algorithm to obtain a zero bias error compensation model.

[0027] Furthermore, the Kalman filter pose optimization module includes the following steps:

[0028] The real-time moving speed is converted into a velocity observation vector in the tunnel coordinate system. The predicted observation value corresponding to the prior estimate is calculated according to the observation model, and the observation residual between the predicted observation value and the velocity observation vector is obtained. The observation residual is weighted and adjusted according to the Kalman filter gain matrix to obtain the state correction amount. The state correction amount is added to the prior estimate and fused to obtain the optimized posterior estimate of the pose state.

[0029] Furthermore, the multipath compensation module includes the following steps:

[0030] Based on the mapping relationship model, the nonlinear relationship between the movement state of the smart tablet device and the changes in various multipath statistical features is analyzed. The expected distortion of multipath features caused by the movement of the smart tablet device is calculated in real time based on the nonlinear relationship using the three-dimensional motion trajectory and real-time attitude change data of the smart tablet device.

[0031] For the number of multipath components, delay spread, angle spread, and signal strength fluctuation variance, corresponding inverse compensation functions are constructed. The expected distortion of the multipath features is then processed by the inverse compensation function to perform inverse calculations on the measured multipath statistical features, resulting in compensated multipath statistical features that are independent of the movement of the smart tablet device and only reflect the inherent characteristics of the tunnel space channel.

[0032] Furthermore, the blind spot matching and determination module includes the following steps:

[0033] Extract the reference range of multipath statistical features corresponding to each blind zone channel mode from the tunnel multipath blind zone fingerprint database, and calculate the feature matching similarity between the number of multipath components, delay spread, angle spread, and signal strength fluctuation variance in the compensated multipath statistical features and the corresponding feature reference range of each blind zone channel mode.

[0034] The similarity scores of each feature are weighted and fused to obtain a comprehensive matching score. When the comprehensive matching score exceeds a preset matching threshold, a matching success message indicating a successful match with a specific blind zone channel mode is generated, and the blind zone confidence score and geographical location information corresponding to the blind zone channel mode are output.

[0035] Furthermore, the intrusion prevention response and log management module includes the following steps:

[0036] The blind zone confidence level and the matched blind zone channel mode in the successful matching information are analyzed, and the intrusion risk level is determined according to the level of the blind zone confidence level. The corresponding intrusion prevention response level is selected from the predefined response strategy library according to the intrusion risk level. The predefined response strategy library contains corresponding communication security strategies under different intrusion risk levels.

[0037] Perform intrusion prevention operations at the selected intrusion prevention response level. These operations include adjusting signal transmission power, limiting communication frequency bands, and enabling enhanced authentication mechanisms. Simultaneously, record the timestamp, geographical location, and matching blind zone pattern information of the current blind zone event, generate a blind zone event log, and store it locally.

[0038] (3) Beneficial effects

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] 1. By periodically transmitting detection signals through a smart tablet device, the response signals of legitimate communication nodes are collected, and multipath feature parameters are extracted to calculate multipath statistical characteristics. Motion compensation is performed by combining the real-time movement speed of the smart tablet device with inertial measurement unit data, making the multipath statistical characteristics more accurately reflect the inherent characteristics of the tunnel spatial channel. Matching the compensated multipath statistical characteristics with a tunnel multipath blind zone fingerprint database allows for precise determination of whether the smart tablet device is in a communication blind zone, effectively distinguishing between multipath attenuation and conventional signal attenuation. This solves the problem of inaccurately identifying the causes of communication blind zones and provides a reliable basis for subsequent intrusion prevention measures.

[0041] 2. When a successful match is detected, and the smart tablet device is determined to be in a communication blind zone, the intrusion risk level can be determined based on the blind zone confidence level and the matched blind zone channel mode type. The corresponding intrusion prevention response level can then be selected from a predefined response strategy library. This tiered intrusion prevention response mechanism can adopt corresponding communication security strategies based on different risk situations. It can ensure communication security while minimizing the impact on normal communication, effectively resisting attacks from intruders in communication blind zones, and enhancing the security of smart tablets used for communication in the field.

[0042] 3. While initiating the intrusion prevention response, the system records the timestamp, geographical location, and matching blind zone pattern information of the current blind zone event, generates a blind zone event log, and stores it locally. The log is uploaded after exiting the intrusion prevention response, which helps to further improve the intrusion prevention strategy and continuously enhance the security and reliability of the smart tablet for outdoor communication in complex environments. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the module composition of an intelligent tablet anti-intrusion system for field communication according to Embodiment 1 of the present invention. Detailed Implementation

[0044] 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.

[0045] Before providing examples, it is necessary to explain the application scenarios of the present invention. The present invention is an anti-intrusion system for smart tablets used for field communication. Because smart tablets used for field communication cannot accurately identify communication blind spots in complex tunnel environments, they trigger anti-intrusion operations, which reduces the stability of communication security and lowers the user experience.

[0046] Example 1: As Figure 1 As shown, this embodiment provides an anti-intrusion system for smart tablets used in field communication. The system includes: a multipath feature parameter extraction module, a multipath statistical feature calculation and compensation module, a blind zone matching and judgment module, and an anti-intrusion response and log management module. Each module is connected to the other in sequence.

[0047] The multipath feature parameter extraction module is used to periodically transmit detection signals in a tunnel environment via a smart tablet device, receive and collect response signals from legitimate communication nodes, and extract multipath feature parameters from the response signals. These multipath feature parameters include signal arrival time difference, signal arrival angle, signal strength distribution, and channel impulse response. The smart tablet device actively transmits standardized radio frequency detection signals into the tunnel space at fixed time intervals (e.g., every 2-5 seconds). These radio frequency detection signals use preset frequency and power parameters to ensure good propagation characteristics in the tunnel environment. When legitimate communication nodes are present, they respond to the detection signals according to predetermined communication protocol specifications, returning a response signal containing node authentication information and channel state data. Upon receiving the response signal, the smart tablet device immediately initiates the multipath feature parameter extraction algorithm. Due to the special geometry of the tunnel, the signal undergoes multiple reflections and refractions with surfaces such as the tunnel walls, tunnel ceiling, and ground during propagation, resulting in signal components with different propagation paths arriving at the receiving end simultaneously. By analyzing the arrival characteristics of signals from different paths, smart flat panel devices extract four key multipath characteristic parameters: signal arrival time difference reflects the time delay difference between different propagation paths, measured by a high-precision clock; signal arrival angle describes the spatial incident direction of each path signal relative to the device's antenna array, calculated using the phase difference information of the antenna array; signal strength distribution characterizes the power attenuation of signals from different paths, measured in real time by a power detection circuit; and channel impulse response comprehensively reflects the frequency and time domain characteristics of the entire propagation channel, reconstructing the complete channel transmission function from the received signal using signal processing algorithms.

[0048] The multipath statistical feature calculation and compensation module is used to calculate the multipath statistical features of the current channel based on the multipath feature parameters. These multipath statistical features include the number of multipath components, delay spread, angle spread, and signal strength fluctuation variance. The module performs motion compensation on the multipath statistical features based on the real-time moving speed and inertial measurement data of the smart tablet device to obtain compensated multipath statistical features. The number of multipath components is obtained by using a peak detection algorithm on the channel impulse response to identify the total number of valid signal paths exceeding a preset threshold. The delay spread is calculated using a root mean square delay spread algorithm to quantify the dispersion of the multipath signal in the time dimension. The angle spread is obtained through angle domain power spectrum analysis, reflecting the signal's diffusion range in spatial angles. The signal strength fluctuation variance is calculated using statistical analysis methods to determine the degree of signal strength variation within a time window. Considering that the movement of the smart tablet device in the tunnel will significantly affect the multipath statistical features, a motion compensation mechanism is introduced.

[0049] The blind zone matching and determination module is used to perform matching analysis between the compensated multipath statistical features and the tunnel multipath blind zone fingerprint database to obtain a comprehensive matching degree. When the comprehensive matching degree exceeds a preset matching threshold, a matching success message is obtained. The tunnel multipath blind zone fingerprint database contains blind zone channel patterns caused by signal multipath effects in the tunnel environment. The database contains typical blind zone channel patterns in tunnels, which are established through systematic channel measurement and analysis of tunnels. Each blind zone channel pattern corresponds to the multipath propagation characteristics generated by specific tunnel geometry and reflection environment. The tunnel multipath blind zone fingerprint database stores the reference value range of multipath statistical features for various blind zone channel patterns and the corresponding geographical location information.

[0050] The intrusion prevention response and log management module determines that the smart tablet device is in a communication blind zone and initiates a tiered intrusion prevention response when a successful match is detected. When the multipath statistical characteristics recover to a non-blind zone channel mode and a secure connection is re-established with a legitimate communication node, the intrusion prevention response is exited and the blind zone event log is uploaded. Once the smart tablet device is confirmed to be in a communication blind zone, the system immediately initiates a tiered intrusion prevention response mechanism. The system synchronously records detailed information about the blind zone event, including the precise timestamp of the event, geographical coordinates, and the type of blind zone mode matched, generating a structured blind zone event log and temporarily storing it in the device's local storage. When the device detects that the multipath statistical characteristics have recovered to a normal non-blind zone channel mode and a securely verified stable connection has been re-established with a legitimate communication node, the intrusion prevention response mechanism is automatically deactivated. Simultaneously, the stored blind zone event log is uploaded to the backend management system, providing data support for communication security analysis and strategy optimization. The upload process employs security measures such as encrypted transmission and integrity verification to prevent the blind zone event log from being tampered with or intercepted during transmission.

[0051] The multipath statistical feature calculation and compensation module includes:

[0052] An inertial measurement data acquisition module is used to continuously acquire inertial measurement data from the smart tablet device via its built-in inertial measurement unit. This inertial measurement data includes acceleration, angular velocity, and azimuth data. The built-in inertial measurement unit includes sensor components such as a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer. These sensors continuously acquire motion state data from the smart tablet device at a high frequency (typically 100-200Hz). The three-axis accelerometer measures the linear acceleration components of the device in three orthogonal directions in space; the three-axis gyroscope detects the angular velocity changes of the device around the three coordinate axes; and the magnetometer provides the azimuth information of the device relative to the Earth's magnetic north pole. This raw inertial measurement data constitutes the basic data source for sensing the motion state of the smart tablet device.

[0053] The motion trajectory and attitude calculation module is used to calculate the three-dimensional motion trajectory and real-time attitude change of the smart tablet device in the tunnel space by using the inertial measurement data and real-time moving speed of the smart tablet device through a sensor fusion algorithm.

[0054] The multipath compensation module is used to establish a mapping model between the movement state of the smart tablet device and the changes in multipath statistical characteristics based on the three-dimensional motion trajectory and real-time attitude changes. According to the mapping model, dynamic inverse compensation calculations are performed on the number of multipath components, time delay spread, angle spread, and signal strength fluctuation variance to obtain the compensated multipath statistical characteristics. The mapping model establishes a mathematical relationship between the movement state of the smart tablet device and the changes in multipath statistical characteristics through extensive experimental data and theoretical analysis. For example, when the smart tablet device moves along the tunnel axis, it changes the relative distance and incident angle with the tunnel wall, thus affecting the arrival time and intensity distribution of the reflected signal; when the device rotates, the antenna directivity causes changes in the angle spread characteristics of the received signal; changes in the device's moving speed produce Doppler frequency shift, affecting the measurement accuracy of signal strength.

[0055] The motion trajectory and attitude calculation module includes:

[0056] An inertial measurement data correction module is used to obtain corrected inertial measurement data from the inertial measurement data of the smart tablet device through data preprocessing.

[0057] The Kalman filter pose optimization module is used to input the corrected inertial measurement data and real-time movement velocity into a Kalman filter for multi-source data fusion. Through the prediction phase of the Kalman filter, a prior estimate of the pose state of the smart tablet device is calculated. In the update phase, the real-time movement velocity is used as the observation input, and the prior estimate is corrected in conjunction with the observation model to obtain the optimized posterior estimate of the pose state. The state equation of the Kalman filter describes the dynamic evolution process of the smart tablet device's motion state. The prediction phase uses the state estimate from the previous moment and the state transition model to predict the prior estimate of the pose state at the current moment.

[0058] A continuous 3D motion trajectory construction module is used to iteratively execute the prediction and update phases, continuously calculating the 3D position coordinates, motion velocity vector, and real-time attitude changes in Euler angles of the smart tablet device in the tunnel coordinate system. Based on the continuously calculated pose state sequence, a continuous 3D motion trajectory of the smart tablet device in the tunnel space is constructed. The iterative execution process ensures that the filter can continuously track the changes in the device's motion state. A complete prediction-update loop is executed in each sampling period, forming a recursive state estimation process. Over time, the filter accumulates more observation information, and the accuracy of the state estimation gradually improves. The final output 3D position coordinates are represented in the tunnel coordinate system, which facilitates correlation analysis with the tunnel's geometry; the motion velocity vector includes the velocity components of the device in the three coordinate axes; the attitude information in Euler angle form intuitively describes the spatial orientation of the device. The construction of the continuous 3D motion trajectory is achieved by connecting discrete pose state estimates in chronological order. The system uses an interpolation algorithm to smooth the discrete trajectory points, generating a continuous and smooth trajectory curve. This trajectory not only reflects the spatial movement path of the device, but also contains kinematic information such as velocity and acceleration, providing a complete description of the motion state for subsequent motion compensation calculations.

[0059] The inertial measurement data correction module includes:

[0060] A two-level filtering module is used to perform sliding window mean filtering on the inertial measurement data of the smart tablet device to suppress high-frequency noise components, followed by wavelet transform filtering to separate and eliminate residual random noise, resulting in pre-filtered inertial measurement data. The quality of the raw output data from the inertial sensor directly affects the performance of the entire intrusion prevention system; therefore, a multi-level data processing method is required. Sliding window mean filtering is a simple and effective time-domain filtering method. The system maintains a fixed-length data buffer window (usually containing 10-20 sampling points), and performs an arithmetic average of the data within the window to obtain the filtered output at the current moment, effectively suppressing high-frequency random noise. The selection of the window length requires a trade-off between noise suppression effect and response speed, and the optimal parameters are usually determined based on the motion characteristics of the smart tablet device and the sensor noise level in the tunnel environment. Wavelet transform filtering is a more advanced signal processing method that can analyze signals simultaneously in the time and frequency domains. Wavelet transform decomposes the signal into wavelet coefficients of different frequency scales, removes high-frequency coefficients corresponding to noise through thresholding, and then reconstructs the denoised signal. Compared to traditional frequency domain filtering methods, wavelet transform can better preserve the local characteristics of a signal, making it particularly suitable for processing non-stationary inertial measurement data. The system selects appropriate wavelet basis functions (such as Daubechies wavelets or bioorthogonal wavelets) and decomposition levels, and removes noise components using soft-thresholding or hard-thresholding algorithms.

[0061] The zero-bias error correction module is used to calculate the zero-bias estimate of each inertial measurement channel at the current moment by using the preliminary filtered inertial measurement data through a zero-bias error compensation model. This model is established based on the statistical characteristics of inertial measurement data from a smart tablet device operating for an extended period in a tunnel environment. The corrected inertial measurement data is obtained by subtracting the corresponding zero-bias estimate from the preliminary filtered inertial measurement data. Zero-bias error is a common systematic deviation in inertial sensors, manifested as a non-zero output even when the sensor is stationary. Causes of zero-bias error include limitations in sensor manufacturing processes, the influence of temperature changes, and device aging. In a tunnel environment, changes in temperature and humidity exacerbate the time-varying characteristics of zero-bias error, thus requiring a dynamic error compensation model. The calculation of the zero-bias estimate uses a combination of statistical methods and adaptive algorithms. The system first identifies time periods when the smart tablet device is relatively stationary, during which the theoretical sensor output should be zero (for accelerometers, the influence of gravity needs to be considered). By statistically analyzing the sensor output during these stationary periods, the current zero-bias value can be estimated. Simultaneously, the system employs adaptive algorithms such as recursive least squares or Kalman filtering to dynamically update the zero-bias estimate based on the real-time sensor output, adapting to the time-varying characteristics of the zero bias. The corrected inertial measurement data is obtained by subtracting the corresponding zero-bias estimate from the initially filtered inertial measurement data. This correction process needs to be performed separately for each sensor channel, including the three channels of the triaxial accelerometer and the three channels of the triaxial gyroscope. The corrected inertial measurement data can more accurately reflect the true motion state of the smart tablet device, providing reliable input data for subsequent sensor fusion and trajectory calculation.

[0062] The zero-bias error correction module includes the following steps:

[0063] Historical inertial measurement data (IGM) output from the intelligent flat panel device under typical tunnel operating conditions is collected. This historical IGM data is preprocessed to remove outliers and aligned before extracting the zero-bias error statistical characteristics of each IGM channel. These zero-bias error statistical characteristics include mean, variance, and drift characteristics over time. The accuracy of the zero-bias error compensation model directly affects the accuracy of the entire motion estimation system; therefore, scientific modeling methods and advanced machine learning techniques are required. Collecting historical IGM data is fundamental to establishing the compensation model. The system needs to undergo long-term operational testing of the intelligent flat panel device in a tunnel environment, collecting sensor output data under different operating conditions. Typical operating states include stationary, uniform linear motion, variable speed motion, rotational motion, and composite motion states. The data collection process needs to cover different environmental conditions, such as different temperatures, humidity levels, and vibration levels, to ensure the robustness and adaptability of the compensation model. Data preprocessing is a crucial step in ensuring modeling quality. Outlier detection uses statistical methods to identify and remove data points that significantly deviate from the normal range. These outliers may be caused by sensor malfunctions, external impacts, or electromagnetic interference. Data alignment requires addressing time synchronization issues between different sensors to ensure consistency of multi-channel data across time bases. Preprocessing also includes standardization operations such as data format unification, unit conversion, and coordinate system transformation. The mean of the zero-bias error statistical characteristics reflects the average level of the zero-bias, obtained by statistically averaging the sensor outputs in a static state. Variance describes the degree of fluctuation in the zero-bias, reflecting its stability. The drift characteristics over time are obtained through trend analysis of long-term observation data, including different drift patterns such as linear drift and exponential drift. These statistical characteristics provide a quantitative basis for subsequent mathematical modeling.

[0064] A mathematical model for zero-bias error is established based on the statistical characteristics of the zero-bias error. This model includes the baseline zero-bias value and adaptive adjustment coefficient for each inertial measurement channel. The zero-bias error compensation model is obtained by training and optimizing this mathematical model using a machine learning algorithm. The zero-bias error mathematical model is established using a parametric modeling method. The basic form of the model is: Zero-bias estimate = Baseline zero-bias value + Adaptive adjustment term. The baseline zero-bias value represents the zero-bias level of the sensor under standard conditions and is a static parameter of the model. The adaptive adjustment term considers the influence of environmental and time factors on the zero-bias, including dynamic parameters such as temperature compensation and time drift. The adaptive adjustment coefficient is determined through regression analysis of historical data, establishing a mathematical relationship between zero-bias changes and influencing factors. The application of machine learning algorithms further improves the accuracy and adaptability of the compensation model. The system uses machine learning methods such as support vector machines, neural networks, or random forests, taking environmental parameters and time information from historical data as input features and the corresponding zero-bias error as the output label, to obtain a nonlinear zero-bias error compensation model through training. The training process includes standard machine learning procedures such as feature selection, model parameter optimization, and cross-validation. The optimized zero-bias error compensation model can predict the zero-bias error of each sensor channel in real time based on the current environmental conditions and operating status, achieving dynamic error compensation.

[0065] The Kalman filter pose optimization module includes the following steps:

[0066] The real-time moving speed is converted into a velocity observation vector in the tunnel coordinate system. Based on the observation model, the predicted observation value corresponding to the prior estimate is calculated, and the observation residual between the predicted and predicted values ​​is obtained. The observation residual is then weighted and adjusted according to the Kalman filter gain matrix to obtain a state correction. This state correction is then added to the prior estimate to obtain the optimized posterior estimate of the pose state. The accurate implementation of the update phase directly determines the accuracy and stability of the motion state estimation. Acquiring real-time moving speed data typically relies on a GPS positioning system or its positioning equipment. The GPS receiver provides the device's velocity information in the geographic coordinate system, including three components: eastward velocity, northward velocity, and vertical velocity. However, in a tunnel environment, GPS signals may be obstructed or subject to multipath interference. Therefore, the system also incorporates other auxiliary positioning methods, such as inertial navigation estimation and odometer measurement. The acquired velocity data needs to undergo coordinate transformation from the geographic coordinate system to the tunnel coordinate system to maintain consistency with the state variables of the Kalman filter. The construction of the velocity observation vector requires consideration of technical details such as coordinate system transformation and unit unification. Geographic coordinate systems typically use latitude and longitude, while tunnel coordinate systems are rectangular coordinate systems based on the tunnel's geometry. The coordinate transformation matrix is ​​determined based on the tunnel's geographical location and tilt angle, converting the velocity vector in the geographic coordinate system to its representation in the tunnel coordinate system through matrix operations. The transformation process also needs to consider the influence of geophysical factors such as Earth's curvature and magnetic declination to ensure the transformation accuracy meets system requirements. The observation model is the mathematical bridge connecting state variables and observations, establishing a linear relationship between the velocity components in the Kalman filter's state vector and external velocity observations. The observation matrix H is a selection matrix used to extract the state components corresponding to the observations from the complete state vector. For velocity observations, the observation matrix selects three velocity components from the state vector, ignoring other state components such as position, acceleration, and attitude. The calculation of predicted observations involves mapping the prior estimate of the current state to the observation space through the observation model. Specifically, multiplying the prior estimate of the state vector by the observation matrix yields the predicted velocity value based on the current state estimate. This predicted observation represents the expected observation value derived from the motion model, reflecting the filter's expectation of the observations at the current moment. Observation residuals are a crucial indicator of the accuracy of a prediction model. They are the difference between the actually observed velocity vector and the predicted observed value calculated from prior estimates. The magnitude of the observation residuals reflects the degree of agreement between the motion model and actual observations. Smaller observation residuals indicate that the motion model accurately describes the motion of the smart tablet device; larger observation residuals may indicate modeling errors, external disturbances, or observation noise. The observation residuals also contain valuable information for correcting state estimates and are essential for Kalman filter state updates. The calculation of the Kalman filter gain matrix is ​​the core step in the update process.The gain matrix K is calculated as follows: K = P⁻H^T(HP⁻H^T + R)⁻¹, where P⁻ is the prior estimate of the state covariance, H is the observation matrix, and R is the observation noise covariance matrix. The gain matrix essentially reflects the trade-off between the reliability of prediction and observation information. When observation noise is low, the weight of observation information increases; when process noise is low, the weight of prediction information increases. The state correction is calculated using a linear weighted method, representing the magnitude and direction of the adjustment to the state estimate based on the current observation information. Each component of the state correction corresponds to an element of the state vector, including position correction, velocity correction, acceleration correction, and attitude correction. The final posterior estimate is obtained by adding the prior estimate to the state correction: Posterior estimate = Prior estimate + State correction. The posterior estimate represents the optimal estimate of the device state at the current moment.

[0067] The multipath compensation module includes the following steps:

[0068] The nonlinear relationship between the movement state of the smart tablet device and the changes in various multipath statistical characteristics is analyzed based on the mapping relationship model. The expected distortion of multipath characteristics caused by the movement of the smart tablet device is calculated in real time using the three-dimensional motion trajectory and real-time attitude change data of the smart tablet device according to the nonlinear relationship. Inverse compensation aims to eliminate the impact of device movement on the measurement of multipath statistical characteristics and restore the true characteristics of the tunnel channel. For the analysis of the impact of the number of multipath components, it is necessary to consider how the movement of the smart tablet device changes the number of distinguishable propagation paths. When the smart tablet device moves along the tunnel axis, some previously blocked reflection paths may become visible, while some previously visible paths may be blocked by new obstacles. The lateral movement of the smart tablet device changes the distance relationship with the tunnel wall, affecting the strength and detectability of the sidewall reflection paths. The rotational movement of the smart tablet device changes the antenna directivity, affecting the receiving sensitivity of signals in different directions. The delay spread value is mainly affected by the movement of the smart tablet device in terms of the change in the propagation path length. The movement of the smart tablet device causes different degrees of change in the length of the direct path and each reflection path, thereby changing the relative delay of the signals on each path. Changes in movement speed also cause Doppler frequency shift, further affecting the accuracy of time delay measurements. The system needs to accurately calculate the changes in the length of each propagation path based on the three-dimensional motion trajectory of the smart tablet device and the geometry of the tunnel. The analysis of the impact of movement on the angular spread value is more complex, requiring consideration of the influence of changes in the smart tablet device's attitude on the antenna receiving characteristics. Smart tablet devices are typically equipped with multi-antenna systems or antenna arrays, and the directivity of the antennas affects the receiving gain of signals at different angles. When the smart tablet device rotates, the pointing of the antenna array changes, causing a shift in the measurement results of the angular power spectrum. In addition, the movement of the smart tablet device also changes its relative positional relationship with the tunnel geometry, affecting the signal incident angle distribution. The mechanism by which the signal strength fluctuation variance is affected by movement is the most complex, involving the coupling of multiple fading mechanisms. Fast fading caused by device movement is mainly due to the interference effect of multipath signals; movement causes a rapid change in the phase relationship of signals along each path, resulting in rapid fluctuations in the received signal strength. Slow fading is mainly caused by the shadowing effect; obstacles encountered during the movement of the smart tablet device cause slow changes in the signal. Furthermore, the movement of the smart tablet device also causes changes in the receiving distance, leading to changes in free space path loss. The calculation of the expected distortion of multipath characteristics requires quantitative modeling of the various influencing factors mentioned above. Based on the real-time three-dimensional motion trajectory and attitude change data of the smart tablet device, combined with the geometric model of the tunnel and electromagnetic propagation theory, the system calculates the expected impact of various movement effects on the multipath statistical characteristics, i.e., the expected distortion of multipath characteristics.

[0069] For each of the multipath component quantity, delay spread, angle spread, and signal strength fluctuation variance, a corresponding inverse compensation function is constructed. The expected distortion of the multipath characteristics is then processed through the inverse compensation function to perform inverse operations on the measured multipath statistical characteristics, resulting in compensated multipath statistical characteristics that are independent of the smart tablet device's movement and only reflect the inherent characteristics of the tunnel space channel. For each multipath statistical characteristic parameter, the system needs to construct a corresponding inverse compensation function. The basic idea of ​​these inverse compensation functions is to establish an inverse mapping relationship between the movement distortion and the compensation amount. For the multipath component quantity, the inverse compensation function needs to determine which detected multipath components are spurious paths caused by movement and which multipath components masked by movement need to be recovered. For the delay spread, the inverse compensation function performs a geometric correction to the measured delay based on the path length change. For the angle spread, the inverse compensation function needs to compensate for the effects of antenna directivity changes and angle measurement offsets. For the signal strength fluctuation variance, the inverse compensation function needs to separate and eliminate various fading components caused by movement. The implementation of the inverse operation is a key step in transforming the theoretical model into a practical algorithm. The system inputs the measured multipath statistical features and the calculated motion distortion into the corresponding inverse compensation function, and obtains the compensated feature values ​​through numerical calculation. The acquisition of the compensated multipath statistical features marks the completion of the motion compensation process. These compensated multipath statistical features accurately reflect the inherent propagation characteristics of the tunnel spatial channel and are independent of the movement state of the smart tablet device. Only such multipath statistical features can be accurately matched with the tunnel multipath blind zone fingerprint database, achieving reliable communication blind zone detection.

[0070] The blind zone matching and determination module includes the following steps:

[0071] The multipath statistical feature reference ranges corresponding to each blind zone channel mode are extracted from the tunnel multipath blind zone fingerprint database. The feature matching similarity between the compensated multipath statistical features (number of multipath components, delay spread, angle spread, and signal strength fluctuation variance) and the corresponding feature reference ranges of each blind zone channel mode is calculated. The data structure design of the tunnel multipath blind zone fingerprint database needs to fully consider the efficiency and accuracy requirements of the matching algorithm. Each blind zone channel mode in the tunnel multipath blind zone fingerprint database contains a complete multipath statistical feature reference range. These reference ranges are not simple numerical intervals, but probability distribution models obtained based on statistical analysis of a large amount of measurement data. For the number of multipath components, the reference range may be represented as a discrete probability distribution, describing the probability of occurrence of different numbers of multipath components. For continuous variables such as delay spread, angle spread, and signal strength fluctuation variance, the reference range is usually described using a normal distribution, log-normal distribution, or a more complex probability density function. The calculation of feature matching similarity uses probabilistic and statistical methods. For each multipath statistical feature parameter, the system calculates the probability or likelihood that the compensated measured value falls within the reference range of each blind zone mode. The specific calculation method depends on the probability distribution type of the reference range. For a normally distributed reference range, similarity can be obtained by calculating the standardized distance between the measured values; the smaller the distance, the higher the similarity. For a non-parametric distribution reference range, similarity can be calculated using methods such as kernel density estimation or histogram interpolation. The matching similarity calculation for the number of multipath components compares the measured number of multipath components with the corresponding probability distribution in the blind zone mode, directly obtaining the probability of that value as the similarity. If the measured value is not within the support interval of the reference distribution, the similarity is set to zero or a very small value. The matching similarity calculation for delay spread, angle spread, and signal strength fluctuation variance needs to consider the characteristics of continuous variables. The system uses the function value of the probability density function at the measured value as the similarity index, or the probability that the measured value falls within the confidence interval of the reference range as the similarity. To improve the robustness of the algorithm, the system also considers the impact of measurement errors, incorporating measurement uncertainty into the similarity calculation through convolution operations.

[0072] The overall matching degree is obtained by weighted fusion of the similarity scores of each feature. When the overall matching degree exceeds a preset matching threshold, a successful matching message indicating a successful match with a specific blind zone channel pattern is generated, and the blind zone confidence score and geographical location information corresponding to the blind zone channel pattern are output. Different multipath statistical features may have different importance in blind zone detection, so appropriate weight coefficients need to be set. The weights can be determined based on factors such as the information content, stability, and discriminative power of the features. Generally speaking, in tunnel environments, delay spread and angle spread values ​​usually have strong environmental specificity, so their weights are relatively high; while signal strength fluctuation variance is easily affected by external interference, so its weight may be relatively low. The formula for calculating the overall matching degree is: Overall matching degree = Σ(wi × Si), where wi is the weight coefficient of the i-th feature, and Si is the similarity of the i-th feature. The weight coefficients need to satisfy the normalization condition, that is, the sum of all weights equals 1. To avoid the extreme values ​​of a certain feature having too much influence on the overall matching degree, the system may also use other fusion methods such as geometric mean or harmonic mean. Setting a preset matching threshold requires balancing detection accuracy and false alarm rate. A threshold set too high will lead to missed detections, meaning the true blind zone location cannot be identified; a threshold set too low will lead to false alarms, misclassifying non-blind zone locations as blind zones. The optimal threshold value is usually determined through extensive experimental data and statistical analysis, using methods such as ROC curve analysis to find the best operating point. A successful match is determined when the overall matching degree exceeds the preset matching threshold. Once this condition is met, the system considers the currently measured multipath statistical characteristics to match a specific blind zone channel pattern. Blind zone confidence is a quantitative assessment of the reliability of successful matching information. Confidence can be calculated based on the overall matching degree; a higher matching degree indicates higher confidence. The system can also consider other factors, such as the number of features involved in the matching and the consistency of historical matching records. Confidence is usually expressed as a value between 0 and 1, or converted to a percentage. Geographical location information is an important attribute of each blind zone pattern in the blind zone fingerprint database, containing the precise location coordinates, coverage area, and geometric features of the blind zone within the tunnel. This geographic information helps validate the validity of successful matches and provides spatial context for subsequent intrusion prevention responses.

[0073] The intrusion prevention response and log management module includes the following steps:

[0074] The system analyzes the blind zone confidence level and the matched blind zone channel pattern in the successful matching information, and determines the intrusion risk level based on the blind zone confidence level. It then selects the corresponding intrusion prevention response level from a predefined response strategy library, which contains communication security strategies for different intrusion risk levels. The system extracts the blind zone confidence level value from the successful matching information and classifies the confidence level into different levels according to a preset threshold range. For example, a confidence level above 0.9 is defined as high confidence, indicating that the blind zone detection result is very reliable; a confidence level between 0.7 and 0.9 is defined as medium confidence, indicating that the detection result is basically reliable but has some uncertainty; and a confidence level between 0.5 and 0.7 is defined as low confidence, indicating that the reliability of the detection result is relatively low. The analysis of blind zone channel pattern types provides detailed information on blind zone characteristics. Different types of blind zones may have different risk characteristics, requiring corresponding protection strategies. For example, blind zones caused by tunnel curves typically exhibit strong spatial locality, resulting in a relatively low intrusion risk; while blind zones caused by tunnel bifurcation or complex geometry may have multiple concealed paths, leading to a higher intrusion risk. The system assigns a corresponding intrusion risk level to each blind zone channel pattern type based on historical statistical data and expert knowledge. In addition to blind zone confidence and channel pattern type, the system also considers contextual information such as current time, location, communication traffic, and historical intrusion records. Risk levels are typically categorized into four levels: low, medium, high, and extremely high. In low-risk situations, the system may only initiate basic monitoring measures; in medium-risk situations, it will enhance authentication and signal monitoring; in high-risk situations, it will restrict certain communication functions and raise the alert level; in extremely high-risk situations, the system will enter emergency protection mode, temporarily suspending communication with the outside world. A predefined response strategy library serves as a knowledge base for the system's protection capabilities, containing standardized response schemes for different risk levels and scenarios. The design of the response strategy library is based on best practices in communication security and the specific requirements of the tunnel environment. Each strategy scheme details the specific operational steps, parameter settings, duration, and other execution details.

[0075] The system executes intrusion prevention actions at the selected intrusion response level. These actions include adjusting signal transmission power, limiting communication frequency bands, and enabling enhanced authentication mechanisms. Simultaneously, it records the timestamp, geographical location, and matching blind spot pattern information of current blind spot events, generating a blind spot event log and storing it locally. Adjusting signal transmission power is a common protective measure. In low-risk situations, the system may appropriately increase transmission power to enhance signal coverage and reduce communication blind spots; in high-risk situations, the system may reduce transmission power to decrease signal propagation range and prevent interception by remote intruders. Power adjustment requires a balance between communication performance and security, ensuring normal communication while preventing information leakage. Limiting communication frequency bands is another important protective measure. The system can dynamically adjust available communication frequency bands according to the risk level. Under normal circumstances, smart tablet devices can use all allocated frequency bands for communication; in high-risk situations, the system will disable certain frequency bands that are easily eavesdropped or interfered with, retaining only the most secure communication channels. Frequency band limitations may affect communication bandwidth and data transmission rate, but effectively improve communication security. Enabling enhanced authentication mechanisms is a key protective measure for high-risk situations. Under normal circumstances, authentication between smart tablet devices may use simplified protocols to improve communication efficiency; however, when an intrusion risk is detected, the system will activate a more stringent authentication mechanism, such as multi-factor authentication, encryption key updates, and digital signature verification. Enhanced authentication increases communication overhead and latency, but significantly improves system security. Recording blind zone event logs is a crucial function of the intrusion prevention system, providing data support for subsequent security analysis and system optimization. Timestamps record the precise time of the event, including complete time information such as year, month, day, hour, minute, and second, providing a basis for time-series analysis. Geographic location records the spatial coordinates of the event, typically represented in a tunnel coordinate system, facilitating correlation analysis with tunnel maps and other spatial information. Matching blind zone pattern information includes detailed technical data such as the detected blind zone type, confidence level, and matching parameters. This information helps analyze the distribution patterns, trends, and influencing factors of blind zones. The log also records execution information such as specific protective measures taken, duration, and effectiveness evaluation. Local storage mechanisms ensure that important security event information is not lost in the event of network connection interruption. The system uses methods such as a circular buffer or database to store log data locally.

[0076] Finally, it should be noted that although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart tablet anti-intrusion system for field communication, characterized in that, The system includes: a multipath feature parameter extraction module, a multipath statistical feature calculation and compensation module, a blind zone matching and judgment module, and an intrusion prevention response and log management module, with each module communicating with the others in sequence. The multipath feature parameter extraction module is used to periodically transmit detection signals in a tunnel environment through a smart tablet device, receive and collect response signals from legitimate communication nodes, and extract multipath feature parameters from the response signals. The multipath feature parameters include signal arrival time difference, signal arrival angle, signal strength distribution, and channel impulse response. The multipath statistical feature calculation and compensation module is used to calculate the multipath statistical features of the current channel based on the multipath feature parameters. The multipath statistical features include the number of multipath components, the time delay spread, the angle spread, and the signal strength fluctuation variance. The multipath statistical features are then compensated by performing motion compensation based on the real-time moving speed and inertial measurement data of the smart tablet device to obtain the compensated multipath statistical features. The blind zone matching and determination module is used to perform matching analysis between the compensated multipath statistical features and the tunnel multipath blind zone fingerprint database to obtain a comprehensive matching degree. When the comprehensive matching degree exceeds a preset matching threshold, a matching success message is obtained. The tunnel multipath blind zone fingerprint database contains blind zone channel patterns caused by signal multipath effects in the tunnel environment. The blind zone channel patterns are established through systematic channel measurement and analysis of the tunnel. Each blind zone channel pattern corresponds to the multipath propagation characteristics generated by specific tunnel geometry and reflection environment. The intrusion prevention response and log management module is used to determine that the smart tablet device is in a communication blind zone and start a hierarchical intrusion prevention response when a successful matching information is detected; when the multipath statistical characteristics are detected to recover to the non-blind zone channel mode and a secure connection is re-established with the legitimate communication node, the intrusion prevention response is exited and the blind zone event log is uploaded. The multipath statistical feature calculation and compensation module includes: An inertial measurement data acquisition module is used to continuously acquire inertial measurement data of the smart tablet device through the inertial measurement unit built into the smart tablet device. The inertial measurement data includes acceleration, angular velocity and azimuth data. The motion trajectory and attitude calculation module is used to calculate the three-dimensional motion trajectory and real-time attitude change of the smart tablet device in the tunnel space by using the inertial measurement data and real-time moving speed of the smart tablet device through a sensor fusion algorithm. The multipath compensation module is used to establish a mapping relationship model between the movement state of the smart tablet device and the changes in multipath statistical characteristics based on the three-dimensional motion trajectory and real-time attitude changes; and to perform dynamic inverse compensation calculations on the number of multipath components, time delay spread, angle spread, and signal strength fluctuation variance according to the mapping relationship model to obtain the compensated multipath statistical characteristics. The multipath compensation module includes the following steps: The nonlinear relationship between the movement state of the smart tablet device and the changes in various multipath statistical features is analyzed based on the mapping relationship model. The expected distortion of multipath features caused by the movement of the smart tablet device is calculated in real time based on the three-dimensional motion trajectory and real-time attitude change data of the smart tablet device according to the nonlinear relationship. For the number of multipath components, delay spread, angle spread and signal strength fluctuation variance, respectively, a corresponding inverse compensation function is constructed. The expected distortion of the multipath features is obtained by performing inverse calculation on the measured multipath statistical features in the inverse compensation function to obtain the compensated multipath statistical features that are independent of the movement of the smart tablet device and only reflect the inherent characteristics of the tunnel space channel. The blind zone matching and determination module includes the following steps: Extract the reference range of multipath statistical features corresponding to each blind zone channel mode from the tunnel multipath blind zone fingerprint database, and calculate the feature matching similarity between the number of multipath components, delay spread, angle spread, and signal strength fluctuation variance in the compensated multipath statistical features and the corresponding feature reference range of each blind zone channel mode. The similarity scores of each feature are weighted and fused to obtain a comprehensive matching score. When the comprehensive matching score exceeds a preset matching threshold, a matching success message indicating a successful match with a specific blind zone channel mode is generated, and the blind zone confidence score and geographical location information corresponding to the blind zone channel mode are output.

2. The intelligent flat-panel anti-intrusion system for field communication according to claim 1, characterized in that, The motion trajectory and attitude calculation module includes: An inertial measurement data correction module is used to obtain corrected inertial measurement data from the inertial measurement data of the smart tablet device through data preprocessing. The Kalman filter pose optimization module is used to input the corrected inertial measurement data and real-time moving speed into the Kalman filter for multi-source data fusion. The prior estimate of the pose state of the smart tablet device is calculated through the prediction stage of the Kalman filter. In the update stage, the real-time moving speed is used as the observation input and the prior estimate is corrected in combination with the observation model to obtain the optimized posterior estimate of the pose state. The continuous three-dimensional motion trajectory construction module is used to iteratively execute the prediction stage and the update stage, continuously calculate the three-dimensional position coordinates, motion velocity vector and real-time attitude change in Euler angles of the smart tablet device in the tunnel coordinate system; and construct the continuous three-dimensional motion trajectory of the smart tablet device in the tunnel space based on the continuously calculated pose state sequence. 3.The intelligent flat panel anti-intrusion system for field communication of claim 2, wherein, The inertial measurement data correction module includes: A dual-level filtering module is used to perform sliding window mean filtering on the inertial measurement data of the smart tablet device to suppress high-frequency noise components, and then use wavelet transform filtering to separate and eliminate residual random noise to obtain the pre-filtered inertial measurement data. The zero-bias error correction module is used to calculate the zero-bias estimate of each inertial measurement channel at the current time by using the zero-bias error compensation model to calculate the inertial measurement data after the initial filtering. The zero-bias error compensation model is established based on the statistical characteristics of inertial measurement data of the smart tablet device operating in the tunnel environment for a long time. The corrected inertial measurement data is obtained by subtracting the zero-bias estimate of the corresponding inertial measurement channel from the initial filtered inertial measurement data.

4. The intelligent flat-panel anti-intrusion system for field communication according to claim 3, characterized in that, The zero-bias error correction module includes the following steps: Historical inertial measurement data output by the inertial measurement unit of the intelligent flat panel device under typical tunnel operating conditions are collected. The historical inertial measurement data is preprocessed to remove outliers and aligned. Then, the zero-bias error statistical characteristics of each inertial measurement channel are extracted. The zero-bias error statistical characteristics include mean, variance and drift characteristics over time. A mathematical model of the zero bias error is established based on the statistical characteristics of the zero bias error. The mathematical model of the zero bias error includes the reference zero bias value and adaptive adjustment coefficient of each inertial measurement channel. The mathematical model of the zero bias error is trained and optimized by a machine learning algorithm to obtain a zero bias error compensation model.

5. The intelligent flat panel anti-intrusion system for field communication according to claim 2, characterized in that, The Kalman filter pose optimization module includes the following steps: The real-time moving speed is converted into a velocity observation vector in the tunnel coordinate system. The predicted observation value corresponding to the prior estimate is calculated according to the observation model, and the observation residual between the predicted observation value and the velocity observation vector is obtained. The observation residual is weighted and adjusted according to the Kalman filter gain matrix to obtain the state correction amount. The state correction amount is added to the prior estimate and fused to obtain the optimized posterior estimate of the pose state.

6. The intelligent flat panel anti-intrusion system for field communication according to claim 1, characterized in that, The intrusion prevention response and log management module includes the following steps: The blind zone confidence level and the matched blind zone channel mode in the successful matching information are analyzed, and the intrusion risk level is determined according to the level of the blind zone confidence level. The corresponding intrusion prevention response level is selected from the predefined response strategy library according to the intrusion risk level. The predefined response strategy library contains corresponding communication security strategies under different intrusion risk levels. Perform intrusion prevention operations at the selected intrusion prevention response level. These operations include adjusting signal transmission power, limiting communication frequency bands, and enabling enhanced authentication mechanisms. Simultaneously, record the timestamp, geographical location, and matching blind zone pattern information of the current blind zone event, generate a blind zone event log, and store it locally.

Images