Portable measurement and acquisition real-time alarm system and device for subway track

Abstract

The invention relates to the technical field of metro track measurement, and discloses a metro track portable measurement and acquisition real-time alarm system and device, which comprises the following modules: a data acquisition module, a data processing and analysis module, a data processing and analysis module, a data processing and analysis module and a data processing and analysis module, the abnormal mode recognition module is used for carrying out abnormal mode recognition after carrying out multi-scale feature extraction on collected data. According to the method, space-time characteristics of three-dimensional data tensor fusion orbit parameters are constructed, deep correlation characteristics are extracted through tensor decomposition, a disease mode is accurately divided through spectral clustering, and then disease development path prediction and risk probability calculation are achieved in combination with a Markov chain and a Bayesian network. According to the method, the comprehensiveness of track abnormal feature extraction is improved, the accuracy of disease pattern recognition is enhanced, the risk can be early warned in advance, a precise basis is provided for track maintenance, and the fault occurrence rate is effectively reduced.

Description

Technical Field

[0001] This invention relates to the field of subway track measurement technology, specifically to a portable subway track measurement and acquisition real-time alarm system and device. Background Technology

[0002] Track geometry parameter monitoring is a core technology for ensuring the safe operation of rail transit systems such as railways and subways. It mainly involves collecting key parameters such as track gauge, level, and direction to analyze the spatial state and temporal evolution of the track structure, providing data support for track maintenance and early warning of defects. Its technical performance directly affects the safety and comfort of train operation.

[0003] In existing technologies, track geometry parameter monitoring often adopts a combination of manual inspection and fixed monitoring points: manual inspection is carried out periodically using tools such as track gauges and levels to record key parameters; fixed monitoring points deploy sensor arrays to collect data in real time and transmit it to the backend. Some systems introduce time series analysis or spatial interpolation algorithms to perform simple fitting of the changing trend of a single parameter to help identify obvious track anomalies.

[0004] Existing technologies have significant limitations: First, traditional methods struggle to simultaneously integrate the spatial distribution characteristics and temporal evolution patterns of parameters, often leading to the neglect of early-stage disease characteristics due to the separation of spatiotemporal correlations; second, disease pattern classification relies on manual experience and lacks a data-driven, precise classification mechanism; and third, prediction models are mostly based on linear extrapolation of a single parameter, making it difficult to quantify the probability risk of disease development, resulting in delayed early warnings or a high false alarm rate, which cannot meet the refined maintenance needs of complex track environments. In view of this, we propose a portable measurement and acquisition real-time alarm system and device for subway tracks. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a portable measurement and acquisition real-time alarm system and device for subway tracks, which solves the problem that existing technologies, in monitoring track geometric parameters, are unable to simultaneously integrate the spatial distribution characteristics and temporal evolution patterns of parameters, thus neglecting early abnormal features.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a portable real-time alarm system and device for measuring and acquiring data on subway tracks, comprising the following modules:

[0007] The data acquisition module includes a multi-channel synchronous sampling unit, which is used to simultaneously acquire track gauge and horizontal data at multiple locations 16mm below the top surface of the rail. The track gauge measurement unit adopts dual-frequency laser interferometric ranging technology, and the horizontal measurement unit integrates a MEMS tilt sensor and a fiber optic gyroscope to form a redundant measurement system.

[0008] The data processing and analysis module, connected to the data acquisition module, includes an adaptive Kalman filter unit, a wavelet transform noise reduction unit, and a spatiotemporal correlation analysis unit, which is used to perform multi-scale feature extraction on the acquired data and then perform abnormal pattern recognition.

[0009] The positioning module, connected to the data processing and analysis module, uses UWB ultra-wideband positioning technology and track electronic map matching algorithm to achieve sub-meter level dynamic positioning. It includes a multi-base station collaborative positioning unit and a trajectory smoothing filter unit.

[0010] The data display and recording module is connected to the data processing and analysis module and the positioning module respectively. It includes a 3D visualization unit and a data compression and storage unit, which are used to generate orbital geometry cloud maps and support TB-level historical data storage.

[0011] The alarm module, connected to the data processing and analysis module, is used to implement intelligent alarm decisions based on risk entropy.

[0012] Preferably, the dual-frequency laser interferometric ranging technology in the data acquisition module uses 1550nm and 1310nm dual-wavelength lasers to eliminate the influence of atmospheric refractive index changes on the measurement results through beat frequency interference principle. The adaptive Kalman filter unit uses a Sage-Husa noise estimator to estimate the statistical characteristics of system noise and measurement noise in real time, and dynamically adjusts the filter gain through a fading factor to suppress the influence of model error on state estimation.

[0013] Preferably, the spatiotemporal correlation analysis unit constructs a spatiotemporal evolution tensor model of orbital geometric parameters, and achieves orbital anomaly evolution feature extraction and trend early warning through the following steps:

[0014] I. Construct a three-dimensional data tensor by combining track gauge, level, and other measurement data with spatial location and time dimensions, and characterize the spatial distribution characteristics and temporal evolution of the measurement parameters.

[0015] Second, tensor decomposition technology is used to reduce the dimensionality of high-dimensional data, extract the spatiotemporal feature patterns of orbital state changes, and identify the implicit correlations between different positional parameters.

[0016] Third, the evolution process of track defects is divided into typical patterns such as gradual and abrupt types using spectral clustering algorithm, and a feature vector library for each pattern is established.

[0017] Fourth, based on the Markov chain model, the disease development path is predicted, and the probability of the disease evolving to a dangerous state is calculated by combining the Bayesian network. When the predicted probability exceeds the threshold, an early warning is triggered.

[0018] Preferably, the multi-scale feature extraction includes:

[0019] Spatial scale: Fine-grained feature extraction of local deformation at the millimeter level on the track, while also covering macroscopic feature analysis of the overall smoothness of track sections at the hundred-meter level;

[0020] Time scale: Extracting short-term dynamic changes, medium-term evolution trends, and long-term deterioration patterns of the orbit;

[0021] Data Dimensions: Features are extracted from multi-source monitoring data across different data dimensions, including simultaneously extracting pixel-level texture features and region-level morphological features from image data, and extracting point cloud-level distance features and cross-sectional contour features from laser data.

[0022] Preferably, the UWB ultra-wideband positioning technology and the track electronic map matching algorithm include: using a multi-base station cooperative positioning network with nanosecond-level pulse signals to achieve centimeter-level ranging, combining the vector data of the three-dimensional track electronic map, and fusing UWB positioning data and inertial navigation data through geometric feature comparison and Dempster-Shafer evidence theory.

[0023] Preferably, the multi-base station cooperative positioning unit includes: at least three UWB positioning base stations deployed along the track line to form a distributed positioning network, each base station synchronously collecting ranging data of the mobile terminal through a synchronous clock and a time division multiple access mechanism, and using a least squares algorithm to fuse the multilateral positioning solution results.

[0024] Preferably, the data-driven intelligent alarm decision-making calculates the risk entropy value of each monitoring point on the track using information entropy theory. This risk entropy value integrates the fluctuation amplitude, rate of change, and spatial correlation of track gauge and horizontal parameters. At the same time, risk entropy classification thresholds are set, and different alarm levels are corresponding to different risk entropy values ​​in different ranges.

[0025] A portable real-time alarm device for measuring and collecting data on subway tracks includes a slide, a measuring chamber slidably connected to the outer surface of the slide, a level fixedly connected to the front surface of the measuring chamber, a signal antenna fixedly connected to the upper surface of the measuring chamber, an LED display screen fixedly connected to the upper surface of the measuring chamber, a drive motor fixedly connected to the outer side of the measuring chamber, a drive wheel fixedly connected to the output shaft of the drive motor, and a battery compartment opened on the back of the measuring chamber.

[0026] Preferably, the slide is rotatably connected to a pivot in the middle, the slide is divided into left and right parts of the same length by the pivot, the measuring chamber has a through groove inside, the measuring chamber has a rectangular groove on the lower surface, and an auxiliary wheel is rotatably connected to the lower surface of the measuring chamber.

[0027] Preferably, the output shaft of the drive motor passes through and rotates inside the measuring chamber, the upper surface of the slide is provided with scale markings, two sets of measuring chambers are provided, and the two sets of measuring chambers are respectively located on the left and right sides of the slide, and through-beam sensors are fixedly connected to the adjacent surfaces of the two sets of measuring chambers.

[0028] Working principle: When using this device, it is necessary to ensure that the power supply inside the battery compartment is normal and start the device. At the same time, place the two sets of measuring chambers on the surface of the rails on both sides of the track, and start the drive motor to drive the drive wheel to rotate synchronously, so that the whole device can move back and forth on the surface of the rails. During the movement, the level will detect the parallelism of the rails on both sides and record the levelness parameters accordingly. In addition, the distance between the rails can be preliminarily judged by man based on the scale information on the surface of the carriage. The through-beam sensor uses dual-frequency laser interferometric ranging technology to measure the specific distance information between the two sets of measuring chambers. At the same time, during the movement of this device, the signal antenna will synchronously transmit parameters including positioning information, track spacing, parallelism and other parameters to the data processing terminal for relevant personnel or equipment to process the data.

[0029] Furthermore, when the device is no longer in use, the two sets of carriages can be folded using the pivot. At this time, the auxiliary wheels of the two sets of measuring chambers will come together, allowing people to...

[0030] This invention provides a portable measurement and acquisition real-time alarm system and device for subway tracks. It features the following:

[0031] Beneficial effects:

[0032] 1. This invention constructs a three-dimensional data tensor to fuse the spatiotemporal characteristics of track parameters, extracts deep correlation features through tensor decomposition, accurately classifies fault patterns using spectral clustering, and then combines Markov chains and Bayesian networks to predict fault development paths and calculate hazard probabilities. This approach improves the comprehensiveness of track anomaly feature extraction, enhances the accuracy of fault pattern identification, enables early risk warning, provides precise basis for track maintenance, and effectively reduces the failure rate.

[0033] 2. The invention features a foldable hinge that makes the entire device easy to carry and can shift the carriage when the rail is not level, improving the accuracy of the level instrument in detecting anomalies. The dual measuring chambers are equipped with a through-beam sensor and utilize dual-frequency laser interferometry. By using dual wavelengths of 1550nm and 1310nm and beat frequency interference to eliminate atmospheric effects, combined with the carriage scale, the device can accurately measure parameters such as track gauge. Data is displayed in real time on an LED screen, and information is transmitted via a signal antenna. An alarm is triggered when the threshold is exceeded, providing efficient and accurate real-time monitoring and early warning support for track inspection.

[0034] 3. This invention achieves centimeter-level initial positioning through multi-base station collaboration and the least squares algorithm. By combining geometric feature comparison with a 3D track electronic map, the positioning results are constrained to a reasonable range to avoid drift. By utilizing Dempster-Shafer evidence theory to fuse UWB and inertial navigation data, the complementary advantages are leveraged to maintain continuous high-precision positioning even in complex scenarios such as occlusion and multipath propagation. This meets the stringent requirements of track mobile equipment for positioning reliability and accuracy, providing precise location support for track monitoring and inspection. Attached Figure Description

[0035] Figure 1 Flowchart of a portable measurement and data acquisition real-time alarm system for local railway tracks;

[0036] Figure 2 This is a schematic diagram of the orbital anomaly evolution feature extraction and trend early warning process of the present invention;

[0037] Figure 3 This is a perspective view of the device of the present invention;

[0038] Figure 4 This is a schematic diagram of the measuring chamber of the present invention.

[0039] The components include: 1. Carriage; 2. Rotary shaft; 3. Through-beam sensor; 4. LED display screen; 5. Signal antenna; 6. Measuring chamber; 7. Level; 8. Drive motor; 9. Auxiliary wheel; 10. Drive wheel; and 11. Battery compartment. Detailed Implementation

[0040] The technical solutions in 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.

[0041] Example:

[0042] Please see the appendix Figure 1 - Appendix Figure 2 This invention provides a portable measurement and acquisition real-time alarm system for subway tracks, comprising the following modules:

[0043] The data acquisition module includes a multi-channel synchronous sampling unit, which is used to simultaneously acquire track gauge and horizontal data at multiple locations 16mm below the top surface of the rail. The track gauge measurement unit adopts dual-frequency laser interferometric ranging technology, and the horizontal measurement unit integrates a MEMS tilt sensor and a fiber optic gyroscope to form a redundant measurement system.

[0044] The data processing and analysis module, connected to the data acquisition module, includes an adaptive Kalman filter unit, a wavelet transform noise reduction unit, and a spatiotemporal correlation analysis unit, which is used to perform multi-scale feature extraction on the acquired data and then perform abnormal pattern recognition.

[0045] The positioning module, connected to the data processing and analysis module, uses UWB ultra-wideband positioning technology and track electronic map matching algorithm to achieve sub-meter level dynamic positioning. It includes a multi-base station collaborative positioning unit and a trajectory smoothing filter unit.

[0046] The data display and recording module is connected to the data processing and analysis module and the positioning module respectively. It includes a 3D visualization unit and a data compression and storage unit, which are used to generate orbital geometry cloud maps and support TB-level historical data storage.

[0047] The alarm module, connected to the data processing and analysis module, is used to implement intelligent alarm decisions based on risk entropy.

[0048] The dual-frequency laser interferometric ranging technology in the data acquisition module uses 1550nm and 1310nm dual-wavelength lasers. It eliminates the influence of atmospheric refractive index changes on the measurement results through beat frequency interference principle. The adaptive Kalman filter unit uses the Sage-Husa noise estimator to estimate the statistical characteristics of system noise and measurement noise in real time. It dynamically adjusts the filter gain through fading factor to suppress the influence of model error on state estimation.

[0049] The spatiotemporal correlation analysis unit constructs a spatiotemporal evolution tensor model of orbital geometric parameters, and achieves orbital anomaly evolution feature extraction and trend early warning through the following steps:

[0050] I. Construct a three-dimensional data tensor from the track gauge, leveling, and other measurement data according to spatial location and time dimensions. This tensor simultaneously characterizes the spatial distribution characteristics and temporal evolution of the measurement parameters. The process of constructing the three-dimensional data tensor is as follows: Construct a three-dimensional data tensor from the track gauge, leveling, and other measurement data according to spatial location and time dimensions. Where m is the number of spatial location points, n is the length of the time series, and p is the dimension of the measurement parameter (such as track gauge, horizontal, etc.), this tensor simultaneously characterizes the spatial distribution characteristics and temporal evolution of the measurement parameter. Each element... This represents the measurement value of the kth parameter at location i and time j, providing a basic data structure for subsequent spatiotemporal correlation analysis;

[0051] II. Tensor decomposition is used to reduce the dimensionality of high-dimensional data, extract spatiotemporal feature patterns of orbital state changes, and identify implicit correlations between different positional parameters. Specifically, tensor decomposition and feature extraction involve using Tucker decomposition to reduce the dimensionality of high-dimensional data, extract spatiotemporal feature patterns of orbital state changes, and identify implicit correlations between different positional parameters.

[0052] The tensor is decomposed into a core tensor and three factor matrices:

[0053]

[0054] And make the algorithm satisfy

[0055]

[0056] in:

[0057] Three-dimensional data tensors, fusing the spatial distribution and temporal evolution characteristics of orbital parameters;

[0058] Core tensors are essential for storing data.

[0059] A, B, and C: Factor matrices, corresponding to features in the spatial, temporal, and parametric dimensions, respectively. (m is the number of spatial location points, r1 is the spatial feature dimension) (n is the length of the time series, r² is the dimension of the time feature) (p represents the dimension of the measurement parameters, and r3 represents the dimension of the parameter features);

[0060] III. The evolution process of track defects is divided into typical patterns such as gradual and abrupt changes using spectral clustering algorithms. A feature vector library for each pattern is established. Spectral clustering and pattern recognition involve: First, a similarity matrix S is constructed, where:

[0061]

[0062] x i and x j σ represents the eigenvectors after tensor decomposition, and σ is the kernel bandwidth parameter.

[0063] Then obtain the degree matrix D

[0064]

[0065] Then construct the Laplace matrix

[0066] L=D -1 / 2 SD -1 / 2

[0067] Perform eigenvalue decomposition on L, and form matrix U by taking the eigenvectors corresponding to the k smallest eigenvalues; finally, perform K-means clustering on each row of U to obtain different disease patterns, where:

[0068] S: Similarity matrix, which measures the degree of similarity between feature vectors;

[0069] S ij The element in the i-th row and j-th column of the similarity matrix is ​​calculated using the Gaussian kernel function, where x i x j σ represents the eigenvectors after tensor decomposition, and σ is the kernel bandwidth parameter.

[0070] D: Degree matrix, diagonal matrix, D ii It is the sum of the elements in the i-th row of the similarity matrix;

[0071] L: Laplacian matrix, used for eigenvalue decomposition in spectral clustering, constructed from the degree matrix and similarity matrix;

[0072] U: Eigenvector matrix, composed of the eigenvectors corresponding to the first k smallest eigenvalues ​​of the Laplacian matrix.

[0073] IV. Based on the Markov chain model, the disease development path is predicted, and combined with the Bayesian network, the probability of the disease evolving to a dangerous state is calculated. When the predicted probability exceeds the threshold, an early warning is triggered. The Markov chain and Bayesian network prediction is as follows: Based on the Markov chain model, the disease development path is predicted, and combined with the Bayesian network, the probability of the disease evolving to a dangerous state is calculated. When the predicted probability exceeds the threshold, an early warning is triggered.

[0074] The Markov chain state transition probability matrix P satisfies

[0075] P ij =P(X) t+1 =j|X t =i)

[0076]

[0077] The transition probability at step t is P (t) =P t In a Bayesian network, the probability of a disease state.

[0078]

[0079] H represents the hypothesis of disease state, and E represents observational evidence;

[0080] Risk Entropy

[0081]

[0082] p i =P(H i |E)

[0083] This quantifies uncertainty, and an alert is triggered when the probability exceeds a threshold, where:

[0084] P: Markov chain state transition probability matrix, P ij This represents the probability of transitioning from state i to state j;

[0085] X t The state variables at time t;

[0086] P (t) The t-step transition probability matrix is ​​obtained by raising P to the power of t.

[0087] H: Disease status assumption (e.g., normal, warning, danger);

[0088] E: Observational evidence (such as spatiotemporal feature vectors);

[0089] P(H|E): Posterior probability, the probability of the disease state given the evidence;

[0090] P(E|H): Likelihood function, the probability of observing evidence given a state;

[0091] P(H): Prior probability, state probability based on historical data or expert knowledge;

[0092] P(E): Marginal probability of the evidence;

[0093] H(p): Risk entropy, quantifying the uncertainty of disease status, where p i =P(H i |E) represents the posterior probability of the i-th state.

[0094] The multi-scale feature extraction includes:

[0095] Spatial scale: Fine-grained feature extraction of local deformation at the millimeter level on the track, while also covering macroscopic feature analysis of the overall smoothness of track sections at the hundred-meter level;

[0096] Time scale: Extracting short-term dynamic changes, medium-term evolution trends, and long-term deterioration patterns of the orbit;

[0097] Data Dimensions: Features are extracted from multi-source monitoring data across different data dimensions, including simultaneously extracting pixel-level texture features and region-level morphological features from image data, and extracting point cloud-level distance features and cross-sectional contour features from laser data.

[0098] The UWB ultra-wideband positioning technology and track electronic map matching algorithm include: using a multi-base station cooperative positioning network with nanosecond-level pulse signals to achieve centimeter-level ranging; combining vector data from a three-dimensional track electronic map; and fusing UWB positioning data with inertial navigation data through geometric feature comparison and Dempster-Shafer evidence theory. The multi-base station cooperative positioning unit includes: at least three UWB positioning base stations deployed along the track line to form a distributed positioning network. Each base station synchronously collects ranging data from the mobile terminal through a synchronized clock and time-division multiple access mechanism. The least squares algorithm is used to fuse the multilateral positioning solution results, including the following algorithms:

[0099] Step 1: Deployment and Synchronous Networking of Multiple Base Stations

[0100] At least three UWB positioning base stations are deployed along the track to form a distributed positioning network. Each base station achieves time synchronization through a synchronous clock system and uses a time division multiple access mechanism to allocate communication time slots, ensuring synchronous acquisition of ranging data from the mobile terminal and obtaining the distance measurement value d from the terminal to each base station. i (i = 1, 2, ..., n, n ≥ 3);

[0101] Step 2: Construction and Solution of the Polygonal Positioning Equation System

[0102] Let the coordinates of the mobile terminal be (x, y, z), and the known coordinates of the i-th base station be (x, y, z). i ,y i ,z i A system of nonlinear equations is established based on the distance formula between two points in space. The system of equations can be linearized to Ax = b (where A is the coefficient matrix, and x = [x, y, z]). T Let b be the coordinate vector to be determined, and let b be a constant term vector. The initial coordinates of the terminal are obtained using the least squares algorithm.

[0103] Step 3: Multi-source positioning data acquisition and preprocessing

[0104] Collect UWB positioning data (including the initial coordinates solved in step two) and inertial navigation data (including acceleration, angular velocity and calculated coordinates), and extract vector data (such as the coordinates of the orbit centerline, curvature, etc.) from the three-dimensional orbit electronic map. Perform noise filtering and outlier removal on the raw data.

[0105] Step 4: Geometric Feature Alignment and Matching Metrics

[0106] Calculate the distance deviation between the UWB positioning point, the inertial navigation calculated point, and the nearest orbital centerline on the electronic orbital map. ((x p ,y p (x) represents the coordinates of the positioning point.m ,y m (where α is the coordinate of the nearest point on the map) and the azimuth deviation θ = |α| p -α m |(α p Let α be the azimuth angle of the positioning point. m (The tangent angle of the corresponding point on the map) is used to quantify the geometric matching degree between the positioning data and the map through the deviation value;

[0107] Step 5: Integration and Optimization of Evidence Theories

[0108] Based on the geometric feature comparison results, a basic probability assignment m is constructed between UWB positioning data and inertial navigation data for each positioning state (e.g., "reliable match" and "suspected match"). uwb (A) and m ins (B), satisfies and (Θ represents the identification frame); calculate the conflict coefficient. K =∑ A∩B-0 m uwb (A)·m ins (B) The fusion probability is obtained by applying the synthesis rule:

[0109]

[0110] Step Six: Map Matching Correction and Final Location Output

[0111] The coordinates corresponding to the location state with the highest credibility after fusion are selected and matched and corrected with the electronic map of the track to ensure that the final location point falls within a reasonable range of the map and output the terminal positioning result that meets the accuracy requirements.

[0112] The data-driven intelligent alarm decision-making system calculates the risk entropy value of each monitoring point on the track using information entropy theory. This risk entropy value takes into account the fluctuation amplitude, rate of change, and spatial correlation of track gauge and horizontal parameters. At the same time, risk entropy classification thresholds are set, and different alarm levels are corresponding to different risk entropy values ​​in different ranges.

[0113] Please see the appendix Figure 3 - Appendix Figure 4 A portable measurement and acquisition real-time alarm device for subway tracks includes a slide 1. A measurement chamber 6 is slidably connected to the outer surface of the slide 1. Protrusions are provided on both the left and right sides of the slide 1 to block and restrict the measurement chamber 6 from detaching from the slide 1. A level 7 is fixedly connected to the front surface of the measurement chamber 6. A signal antenna 5 is fixedly connected to the upper surface of the measurement chamber 6. An LED display screen 4 is fixedly connected to the upper surface of the measurement chamber 6. A drive motor 8 is fixedly connected to the outer side of the measurement chamber 6. A drive wheel 10 is fixedly connected to the output shaft of the drive motor 8. A battery compartment 11 is provided on the back of the measurement chamber 6.

[0114] A pivot 2 is rotatably connected to the middle of the slide 1. This pivot 2 allows the device to be folded and rotated when not in use. During levelness measurement, the non-level measuring chamber 6 will cause the slide 1 to shift at an angle, allowing the level instrument 7 to more accurately capture levelness anomalies. The slide 1 is divided into two equal parts by the pivot 2. A through-groove is provided inside the measuring chamber 6, allowing it to slide on the surface of the slide 1. A rectangular groove is provided on the lower surface of the measuring chamber 6 to accommodate the drive wheel 10 and the auxiliary wheel 9. An auxiliary wheel 9 is rotatably connected to the slide 1. The output shaft of the drive motor 8 passes through and rotates inside the measuring chamber 6. The upper surface of the slide 1 is marked with scale markings, so the rails on both sides can be roughly evaluated through these markings. There are two sets of measuring chambers 6, and the two sets of measuring chambers 6 are located on the left and right sides of the slide 1, respectively. A through-beam sensor 3 is fixedly connected to the close surfaces of the two sets of measuring chambers 6. The through-beam sensor realizes the dual-frequency laser interferometric ranging technology. The two sets of through-beam sensors 3 are 1550nm and 1310nm dual-wavelength lasers, respectively. The influence of atmospheric refractive index changes on the measurement results is eliminated by the beat frequency interference principle.

[0115] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A portable measurement and acquisition real-time alarm system for subway tracks, characterized in that, Includes the following modules: The data acquisition module includes a multi-channel synchronous sampling unit, which is used to simultaneously acquire track gauge and horizontal data at multiple locations 16mm below the top surface of the rail. The track gauge measurement unit adopts dual-frequency laser interferometric ranging technology, and the horizontal measurement unit integrates a MEMS tilt sensor and a fiber optic gyroscope to form a redundant measurement system. The data processing and analysis module, connected to the data acquisition module, includes an adaptive Kalman filter unit, a wavelet transform noise reduction unit, and a spatiotemporal correlation analysis unit, which is used to perform multi-scale feature extraction on the acquired data and then perform abnormal pattern recognition. The positioning module, connected to the data processing and analysis module, uses UWB ultra-wideband positioning technology and track electronic map matching algorithm to achieve sub-meter level dynamic positioning. It includes a multi-base station collaborative positioning unit and a trajectory smoothing filter unit. The data display and recording module is connected to the data processing and analysis module and the positioning module respectively. It includes a 3D visualization unit and a data compression and storage unit, which are used to generate orbital geometry cloud maps and support TB-level historical data storage. The alarm module, connected to the data processing and analysis module, is used to implement intelligent alarm decisions based on risk entropy.

2. The portable measurement and acquisition real-time alarm system for subway tracks according to claim 1, characterized in that, The dual-frequency laser interferometric ranging technology in the data acquisition module uses 1550nm and 1310nm dual-wavelength lasers. It eliminates the influence of atmospheric refractive index changes on the measurement results through beat frequency interference principle. The adaptive Kalman filter unit uses the Sage-Husa noise estimator to estimate the statistical characteristics of system noise and measurement noise in real time. It dynamically adjusts the filter gain through fading factor to suppress the influence of model error on state estimation.

3. The portable measurement and acquisition real-time alarm system for subway tracks according to claim 1, characterized in that, The spatiotemporal correlation analysis unit constructs a spatiotemporal evolution tensor model of orbital geometric parameters, and achieves orbital anomaly evolution feature extraction and trend early warning through the following steps: I. Construct a three-dimensional data tensor by combining track gauge, level, and other measurement data with spatial location and time dimensions, and characterize the spatial distribution characteristics and temporal evolution of the measurement parameters. Second, tensor decomposition technology is used to reduce the dimensionality of high-dimensional data, extract the spatiotemporal feature patterns of orbital state changes, and identify the implicit correlations between different positional parameters. Third, the evolution process of track defects is divided into typical patterns such as gradual and abrupt types using spectral clustering algorithm, and a feature vector library for each pattern is established. Fourth, based on the Markov chain model, the disease development path is predicted, and the probability of the disease evolving to a dangerous state is calculated by combining the Bayesian network. When the predicted probability exceeds the threshold, an early warning is triggered.

4. The portable measurement and acquisition real-time alarm system for subway tracks according to claim 1, characterized in that, The multi-scale feature extraction includes: Spatial scale: Fine-grained feature extraction of local deformation at the millimeter level on the track, while also covering macroscopic feature analysis of the overall smoothness of track sections at the hundred-meter level; Time scale: Extracting short-term dynamic changes, medium-term evolution trends, and long-term deterioration patterns of the orbit; Data Dimensions: Features are extracted from multi-source monitoring data across different data dimensions, including simultaneously extracting pixel-level texture features and region-level morphological features from image data, and extracting point cloud-level distance features and cross-sectional contour features from laser data.

5. The portable measurement and acquisition real-time alarm system for subway tracks according to claim 1, characterized in that, The UWB ultra-wideband positioning technology and track electronic map matching algorithm include: using a multi-base station cooperative positioning network with nanosecond-level pulse signals to achieve centimeter-level ranging, combining vector data from a three-dimensional track electronic map, and fusing UWB positioning data and inertial navigation data through geometric feature comparison and Dempster-Shafer evidence theory.

6. The portable measurement and acquisition real-time alarm system for subway tracks according to claim 1, characterized in that, The multi-base station cooperative positioning unit includes: at least three UWB positioning base stations deployed along the track line to form a distributed positioning network. Each base station synchronously collects ranging data of the mobile terminal through a synchronous clock and a time division multiple access mechanism, and uses the least squares algorithm to fuse the multilateral positioning solution results.

7. The portable measurement and acquisition real-time alarm system for subway tracks according to claim 1, characterized in that, The data-driven intelligent alarm decision-making system calculates the risk entropy value of each monitoring point on the track using information entropy theory. This risk entropy value integrates the fluctuation amplitude, rate of change, and spatial correlation of track gauge and horizontal parameters. At the same time, risk entropy classification thresholds are set, and different alarm levels are corresponding to different risk threshold ranges.

8. The portable measurement and acquisition real-time alarm device for subway tracks according to claim 1, characterized in that, The device includes a slide (1), a measuring chamber (6) which is slidably connected to the outer surface of the slide (1), a level (7) which is fixedly connected to the front surface of the measuring chamber (6), a signal antenna (5) which is fixedly connected to the upper surface of the measuring chamber (6), an LED display screen (4) which is fixedly connected to the upper surface of the measuring chamber (6), a drive motor (8) which is fixedly connected to the outer side of the measuring chamber (6), a drive wheel (10) which is fixedly connected to the output shaft of the drive motor (8), and a battery compartment (11) which is opened on the back of the measuring chamber (6).

9. The portable measurement and acquisition real-time alarm device for subway tracks according to claim 8, characterized in that, The slide (1) is rotatably connected to a rotating shaft (2) in the middle. The slide (1) is divided into two parts of equal length by the rotating shaft (2). A through groove is provided inside the measuring chamber (6). A rectangular groove is provided on the lower surface of the measuring chamber (6). An auxiliary wheel (9) is rotatably connected to the lower surface of the measuring chamber (6).

10. The portable measurement and acquisition real-time alarm device for subway tracks according to claim 8, characterized in that, The output shaft of the drive motor (8) passes through and rotates inside the measuring chamber (6). The upper surface of the slide (1) is marked with scale. There are two sets of measuring chambers (6), and the two sets of measuring chambers (6) are located on the left and right sides of the slide (1) respectively. The two sets of measuring chambers (6) are fixedly connected to the adjacent surfaces of the two sets of measuring chambers (6).

Images