A railway communication operation and maintenance monitoring method and system and a storage medium

Abstract

This invention discloses a railway communication operation and maintenance monitoring method, system, and storage medium, belonging to the field of railway communication operation and maintenance monitoring technology. This invention performs unified spatiotemporal processing on multi-source monitoring data of railway communication to form a unified spatiotemporal multimodal feature sequence. Then, a multi-level fusion strategy is used to deeply fuse the unified spatiotemporal multimodal feature sequence to obtain comprehensive health characterization vector information. Diagnostic analysis of the comprehensive health characterization vector information can reflect the current operating status and potential fault risks of railway communication. It can not only accurately determine the current operating status of railway communication but also effectively predict the probability distribution of faults or intrusion events. Subsequently, it can automatically generate and issue targeted operation and maintenance early warning information, guiding operation and maintenance personnel to take timely measures to ensure the stable operation of the railway communication system, thereby significantly improving the safety and efficiency of railway transportation.

Description

Technical Field

[0001] This invention relates to the field of railway communication operation and maintenance monitoring technology, and in particular to a railway communication operation and maintenance monitoring method, system and storage medium. Background Technology

[0002] Railway communication operation and maintenance (O&M) is a crucial link in ensuring railway transportation safety and improving efficiency, and its stability and reliability are of paramount importance. However, traditional railway communication O&M monitoring methods often rely on manual inspections and periodic maintenance. This approach is not only inefficient but also struggles to detect and resolve potential problems in real time, resulting in issues such as delayed fault detection, high false alarm rates, difficulty in root cause localization, and low prediction accuracy. With the deployment of 5G-R and FRMCS, communication data exhibits characteristics of multi-source heterogeneity, long time series, and high noise, making it difficult for traditional methods to meet the high reliability and low latency O&M requirements of high-speed rail. Summary of the Invention

[0003] The purpose of this invention is to address the shortcomings of existing railway communication operation and maintenance methods that are difficult to meet the current needs of railway communication operation and maintenance, and to propose a railway communication operation and maintenance monitoring method, system and storage medium.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] The present invention provides a railway communication operation and maintenance monitoring method in a first aspect, comprising:

[0006] Based on the railway communication operation and maintenance needs, collect multi-source heterogeneous data information of railway communication in real time;

[0007] Spatiotemporal alignment preprocessing is performed on multi-source heterogeneous data information from railway communications to obtain a unified spatiotemporal multimodal feature sequence;

[0008] Multi-level fusion of unified spatiotemporal multimodal feature sequences is performed to obtain comprehensive health representation vector information;

[0009] Based on the railway communication operation and maintenance requirements, the comprehensive health representation vector information is monitored and diagnosed to obtain the current railway communication operation status and fault probability distribution information.

[0010] Based on the railway communication operation and maintenance requirements, and combined with the current railway communication operation status and fault probability distribution information, the current railway communication operation and maintenance early warning information is determined.

[0011] In one feasible approach, the method for obtaining a unified spatiotemporal multimodal feature sequence includes:

[0012] Dynamic time warping The algorithm performs time-axis alignment on time-series data with different sampling frequencies. Assume there are two time series that need to be spatiotemporally aligned. and The calculation formula is as follows:

[0013] Formula 1;

[0014] In Equation 1, Two time series and The dynamic time-normalized distance between them; Time series The first in Observations at each time point; Time series The first in Observations at each time point; Time series No. Time points With time series The first in Time points Local distance between them; The optimal alignment path is determined to satisfy monotonicity, continuity, and boundary conditions.

[0015] In one feasible approach, the multi-level fusion includes one or more of the following: temporal feature-level fusion, cross-modal attention mechanism fusion, and decision-level weighted fusion.

[0016] In one feasible approach, the method for multi-level fusion of unified spatiotemporal multimodal feature sequences includes:

[0017] Based on railway operation and maintenance requirements, feature splicing processing is performed on unified spatiotemporal multimodal feature sequences to obtain preliminary fused feature vectors.

[0018] A multi-head attention mechanism is used to interactively model features from different modalities on the initially fused feature vectors, completing the cross-modal attention mechanism fusion. The calculation process is as follows:

[0019] Formula 2;

[0020] Formula 3;

[0021] In equations 2 and 3, , , These are the query, key, and value matrices, obtained by linear projection of different modal features. For each dimension of attention head, For the number of attention heads, To output the projection matrix; This is the normalization function; For splicing operations;

[0022] After completing the cross-modal attention mechanism fusion, according to the railway communication operation and maintenance requirements, the cross-modal attention mechanism fusion results are assigned corresponding weights, and the comprehensive health representation vector information is obtained by weighted summation.

[0023] In one feasible approach, the method for obtaining the current operating status and fault probability distribution information of railway communication includes:

[0024] Based on the multi-source heterogeneous data information of railway communication, the comprehensive health characterization vector information is diagnosed to obtain the current railway communication operation status classification results and the probability distribution of corresponding fault / intrusion events;

[0025] The multi-source heterogeneous data information includes: based on Vibration disturbance data from distributed fiber optic sensing signals are used for precise identification and location of intrusion events along railway lines.

[0026] In one feasible approach, the method for obtaining the current railway communication operation status classification results and the probability distribution of corresponding fault / intrusion events includes:

[0027] The backscattered Rayleigh scattering signal from the Φ-OTDR is collected to form a two-dimensional spatiotemporal perturbation characteristic signal;

[0028] The spatiotemporal two-dimensional perturbation feature signal is subjected to denoising preprocessing to obtain the denoised phase signal;

[0029] Intrusion event detection and spatiotemporal region segmentation are performed based on the denoised phase signal to obtain at least one event spatiotemporal block;

[0030] For each event spatiotemporal block, multi-domain spatiotemporal features are extracted to form a feature vector;

[0031] Based on the railway communication operation and maintenance requirements, pattern recognition is performed on the feature vectors to obtain the intrusion event detection category;

[0032] High-precision localization is performed on the spatiotemporal blocks of events identified as real intrusion events to obtain the location of the intrusion events.

[0033] In one feasible approach, the method for forming the spatiotemporal two-dimensional perturbation feature signal includes:

[0034] Let the backscattered Rayleigh signal acquired by the Φ-OTDR be... At this point, the phase vibration signal is obtained by differential phase demodulation between adjacent pulses, and the phase change is calculated:

[0035] Equation 4;

[0036] in, This represents the amount of phase change due to the disturbance. These are the spatial coordinates on the optical fiber; The time coordinates for signal acquisition; For complex conjugate; Spatial sampling interval; To obtain the phase angle.

[0037] In one feasible approach, the method for obtaining the intrusion event detection category includes:

[0038] The temporal energy of each event spatiotemporal block is calculated using the following formula:

[0039] Formula 5;

[0040] In Equation 5, In fiber optic location and time The local time-domain energy value at that location; The length of the sliding window; For the coordinates of the fiber optic position Location and time coordinates Phase vibration signal at the location;

[0041] Based on the needs of railway communication operation and maintenance, an adaptive threshold is set. , wherein the adaptive threshold for:

[0042] Formula 6;

[0043] In Equation 6, Background noise or time-domain energy under normal conditions The mean; Time-domain energy under background noise The standard deviation of ; m is the empirical coefficient.

[0044] In a second aspect, the present invention provides a railway communication operation and maintenance monitoring system, which employs the railway communication operation and maintenance monitoring method described in any one of the first aspects, and the monitoring system further includes:

[0045] The acquisition module collects multi-source heterogeneous data information from railway communications in real time.

[0046] The preprocessing module is used to perform spatiotemporal alignment and normalization preprocessing on the multi-source heterogeneous data information of railway communication transmitted from the acquisition module to obtain a unified spatiotemporal multimodal feature sequence.

[0047] The fusion module is used to perform multi-level fusion of unified spatiotemporal multimodal feature sequences to obtain comprehensive health representation vector information;

[0048] The monitoring and diagnosis module is used to monitor and diagnose the comprehensive health representation vector information transmitted from the fusion module, and obtain the current operating status and fault probability distribution information of railway communication.

[0049] The early warning and maintenance module is used to determine the current railway communication operation and maintenance early warning information based on the current operating status and fault probability distribution information of the railway communication.

[0050] In a third aspect, the present invention provides a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements a railway communication operation and maintenance monitoring method as described in any one of the first aspects.

[0051] The beneficial effects of this invention are as follows:

[0052] This invention performs unified spatiotemporal processing on multi-source monitoring data of railway communication to form a unified spatiotemporal multimodal feature sequence. A multi-level fusion strategy is then employed to deeply fuse this sequence, obtaining comprehensive health representation vector information. Diagnostic analysis of this comprehensive health representation vector information reflects the current operational status and potential fault risks of railway communication. It not only accurately determines the current operational status of railway communication but also effectively predicts the probability distribution of faults or intrusion events. Consequently, it automatically generates and issues targeted maintenance warnings, guiding maintenance personnel to take timely measures to ensure the stable operation of the railway communication system, thereby significantly improving the safety and efficiency of railway transportation. In short, it effectively addresses the shortcomings of existing railway communication maintenance methods, which are insufficient to meet the current needs of railway communication maintenance. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of the overall process of a railway communication operation and maintenance monitoring method provided in an embodiment of the present invention;

[0054] Figure 2 This is a schematic diagram of the intrusion event location process of a railway communication operation and maintenance monitoring method provided in an embodiment of the present invention. Detailed Implementation

[0055] 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 a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0056] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0057] In this invention, unless otherwise explicitly specified and limited, the terms "connection," "fixed," etc., should be interpreted broadly. For example, "fixed" can mean a fixed connection, a detachable connection, or an integral part; it can mean a mechanical connection or an electrical connection; it can mean a direct connection or an indirect connection through an intermediate medium; it can mean the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0058] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the meaning of "and / or" throughout the text includes three parallel solutions; for example, "A and / or B" includes solution A, solution B, or a solution where both A and B are satisfied simultaneously. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0059] Reference Figures 1 to 2To address the shortcomings of existing railway communication operation and maintenance methods that fail to meet current railway communication operation and maintenance needs, this invention provides a railway communication operation and maintenance monitoring method. This method performs unified spatiotemporal processing on multi-source monitoring data of railway communication to form a unified spatiotemporal multimodal feature sequence. Furthermore, a multi-level fusion strategy is employed to deeply fuse this unified spatiotemporal multimodal feature sequence, obtaining comprehensive health representation vector information. Diagnostic analysis of this comprehensive health representation vector information reflects the current operational status and potential fault risks of railway communication. It not only accurately determines the current operational status of railway communication but also effectively predicts the probability distribution of faults or intrusion events. Consequently, it automatically generates and issues targeted operation and maintenance early warning information, guiding maintenance personnel to take timely measures to ensure the stable operation of the railway communication system, thereby significantly improving the safety and efficiency of railway transportation. In short, it effectively solves the shortcomings of existing railway communication operation and maintenance methods that fail to meet current railway communication operation and maintenance needs.

[0060] Specifically, in its first aspect, the present invention provides a railway communication operation and maintenance monitoring method, the monitoring method comprising: real-time collection of multi-source heterogeneous data information of railway communication according to railway communication operation and maintenance needs; wherein, the multi-source heterogeneous data information includes, but is not limited to: network element equipment performance index data, alarm event data, signaling tracing data, and optical transmission layer data. The system collects optical power data and environmental / energy consumption monitoring data. Then, it performs spatiotemporal alignment preprocessing on the multi-source heterogeneous data of railway communication (including but not limited to spatiotemporal alignment and standardization preprocessing of the multi-source heterogeneous data to obtain a unified spatiotemporal multimodal feature sequence). This unified spatiotemporal multimodal feature sequence is then fused at multiple levels to obtain comprehensive health representation vector information. It should be noted that the multi-level fusion includes: temporal feature-level fusion, cross-modal attention mechanism fusion, and decision-level weighted fusion. Subsequently, based on the railway communication operation and maintenance requirements, the comprehensive health representation vector information can be monitored and diagnosed to obtain the current operating status and fault probability distribution information of the railway communication.

[0061] Based on the operational and maintenance needs of railway communication, and combined with the current operational status and fault probability distribution information of railway communication, early warning information for current railway communication operation and maintenance is determined. In this embodiment, through comprehensive collection and in-depth processing of multi-source heterogeneous data, the railway communication operation and maintenance monitoring method proposed in this invention can accurately capture subtle changes in the railway communication system. The spatiotemporal alignment preprocessing step ensures that data from different sources and in different formats can be analyzed within a unified temporal and spatial framework, thereby eliminating spatiotemporal differences between data and laying a solid foundation for subsequent multi-level fusion. The application of multi-level fusion strategies, especially the organic combination of temporal feature-level fusion, cross-modal attention mechanism fusion, and decision-level weighted fusion, enables the comprehensive health representation vector information to comprehensively and accurately reflect the current status and potential risks of the railway communication system. This innovative data processing flow not only improves the accuracy of fault detection but also significantly enhances the accuracy of prediction, providing strong technical support for railway communication operation and maintenance. Ultimately, based on the monitoring and diagnostic results of the comprehensive health representation vector information, the system can automatically generate and issue targeted operation and maintenance early warning information, guiding operation and maintenance personnel to respond quickly and take timely measures to ensure the continuous and stable operation of the railway communication system, providing a solid guarantee for the safety and efficiency of railway transportation.

[0062] In this embodiment, to facilitate understanding of how to perform unified spatiotemporal preprocessing, the method for obtaining unified spatiotemporal multimodal feature sequences includes:

[0063] Dynamic time warping The algorithm performs time-axis alignment on time-series data with different sampling frequencies. Assume there are two time series that need to be spatiotemporally aligned. and The calculation formula is as follows:

[0064] Formula 1;

[0065] In Equation 1, , These are two time series to be aligned (e.g.) It can be a sequence of performance indicators for network elements. This can be an alarm event sequence, etc., and can be aligned according to actual processing needs. Two time series and The dynamic time-normalized distance between them; Time series The first in Observations at each time point; Time series The first in Observations at each time point; Time series No. Time points With time series The first in Time points Local distance between them; To satisfy the optimal alignment path that meets monotonicity, continuity, and boundary conditions, in this embodiment, after completing time axis alignment, spatial alignment processing is also required. Since data from different sources may have different spatial reference frames or resolutions, methods such as spatial interpolation and coordinate transformation are needed to unify the data under the same spatial reference frame, ensuring consistency in spatial dimensions. For example, environmental monitoring data from different monitoring stations can be interpolated onto a unified grid using spatial interpolation methods for subsequent analysis. After completing the spatiotemporal alignment preprocessing, a unified spatiotemporal multimodal feature sequence can be obtained.

[0066] In this embodiment, to ensure the accuracy and effectiveness of multi-level fusion of the aforementioned unified spatiotemporal multimodal feature sequences, the multi-level fusion includes one or more combinations of temporal feature-level fusion, cross-modal attention mechanism fusion, and decision-level weighted fusion. Specifically, to facilitate understanding of how to perform multi-level fusion of the unified spatiotemporal multimodal feature sequences, the method for multi-level fusion of the unified spatiotemporal multimodal feature sequences includes:

[0067] Based on railway operation and maintenance requirements, a unified spatiotemporal multimodal feature sequence is processed by feature concatenation to obtain a preliminary fused feature vector. Specifically, the temporal feature-level fusion steps involve concatenating features from each modality within the same time window, using weighted averaging or principal component analysis for dimensionality reduction, to obtain a preliminary fused feature vector. This temporal feature-level fusion employs a sliding window technique to extract temporal features, and then utilizes a convolutional neural network to perform preliminary fusion of the extracted temporal features, resulting in a temporal feature fusion vector. Specifically, a fixed-length sliding window is set and slides across the unified spatiotemporal multimodal feature sequence to extract temporal features within each window, such as mean, variance, maximum, and minimum values. These temporal features are then input into a convolutional neural network, where convolutional layers, pooling layers, and other operations are used to perform preliminary fusion and dimensionality reduction, yielding a temporal feature fusion vector. This vector can preliminarily reflect the temporal variation characteristics of the railway communication system.

[0068] A multi-head attention mechanism is used to interactively model features from different modalities in the initial fused feature vectors, completing the cross-modal attention mechanism fusion. The core calculation process is as follows:

[0069] Formula 2;

[0070] Formula 3;

[0071] In equations 2 and 3, , , These are the query, key, and value matrices, obtained by linear projection of different modal features, such as... It can be obtained from the text modality. , It can be obtained from images, audio, or other modalities. For each dimension of attention head, For the number of attention heads, To output the projection matrix; This is the normalization function; The process involves a splicing operation; specifically, the steps of this cross-modal attention mechanism fusion are as follows: a multi-head attention mechanism can be used to interactively model features from different modalities, achieving deep fusion of cross-modal features and enhancing the richness and accuracy of feature representation. Specifically, this cross-modal attention mechanism fusion constructs an attention mechanism model to weightedly fuse features from different modalities, highlighting the contribution of key modal features to the overall health representation. Specifically, firstly, different modal features (such as network element performance indicators and alarm event features) in the unified spatiotemporal multimodal feature sequence are processed independently to extract their respective key features. Then, an attention mechanism model is constructed, which automatically assigns different weights based on the degree of influence of different modal features on the state of the railway communication system. Through the calculation of the attention mechanism model, the weighted modal features are obtained, and these weighted features are fused to obtain a cross-modal attention mechanism fusion vector. This vector can more accurately reflect the overall state of the railway communication system under different modalities.

[0072] After completing the cross-modal attention mechanism fusion, based on the railway communication operation and maintenance requirements, corresponding weights are assigned to the fusion results, and a comprehensive health representation vector is obtained through weighted summation. That is, decision-level weighted fusion can, based on the results of temporal feature-level fusion and cross-modal attention mechanism fusion, combined with expert experience or preset rules, finally fuse the comprehensive health representation vector information to obtain the current railway communication operation status and fault probability distribution information. Specifically, based on the temporal feature-level fusion vector and the cross-modal attention mechanism fusion vector, combined with historical operation data of the railway communication system and expert experience, different weight allocation rules are set. For example, for areas with large temporal feature variations and high weights for key modal features in the cross-modal attention mechanism, these areas can be considered to have a high fault risk and are assigned a larger weight. Through decision-level weighted fusion, the results of temporal feature-level fusion and cross-modal attention mechanism fusion are combined to obtain the current railway communication operation status (e.g., normal, abnormal, faulty, etc.) and fault probability distribution information (e.g., fault probability of each area or each piece of equipment). This information can provide maintenance personnel with intuitive and accurate decision-making basis, guiding them to take timely measures to ensure the stable operation of the railway communication system.

[0073] In this embodiment, to facilitate understanding of how the current railway communication operation status and fault probability distribution information are determined based on the multi-level fusion of the unified spatiotemporal multimodal feature sequences described above, the following is used: Distributed fiber optic sensing monitoring is used as an example to illustrate this. Specifically, the method for obtaining the current operating status and fault probability distribution information of railway communication includes:

[0074] Collectable Vibration disturbance data from distributed optical fiber sensing signals can be collected by acquiring Φ-OTDR backscattered Rayleigh signals to form a spatiotemporal two-dimensional disturbance feature signal. This spatiotemporal two-dimensional disturbance feature signal is then preprocessed with denoising to obtain a denoised phase signal. Subsequently, intrusion event detection and spatiotemporal region segmentation are performed based on the denoised phase signal to obtain at least one event spatiotemporal block. Then, multi-domain spatiotemporal features (including at least one of time-domain, frequency-domain, time-frequency, and spatiotemporal features) are extracted from each event spatiotemporal block to form a feature vector. Based on railway communication operation and maintenance requirements, pattern recognition can be performed on the feature vector to obtain the intrusion event detection category. High-precision positioning is then performed on the event spatiotemporal blocks identified as genuine intrusion events to obtain the intrusion event location. In this embodiment, based on multi-source heterogeneous data information from railway communication (such as Φ-OTDR backscattered Rayleigh signals), a deep learning diagnostic model using a Transformer architecture with an attention mechanism or a convolutional neural network architecture can be used to diagnose the comprehensive health representation vector information. The output layer adopts... The function obtains the classification results of the current railway communication operation status and the probability distribution of corresponding fault / intrusion events. It should be noted that the classification results can be divided into categories according to the actual situation, such as excavator, manual digging / knocking, vehicle passing, mechanical construction, natural wind noise, rain noise, and no event. That is, through detailed analysis of the comprehensive health representation vector information, it is possible to accurately distinguish different types of events, whether caused by human activities or natural environmental factors.

[0075] Specifically, the method for forming a two-dimensional spatiotemporal perturbation feature signal includes:

[0076] Let the backscattered Rayleigh signal acquired by the Φ-OTDR be... At this point, the phase vibration signal is obtained by differential phase demodulation between adjacent pulses, and the phase change is calculated:

[0077] Equation 4;

[0078] in, This represents the amount of phase change due to the disturbance. These are the spatial coordinates on the optical fiber; The time coordinates for signal acquisition; For complex conjugate; Spatial sampling interval; To obtain the phase angle, this step converts the intensity signal into a sensitive phase vibration signal. This improves sensitivity to weak vibrations. Furthermore, in this embodiment, to reduce interference affecting the location accuracy of intrusion events, wavelet denoising can be used on the demodulated phase signal to improve the signal-to-noise ratio. Specifically, the operation is as follows:

[0079] ;

[0080] In the formula, and These are wavelet transform and its inverse transform, respectively. This is a soft thresholding function, and the threshold value is... ; The standard deviation of noise. This is the signal length.

[0081] In this embodiment, the method for obtaining the intrusion event detection category includes:

[0082] The temporal energy of each event spatiotemporal block is calculated using the following formula:

[0083] Formula 5;

[0084] In Equation 5, In fiber optic location and time The local time-domain energy value at that location; The length of the sliding window; For the coordinates of the fiber optic position Location and time coordinates The phase vibration signal at the location, the It can be obtained through calculation using Equation 4;

[0085] Based on the needs of railway communication operation and maintenance, an adaptive threshold is set. Adaptive threshold Used for judgment Whether it exceeds a threshold, thus marking it as a potential intrusion point, specifically, the adaptive threshold. :

[0086] Formula 6;

[0087] In Equation 6, Background noise or time-domain energy under normal conditions The mean; Time-domain energy under background noise The standard deviation; m is an empirical coefficient that can be determined based on the actual situation, and is generally taken as a value. That is, in If so, it can be considered abnormal (there is a potential risk of intrusion).

[0088] In this embodiment, to facilitate understanding of how to accurately locate intrusion risks, the following explanation is provided: the location method can be achieved through a combination of dual-end energy peaks and phase gradients. Specifically, the phase gradient is calculated as follows:

[0089] Formula 7;

[0090] In Equation 7, Position along the fiber Phase gradient in the direction; The denoised phase vibration signal at position The differential phase value at that point; To sample points in adjacent spaces The phase vibration signal value at that location; The spatial sampling interval (i.e., Spatial resolution is typically determined by the optical pulse width and sampling rate.

[0091] The energy peak position is obtained from Equation 7 above. and the position of the phase gradient extremum Then, the final location can be obtained through weighted fusion:

[0092] ;

[0093] In the formula, To pinpoint the location of the intrusion event with high precision. This is a location estimate obtained by calculating the peak location of the signal energy distribution within the spatiotemporal block of the event. To find the phase gradient The location of the maximum or minimum value is determined. Weighting coefficients for the energy peak location results. These are the weighting coefficients for the phase gradient localization results. This is the location correction value. Using the above positioning method, the specific location of an intrusion event on the optical fiber can be accurately determined, providing crucial support for the safe operation and maintenance of railway communication systems. In practical applications, this positioning method is not only suitable for detecting single intrusion events but can also effectively handle multiple simultaneous intrusion events. By distinguishing the spatiotemporal characteristics of different events, parallel positioning and identification of multiple events can be achieved.

[0094] In a second aspect, this invention also provides a railway communication operation and maintenance monitoring system, employing the railway communication operation and maintenance monitoring method described in any one of the first aspects. The monitoring system further includes: a data acquisition module, a preprocessing module, a fusion module, a monitoring and diagnosis module, and an early warning operation and maintenance module. The data acquisition module acquires multi-source heterogeneous data information from railway communication in real time. The preprocessing module performs spatiotemporal alignment and normalization preprocessing on the multi-source heterogeneous data information from railway communication transmitted from the data acquisition module to obtain a unified spatiotemporal multimodal feature sequence. The fusion module performs multi-level fusion on the unified spatiotemporal multimodal feature sequence to obtain comprehensive health representation vector information. The monitoring and diagnosis module monitors and diagnoses the comprehensive health representation vector information transmitted from the fusion module to obtain the current operating status and fault probability distribution information of the railway communication. The early warning operation and maintenance module determines the current railway communication operation and maintenance early warning information based on the current operating status and fault probability distribution information of the railway communication. In this embodiment, the railway communication operation and maintenance monitoring system acquires various multi-source heterogeneous data from railway communication in real time through the data acquisition module, covering network element equipment performance indicators, alarm events, signaling tracking, and optical transmission layer data. The monitoring module collects data on optical power, environmental and energy consumption, providing a comprehensive and detailed foundation for subsequent precise analysis. The preprocessing module performs spatiotemporal alignment and normalization on the aforementioned monitoring data, effectively resolving the differences in time and space between different data sources and generating a unified spatiotemporal multimodal feature sequence, ensuring data consistency and comparability. The fusion module employs multi-level fusion strategies, including time-domain feature-level fusion, cross-modal attention mechanism fusion, and decision-level weighted fusion, to deeply mine the inherent correlations and complementary information between data, extracting comprehensive health representation vector information that reflects the current operating status and potential fault risks of the railway communication system. The monitoring and diagnosis module, through in-depth analysis of the comprehensive health representation vector information, can not only accurately determine the operating status of railway communication but also predict the probability distribution of faults or intrusion events, providing a scientific basis for operation and maintenance decisions. Finally, the early warning and maintenance module generates and releases timely operation and maintenance early warning information based on the monitoring and diagnosis results, guiding maintenance personnel to take targeted measures to effectively prevent and respond to potential faults, ensuring the stable operation of the railway communication system.

[0095] In some implementations, the monitoring system can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed networks.

[0096] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), etc.

[0097] A third aspect of the present invention provides a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements a railway communication operation and maintenance monitoring method as described in any one of the first aspects. The computer-readable medium in this embodiment can be written in one or more programming languages ​​or a combination thereof to perform computer program code for carrying out operations of some embodiments of the present disclosure. These programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—and conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on a user's computer, partially on a user's computer, as a standalone software package, partially on a user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0098] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0099] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts.

[0100] A fourth aspect of the present invention provides an electronic device, comprising: one or more processors; a storage device storing one or more programs thereon; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement a railway communication operation and maintenance monitoring method as described in the first aspect. The computer-readable medium may be included in the electronic device or may exist independently, i.e., not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, enable the electronic device to implement the railway communication operation and maintenance monitoring method as described in the first aspect.

[0101] The fifth aspect of the present invention provides a computer program product, including a computer program that, when executed by a processor, implements a railway communication operation and maintenance monitoring method as described in the first aspect.

[0102] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A railway communication operation and maintenance monitoring method, characterized in that, include: Based on the railway communication operation and maintenance needs, collect multi-source heterogeneous data information of railway communication in real time; Spatiotemporal alignment preprocessing is performed on multi-source heterogeneous data information from railway communications to obtain a unified spatiotemporal multimodal feature sequence; Multi-level fusion of unified spatiotemporal multimodal feature sequences is performed to obtain comprehensive health representation vector information; Based on the railway communication operation and maintenance requirements, the comprehensive health representation vector information is monitored and diagnosed to obtain the current railway communication operation status and fault probability distribution information. Based on the railway communication operation and maintenance requirements, and combined with the current railway communication operation status and fault probability distribution information, determine the current railway communication operation and maintenance early warning information; The method for obtaining the current operating status and fault probability distribution information of railway communication includes: Based on the multi-source heterogeneous data information of railway communication, the comprehensive health characterization vector information is diagnosed to obtain the current railway communication operation status classification results and the probability distribution of corresponding fault / intrusion events; The multi-source heterogeneous data information includes: based on Vibration disturbance data from distributed fiber optic sensing signals are used for precise identification and location of intrusion events along railway lines; Specifically, the method for obtaining the current railway communication operation status classification results and the probability distribution of corresponding fault / intrusion events includes: The backscattered Rayleigh scattering signal from the Φ-OTDR is collected to form a two-dimensional spatiotemporal perturbation characteristic signal; The spatiotemporal two-dimensional perturbation feature signal is subjected to denoising preprocessing to obtain the denoised phase signal; Intrusion event detection and spatiotemporal region segmentation are performed based on the denoised phase signal to obtain at least one event spatiotemporal block; For each event spatiotemporal block, multi-domain spatiotemporal features are extracted to form a feature vector; Based on the railway communication operation and maintenance requirements, pattern recognition is performed on the feature vectors to obtain the intrusion event detection category; High-precision localization is performed on the spatiotemporal blocks of events identified as real intrusion events to obtain the location of the intrusion events.

2. The railway communication operation and maintenance monitoring method according to claim 1, characterized in that, The method for obtaining a unified spatiotemporal multimodal feature sequence includes: Dynamic time warping The algorithm performs time-axis alignment on time-series data with different sampling frequencies. Assume there are two time series that need to be spatiotemporally aligned. and The calculation formula is as follows: Formula 1; In Equation 1, Two time series and The dynamic time-normalized distance between them; Time series The first in Observations at each time point; Time series The first in Observations at each time point; Time series No. Time points With time series The first in Time points Local distance between them; The optimal alignment path is determined to satisfy monotonicity, continuity, and boundary conditions.

3. The railway communication operation and maintenance monitoring method according to claim 1, characterized in that, The multi-level fusion includes one or more of the following: temporal feature-level fusion, cross-modal attention mechanism fusion, and decision-level weighted fusion.

4. The railway communication operation and maintenance monitoring method according to claim 3, characterized in that, The method for multi-level fusion of unified spatiotemporal multimodal feature sequences includes: Based on railway operation and maintenance requirements, feature splicing processing is performed on unified spatiotemporal multimodal feature sequences to obtain preliminary fused feature vectors. A multi-head attention mechanism is used to interactively model features from different modalities on the initially fused feature vectors, completing the cross-modal attention mechanism fusion. The calculation process is as follows: Formula 2; Formula 3; In equations 2 and 3, , , These are the query, key, and value matrices, obtained by linear projection of different modal features. For each dimension of attention head, For the number of attention heads, To output the projection matrix; This is the normalization function; For splicing operations; After completing the cross-modal attention mechanism fusion, according to the railway communication operation and maintenance requirements, the cross-modal attention mechanism fusion results are assigned corresponding weights, and the comprehensive health representation vector information is obtained by weighted summation.

5. A railway communication operation and maintenance monitoring method according to claim 4, characterized in that, The method for forming a spatiotemporal two-dimensional perturbation feature signal includes: Let the backscattered Rayleigh signal acquired by the Φ-OTDR be... At this point, the phase vibration signal is obtained by differential phase demodulation between adjacent pulses, and the phase change is calculated: Equation 4; in, This represents the amount of phase change due to the disturbance. These are the spatial coordinates on the optical fiber; The time coordinates for signal acquisition; For complex conjugate; Spatial sampling interval; To obtain the phase angle.

6. A railway communication operation and maintenance monitoring method according to claim 5, characterized in that, The method for obtaining the intrusion event detection category includes: The temporal energy of each event spatiotemporal block is calculated using the following formula: Formula 5; In Equation 5, In fiber optic location and time The local time-domain energy value at that location; The length of the sliding window; For the coordinates of the fiber optic position Location and time coordinates Phase vibration signal at the location; Based on the needs of railway communication operation and maintenance, an adaptive threshold is set. , wherein the adaptive threshold for: Formula 6; In Equation 6, Background noise or time-domain energy under normal conditions The mean; Time-domain energy under background noise The standard deviation of ; m is the empirical coefficient.

7. A railway communication operation and maintenance monitoring system, characterized in that, The railway communication operation and maintenance monitoring method according to any one of claims 1 to 6 is adopted, wherein the detection system further includes: The acquisition module collects multi-source heterogeneous data information from railway communications in real time. The preprocessing module is used to perform spatiotemporal alignment and normalization preprocessing on the multi-source heterogeneous data information of railway communication transmitted from the acquisition module to obtain a unified spatiotemporal multimodal feature sequence. The fusion module is used to perform multi-level fusion of unified spatiotemporal multimodal feature sequences to obtain comprehensive health representation vector information; The monitoring and diagnosis module is used to monitor and diagnose the comprehensive health representation vector information transmitted from the fusion module, and obtain the current operating status and fault probability distribution information of railway communication. The early warning and maintenance module is used to determine the current railway communication operation and maintenance early warning information based on the current operating status and fault probability distribution information of the railway communication.

8. A computer-readable medium having a computer program stored thereon, characterized in that, in, When the program is executed by the processor, it implements a railway communication operation and maintenance monitoring method as described in any one of claims 1 to 6.

Images