Train positioning and tunnel safety real-time monitoring method and system, storage medium

By deploying fiber optic sensor networks, acoustic fingerprint sensors and geomagnetic field sensors in ultra-long tunnels in high altitude areas, combined with machine learning algorithms, the problems of train positioning and tunnel safety monitoring are solved, and precise positioning and full-process safety monitoring are achieved, improving emergency response speed and safety.

CN118182574BActive Publication Date: 2024-11-19CHINA RAILWAY TUNNEL GROUP CO LTD +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410372910.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-03-29
Publication Date
2024-11-19
Estimated Expiration
2044-03-29

AI Technical Summary

Technical Problem

In ultra-long tunnels in high altitude areas, it is difficult for the prior art to achieve accurate positioning of trains and full-course tunnel safety monitoring, especially when GPS signals cannot penetrate deep underground and there are blind spots in traditional monitoring systems, resulting in discontinuous positioning and timely discover and respond to safety hazards and difficulties.

Method used

Using fiber optic sensing network, acoustic fingerprint sensor and geomagnetic field sensor combined with modern sensing technology and machine learning algorithms, a monitoring network is established for environmental data collection and preprocessing, and using pre-trained models to evaluate train positioning and tunnel safety status to generate real-time Monitor data and security alerts.

Benefits of technology

It realizes accurate positioning of trains in tunnels with unreachable GPS signals and full-process tunnel safety monitoring, eliminates monitoring blind spots, responds to emergencies in a timely manner, and significantly improves the safety and efficiency of train operations in the tunnel.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118182574B_ABST
    Figure CN118182574B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of extreme environment train and tunnel safety, and in particular to a train positioning and tunnel safety real-time monitoring method and system, and a storage medium. The method comprises the following steps: setting monitoring points in a tunnel and deploying an optical fiber sensor network, an acoustic wave fingerprint sensor, and a geomagnetic field sensor, establishing a monitoring network layout, collecting environmental monitoring data through an initialized monitoring network, and preprocessing the environmental monitoring data, wherein the environmental monitoring data comprises: acoustic wave reflection and temperature change data, acoustic wave fingerprint data, and geomagnetic field change data; using a pre-trained first model, taking the environmental monitoring data as input, and outputting train positioning information; based on the current train positioning information and environmental monitoring data, predicting the train position and evaluating the tunnel safety status through a pre-trained second model, generating tunnel real-time monitoring data and safety alarms. The present invention significantly improves the safety and efficiency of train operation in the tunnel.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of extreme environment train and tunnel safety technology, and in particular to a train positioning and tunnel safety real-time monitoring method and system, and a storage medium. Background Art

[0002] In my country's Qinghai-Tibet Plateau, especially in an environment with an altitude of 3,000-5,000 meters, building and operating ultra-long train tunnels (over 36 kilometers) is of great strategic significance and challenge. These tunnels not only provide a feasible path for crossing complex terrain, but also greatly shorten the transportation distance between regions, promote regional economic development and the rapid flow of people. However, in such a plateau environment, train positioning and tunnel safety face unique problems and challenges.

[0003] In super-long tunnels, conventional GPS signals cannot penetrate deep underground, making it difficult to locate trains. Accurate train positioning is crucial to ensuring that trains pass through tunnels efficiently and safely, especially in emergency situations where it is necessary to quickly and accurately locate the train for rescue. There may be a variety of safety hazards inside super-long tunnels, including but not limited to fire, poor ventilation, structural damage, etc. These safety issues may be caused by a variety of reasons, such as equipment failure, improper maintenance, natural disasters, etc. In plateau areas, complex geological conditions and extreme weather may also exacerbate these safety hazards.

[0004] Although there are many solutions (such as Chinese invention patent, publication number: CN115923879A) that attempt to solve the problems of train positioning and tunnel safety, these solutions often face the following deficiencies or defects in the specific environment of ultra-long tunnels in plateau areas:

[0005] Existing technologies (such as positioning systems based on ground signals) often have difficulty covering the entire length of an ultra-long tunnel, resulting in discontinuous positioning information and the inability to achieve accurate positioning throughout the tunnel.

[0006] Traditional tunnel monitoring technologies (such as video surveillance) may have blind spots due to limited vision in the tunnel or the inability of monitoring equipment to cover the entire tunnel, making it difficult to detect and respond to safety hazards in a timely manner.

[0007] In the event of an emergency, existing solutions often have difficulty conveying alarms and location information in real time, resulting in delayed response measures and increased safety risks. Summary of the invention

[0008] In view of the many problems existing in the above-mentioned prior art when building and operating trains in high-altitude and ultra-long tunnels, the present invention provides a train positioning and tunnel safety real-time monitoring method and system, and storage medium. The present invention combines modern sensing technology and machine learning algorithms to provide a comprehensive, accurate and real-time train positioning and tunnel safety monitoring solution for the special challenges in the ultra-long tunnel environment in plateau areas, which significantly improves the safety and efficiency of train operation in tunnels.

[0009] A method for real-time monitoring of train positioning and tunnel safety comprises the following steps:

[0010] Setting monitoring points in the tunnel and deploying optical fiber sensor networks, acoustic fingerprint sensors and geomagnetic field sensors, establishing the monitoring network layout, collecting environmental monitoring data through the initialized monitoring network, and pre-processing the environmental monitoring data, the environmental monitoring data including: acoustic reflection and temperature change data, acoustic fingerprint data and geomagnetic field change data;

[0011] Using the pre-trained first model, the environmental monitoring data is used as input and the train positioning information is output;

[0012] Based on the current train positioning information and environmental monitoring data, the pre-trained second model is used to predict the train position and evaluate the tunnel safety status, generating real-time tunnel monitoring data and safety alerts.

[0013] Preferably, a fiber optic sensor network is deployed in the tunnel to monitor the changes in sound waves and temperature in real time, capture the changes in light signals caused by sound wave reflection and temperature respectively, and obtain sound wave reflection and temperature distribution data; an acoustic fingerprint sensor is used to collect the unique characteristics of the sound wave echoes generated by the movement of the train in the tunnel to generate acoustic fingerprint data; and a geomagnetic field sensor is used to regularly measure the changes in the geomagnetic field intensity in the tunnel to capture the geomagnetic field disturbance caused by the metal body of the train to obtain geomagnetic field change data.

[0014] Preferably, the preprocessing of the environmental monitoring data includes:

[0015] The preprocessing of acoustic wave reflection and temperature change data includes: applying digital signal processing technology to remove background noise and extract effective acoustic wave reflection signals and temperature change signals; normalizing the acoustic wave reflection intensity and temperature data to eliminate the influence of differences between devices and measurement range; using signal enhancement technology to enhance the characteristics of acoustic wave signals;

[0016] The preprocessing of acoustic fingerprint data includes: extracting acoustic features from the collected acoustic echo data, compressing the acoustic feature data and identifying and processing abnormal values, wherein the acoustic features include: frequency, amplitude and echo delay;

[0017] The preprocessing of geomagnetic field change data includes: performing baseline calibration on geomagnetic field data based on the measured values ​​of geomagnetic field strength at monitoring points when no trains pass, so as to measure the geomagnetic field changes caused by trains; applying filtering technology to remove short-term fluctuations and electromagnetic interference in geomagnetic field data, and retaining the geomagnetic field change signals caused by trains; analyzing the time series trend of geomagnetic field data, and identifying the geomagnetic field change patterns caused by train movement.

[0018] Preferably, the training of the first model includes:

[0019] Collect historical environmental monitoring data and corresponding actual train location data, and mark the relationship between the environmental monitoring data and the actual train location, wherein the historical environmental monitoring data includes: historical environmental monitoring data and pre-processed data thereof;

[0020] Extract key features from the data, the key features including: temperature change trend; frequency, amplitude, and delay of the acoustic wave signal; and intensity change of the geomagnetic field;

[0021] The first model is trained using the key features as a training set to learn the relationship between the data features and the train positions.

[0022] Preferably, the first model predicts the position of the train based on new environmental monitoring data by learning the relationship between environmental monitoring data and the train position; in the prediction stage, the first model receives new environmental monitoring data input, and outputs corresponding train positioning information through the rules and relationships learned in the training stage, and the train positioning information includes: the position coordinates and travel speed of the train.

[0023] Preferably, the training of the second model includes:

[0024] Through the current train positioning information and environmental monitoring data, historical train positioning information and environmental monitoring data, historical train operation mode and special event history records, each data instance is annotated with the predicted train position and tunnel safety status, and then organized as the training set of the second model, where:

[0025] The historical train operation mode is used to help the second model understand the typical operation trajectory and change law of the train. The historical train operation mode includes:

[0026] The average speed and position changes of trains in different time periods;

[0027] The train’s behavior patterns in response to different environmental conditions;

[0028] The impact of specific times on train speed and route selection;

[0029] The special event history records are used to help the second model identify special situations that affect train safety and positioning accuracy, and the special event history records include: historical safety accidents and maintenance records in the tunnel;

[0030] The second model uses the training set to learn to predict the future position of the train and assess the safety status of the tunnel.

[0031] Preferably, the second model predicts the position of the train at a specific time point in the future by analyzing the learned spatiotemporal relationship and train operation mode, using current train positioning information and environmental monitoring data and historical train operation mode;

[0032] The second model predicts safety hazards in the tunnel and evaluates the safety status of the tunnel by analyzing the current environmental monitoring data and the predicted position of the train, combined with the historical records of special events;

[0033] The generated tunnel real-time monitoring data and security alerts are output.

[0034] A system for executing the train positioning and tunnel safety real-time monitoring method, comprising:

[0035] A data acquisition module, wherein the data acquisition module comprises:

[0036] The fiber-optic sensor network is deployed in the tunnel to capture the changes in light signals caused by sound wave reflection and temperature, providing real-time sound wave reflection and temperature distribution data;

[0037] Acoustic fingerprint sensors, installed at specific locations in the tunnel, are used to collect the unique characteristics of the acoustic echoes generated by the movement of the train and generate acoustic fingerprint data for identifying and locating the train's position;

[0038] Geomagnetic field sensors are arranged at various monitoring points in the tunnel to regularly measure the geomagnetic field change data caused by the geomagnetic field disturbance caused by the metal body of the train in the tunnel;

[0039] The data processing center, as the operating carrier of the first model and the second model, is used to receive the environmental monitoring data collected from the optical fiber sensor network, the acoustic fingerprint sensor and the geomagnetic field sensor, process it through the first model, and output the train positioning information; based on the current train positioning information and environmental monitoring data, the second model is used to predict the train position and evaluate the tunnel safety status, and output the real-time monitoring data and safety alarm of the tunnel.

[0040] A storage medium stores a computer program, which, when executed by a processor, implements the steps of the train positioning and tunnel safety real-time monitoring method.

[0041] Compared with the prior art, the advantages and beneficial effects of the present invention are:

[0042] By using environmental monitoring data and machine learning models, the present invention can accurately locate trains in tunnels where GPS signals cannot reach, ensuring the continuity and accuracy of train location information;

[0043] Through the monitoring network deployed throughout the whole process and combined with multi-source data fusion technology, the present invention can comprehensively monitor the safety status in the tunnel, effectively eliminate monitoring blind spots, and discover safety hazards in real time;

[0044] Based on real-time monitoring data and a safety alarm system, the present invention can respond promptly and accurately when an emergency is discovered, greatly shortening the time for emergency handling. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the process of the present invention;

[0046] Figure 2 It is a schematic diagram of the data processing flow in the present invention;

[0047] Figure 3 A schematic diagram of the functional relationship and data flow between the first model and the second model in the present invention;

[0048] Figure 4 It is a structural block diagram of the system of the present invention. DETAILED DESCRIPTION

[0049] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is obvious that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0050] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0051] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0052] In the case of using expressions such as "at least one of A, B, and C, etc.", it should generally be interpreted in accordance with the meaning of the expression generally understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.). In the case of using expressions such as "at least one of A, B, or C, etc.", it should generally be interpreted in accordance with the meaning of the expression generally understood by those skilled in the art (for example, "a system having at least one of A, B, or C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0053] Some block diagrams and / or flow charts are shown in the accompanying drawings. It should be understood that some boxes or combinations thereof in the block diagrams and / or flow charts can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that these instructions can create a device for implementing the functions / operations described in these block diagrams and / or flow charts when executed by the processor. The technology of the present disclosure can be implemented in the form of hardware and / or software (including firmware, microcode, etc.). In addition, the technology of the present disclosure can take the form of a computer program product on a computer-readable storage medium storing instructions, which can be used by an instruction execution system or used in conjunction with an instruction execution system.

[0054] like Figure 1 - Figure 3 As shown, a method for real-time monitoring of train positioning and tunnel safety includes the following steps:

[0055] Setting monitoring points in the tunnel and deploying optical fiber sensor networks, acoustic fingerprint sensors and geomagnetic field sensors, establishing the monitoring network layout, collecting environmental monitoring data through the initialized monitoring network, and pre-processing the environmental monitoring data, the environmental monitoring data including: acoustic reflection and temperature change data, acoustic fingerprint data and geomagnetic field change data;

[0056] In high altitude and ultra-long tunnel environments, traditional train positioning and tunnel safety monitoring methods are subject to multiple limitations. Therefore, the method of combining the optical fiber sensor network, the acoustic fingerprint sensor and the geomagnetic field sensor of the present invention has its unique applicability and advantages.

[0057] Utilizing fiber optic sensing technology, distributed acoustic detection (DAS) and distributed temperature detection (DTS) can realize continuous monitoring of acoustic waves and temperature changes in tunnels. The optical signal in the optical fiber line is affected by physical changes (such as acoustic vibrations and temperature changes), which causes changes in the reflectivity and refractive index of light, so that the distribution and changes of acoustic waves and temperature can be accurately monitored. Optical fiber cables are laid along the tunnel, and the optical time domain reflectometer (OTDR) technology in the optical fiber monitors and analyzes the optical fiber sensing signals in real time to obtain acoustic wave and temperature data. Realizing continuous and real-time monitoring of the entire length of the tunnel and providing accurate acoustic wave and temperature change data are crucial for identifying abnormal events and environmental changes in the tunnel.

[0058] The acoustic fingerprint sensor captures the acoustic echo characteristics generated by the movement of the train and uses these unique acoustic characteristics (such as frequency, amplitude, and echo delay) to locate the train. When the sound wave propagates in the tunnel, it will produce a specific echo pattern due to the specific structure of the tunnel and the movement state of the train, forming a unique "acoustic fingerprint". Acoustic fingerprint sensors are installed at key locations in the tunnel to capture the acoustic signals generated by the train passing through the tunnel, and these signals are analyzed to extract the acoustic fingerprint data. The uniqueness of the acoustic fingerprint can accurately identify the location of the train, which is particularly suitable for tunnel environments that cannot be covered by GPS signals.

[0059] The geomagnetic field sensor locates the train by measuring the changes in the magnetic field in the tunnel. The metal structure of the train will disturb the surrounding geomagnetic field when it moves, and these changes can be captured by the geomagnetic field sensor. Geomagnetic field sensors are arranged along the tunnel to regularly measure and record the geomagnetic field strength and analyze the changes in the geomagnetic field caused by the movement of the train. This provides an effective means for train positioning, especially when combined with acoustic fingerprints and temperature change data for comprehensive analysis, which can greatly improve the accuracy and reliability of positioning.

[0060] Preferably, a fiber optic sensor network is deployed in the tunnel to monitor the changes in sound waves and temperature in real time, capture the changes in light signals caused by sound wave reflection and temperature respectively, and obtain sound wave reflection and temperature distribution data; an acoustic fingerprint sensor is used to collect the unique characteristics of the sound wave echoes generated by the movement of the train in the tunnel to generate acoustic fingerprint data; and a geomagnetic field sensor is used to regularly measure the changes in the geomagnetic field intensity in the tunnel to capture the geomagnetic field disturbance caused by the metal body of the train to obtain geomagnetic field change data.

[0061] The sensitivity of fiber-optic sensor networks to monitor temperature changes in high-altitude ultra-long tunnels is affected by many factors, the key being the unique physical and climatic conditions in this environment. Temperatures in high-altitude areas vary greatly, and climatic conditions are complex and changeable, all of which may affect the performance of fiber-optic sensor networks. Fiber-optic sensing technology uses changes in optical signals in optical fibers to sense physical changes in the outside world (such as temperature and sound waves). Temperature changes can cause changes in the physical properties of optical fiber materials (such as changes in refractive index), which in turn affects optical signals. In high-altitude areas, especially in environments with large temperature differences, the sensitivity of optical fibers to temperature is particularly critical, because rapid changes in temperature may cause rapid fluctuations in fiber-optic sensor data, thereby affecting the accurate interpretation of the data.

[0062] Although the high-altitude ultra-long tunnel environment poses challenges to the temperature monitoring of the fiber optic sensor network, these challenges can be overcome through the selection and optimization of technology and materials. For example, specially designed fiber optic sensors can be used to improve the monitoring accuracy of temperature changes, or data processing algorithms can be used to correct monitoring errors caused by extreme temperature changes to ensure data accuracy and reliability.

[0063] The fiber optic sensor network uses the sensitivity of light to environmental changes (such as acoustic vibration and temperature changes) when it propagates in the optical fiber, and detects acoustic waves and temperature by monitoring the changes in optical signals. When acoustic vibration or temperature changes act on the optical fiber, it will cause changes in the phase, intensity or polarization state of the light in the optical fiber, and these changes are detected by equipment such as optical time domain reflectometer (OTDR). Fiber optic sensor cables are arranged along the tunnel, and the entire fiber optic network covers the entire length of the tunnel. Use equipment such as OTDR to monitor the optical signal in the optical fiber in real time, record the changes in the optical signal, and then analyze the data of acoustic wave reflection and temperature distribution. The fiber optic sensor network can realize continuous monitoring of the entire length of the tunnel and provide accurate data on acoustic waves and temperature changes. This is crucial for detecting abnormal conditions inside the tunnel (such as acoustic abnormalities caused by structural damage and temperature increases caused by fire), and has a significant effect on maintaining tunnel safety and preventing accidents.

[0064] The acoustic fingerprint sensor locates the train based on the unique echo pattern generated when the sound waves propagate in the tunnel. Each tunnel structure and the sound wave propagation characteristics when the train passes through are different, forming a kind of "acoustic fingerprint". By analyzing these acoustic fingerprints, the position of the train can be accurately identified. Acoustic fingerprint sensors are installed at key locations in the tunnel. These sensors are specially designed to capture the acoustic echoes generated when the train passes. The acoustic wave data collected by the sensor is sent to the analysis system, and the acoustic fingerprint data is generated by analyzing the acoustic wave characteristics (frequency, amplitude, delay). The acoustic fingerprint sensor provides a way to accurately locate the train in the tunnel without external signals. This is especially important in the tunnel environment where GPS signals are not accessible, and enhances the accuracy and reliability of train positioning.

[0065] Geomagnetic field sensors monitor the position of trains by measuring changes in the magnetic field within the tunnel. The metal structure of the train affects the surrounding geomagnetic field during movement, causing detectable changes in the geomagnetic field. Geomagnetic field sensors are deployed in the tunnel, and these sensors regularly measure the strength of the geomagnetic field in the tunnel. By analyzing the change data of the geomagnetic field, especially the changes when the train passes, the geomagnetic field disturbance caused by the train can be captured. Geomagnetic field sensors provide another important means for train positioning, especially when used in combination with data from acoustic fingerprint sensors and fiber optic sensor networks, the accuracy of positioning can be greatly improved. In addition, geomagnetic field monitoring is also helpful in identifying other metal structures or obstacles that may exist in the tunnel.

[0066] Preferably, the preprocessing of the environmental monitoring data includes:

[0067] The preprocessing of acoustic wave reflection and temperature change data includes: applying digital signal processing technology to remove background noise and extract effective acoustic wave reflection signals and temperature change signals; normalizing the acoustic wave reflection intensity and temperature data to eliminate the influence of differences between devices and measurement range; using signal enhancement technology to enhance the characteristics of acoustic wave signals;

[0068] The preprocessing of acoustic fingerprint data includes: extracting acoustic features from the collected acoustic echo data, compressing the acoustic feature data and identifying and processing abnormal values, wherein the acoustic features include: frequency, amplitude and echo delay;

[0069] The preprocessing of geomagnetic field change data includes: performing baseline calibration on geomagnetic field data based on the measured values ​​of geomagnetic field strength at monitoring points when no trains pass, so as to measure the geomagnetic field changes caused by trains; applying filtering technology to remove short-term fluctuations and electromagnetic interference in geomagnetic field data, and retaining the geomagnetic field change signals caused by trains; analyzing the time series trend of geomagnetic field data, and identifying the geomagnetic field change patterns caused by train movement.

[0070] In addition, integrating the data collected by the fiber optic sensor network, acoustic fingerprint sensor and geomagnetic field sensor, and applying data fusion technology are the key to achieving efficient train positioning, train position prediction and tunnel safety monitoring. Data integration and data fusion include:

[0071] Data integration strategy:

[0072] (1) Unified data format: First, the data collected by various sensors needs to be converted into a unified format, including a unified timestamp format, data unit, and data structure. This step is the prerequisite for ensuring that the data can be smoothly integrated and processed.

[0073] (2) Time synchronization: Ensure that the time of all sensor systems is precisely synchronized. This is critical for accurately correlating data from different sources, especially when dynamically analyzing train positions and tunnel safety conditions.

[0074] (3) Spatial calibration: Calibrate the specific location of each sensor in the tunnel to ensure that the spatial information of the data collection point is accurate. This helps to accurately understand the source of the data and the relevant spatial relationship in subsequent analysis.

[0075] Data fusion technology application:

[0076] (1) Hierarchical data processing: The data fusion process can be divided into several levels, from simple data preprocessing (such as denoising and normalization) to complex pattern recognition and machine learning. At the initial level, each sensor data is processed independently to remove noise and extract features. Then, the processed data from different sensors are merged.

[0077] (2) Feature fusion: The feature data from different sensors are integrated, including acoustic reflection characteristics, temperature change trends, and geomagnetic field change patterns. Feature fusion aims to integrate information provided by different data sources to provide a more comprehensive perspective for train positioning and tunnel safety assessment.

[0078] (3) Model-level fusion: Use machine learning models (such as the first model and the second model) to process the fused feature data. These models can learn the relationships and patterns between different data features, thereby achieving more accurate train positioning, train position prediction, and tunnel safety status assessment. Model-level fusion can use ensemble learning methods, such as random forests or deep learning techniques, to enhance the accuracy and robustness of predictions.

[0079] (4) Decision-level fusion: In the final decision-making stage, the outputs of different models can be integrated by using methods such as voting and weighted averaging to comprehensively consider the prediction results of multiple models to obtain the final train positioning information, train position prediction, and tunnel safety status assessment.

[0080] It should be noted that the environmental monitoring data of the present invention needs to be integrated as described above after preprocessing, and data fusion is an essential step when the environmental monitoring data is used as the input of the first model and the second model. Through the application of the above data fusion technology, the present invention can make full use of the advantages of each monitoring technology, make up for the shortcomings of a single data source, and improve the accuracy of train positioning and the efficiency of train position prediction and tunnel safety monitoring. Data fusion not only enhances the real-time tracking capability of the dynamic position of the train, but also enhances the early warning and response capabilities for potential safety hazards in the tunnel, providing strong technical support for train operation and tunnel safety management in high-altitude ultra-long tunnels. The application of data fusion of environmental monitoring data in the present invention will be described in detail below.

[0081] Using the pre-trained first model, the environmental monitoring data is used as input and the train positioning information is output;

[0082] The core challenge of achieving train positioning in high-altitude and ultra-long tunnels is the inapplicability of traditional GPS systems. Therefore, the present invention uses environmental monitoring data (acoustic wave reflection and temperature change data, acoustic wave fingerprint data, and geomagnetic field change data) to locate trains, which becomes an innovative and effective solution. The core principle of the first model is to learn and identify the location characteristics of the train by analyzing these environmental monitoring data, and then accurately infer the real-time position and speed of the train.

[0083] In practical applications, the propagation characteristics of sound waves may change under extreme climate conditions, especially in low temperature environments at high altitudes. Low temperatures may affect the speed and attenuation rate of sound waves, thereby affecting the accuracy of the acoustic fingerprint data. The adaptability of the first model lies in its ability to learn the characteristic changes of sound wave propagation under different environmental conditions through a large amount of training data, including changes in sound wave characteristics caused by extreme temperature changes. By introducing diverse environmental monitoring data during the training process, the model can recognize and adapt to these changes and ensure the accuracy of acoustic data analysis.

[0084] The first model's ability to distinguish the acoustic signals generated by train movement from other interfering signals (such as construction sounds in tunnels, other mechanical equipment running sounds, etc.) mainly relies on advanced signal processing and pattern recognition technologies. By learning the characteristic patterns of acoustic fingerprints when trains pass, the model can identify specific frequency ranges, amplitude patterns, and delay characteristics, which are all unique acoustic characteristics of train movement. By including data on various interfering signals in the training set, the model can learn to distinguish the acoustic signals generated by trains from other non-train-related acoustic signals.

[0085] Geomagnetic field data is less sensitive to temperature changes than acoustic wave data, but extreme climate conditions may affect the accuracy of geomagnetic field measurements in indirect ways, such as changes in equipment performance under low temperature conditions. The first model is trained by including geomagnetic field data under various climate conditions, which improves the adaptability to data changes under these non-ideal conditions. In addition, the physical design and material selection of the geomagnetic field sensor can also be optimized for extreme climate conditions to reduce the impact of climate on measurement accuracy.

[0086] In the analysis of geomagnetic field data, the model also needs to distinguish between geomagnetic field changes caused by trains and other possible interference sources. This can be achieved by analyzing the time series characteristics of geomagnetic field changes. The geomagnetic field changes caused by trains usually have specific time characteristics and amplitude changes, which are different from the changes caused by other interference sources. Including geomagnetic field change data of different types of interference sources during training helps the model learn how to distinguish changes caused by trains and other interference based on the time series characteristics of geomagnetic field data.

[0087] In summary, in actual application, the first model improves its adaptability to extreme climate change and its ability to distinguish interference signals by learning from a training data set containing data under extreme climate conditions and various interference signals. This enables the model to maintain high accuracy and robustness when implementing train positioning in high-altitude and ultra-long tunnels, even in the face of complex and changing environmental conditions and potential interference sources. By continuously optimizing the model and enhancing the diversity of training data, the overall performance of the positioning system can be further improved.

[0088] Preferably, the training of the first model includes:

[0089] Collect historical environmental monitoring data and corresponding actual train location data, and mark the relationship between the environmental monitoring data and the actual train location, wherein the historical environmental monitoring data includes: historical environmental monitoring data and pre-processed data thereof;

[0090] Extract key features from the data, the key features including: temperature change trend; frequency, amplitude, and delay of the acoustic wave signal; and intensity change of the geomagnetic field;

[0091] The first model is trained using the key features as a training set to learn the relationship between the data features and the train positions.

[0092] The first model training is based on learning the relationship between environmental monitoring data and train location, aiming to achieve accurate train positioning in high-altitude and ultra-long tunnel environments. The key is to use machine learning technology to analyze and understand the complex relationship between different environmental variables such as sound waves, temperature and geomagnetic field changes and the specific location of the train.

[0093] Acoustic wave data: The sound waves generated by a train passing through a tunnel will produce unique reflection patterns in the tunnel, which can be captured by the acoustic fingerprint sensor and used for positioning.

[0094] Temperature data: Train operation may cause local changes in the surrounding temperature, which are recorded by the fiber optic sensor network and can assist in positioning analysis.

[0095] Geomagnetic field data: The metal structure of the train affects the distribution of the geomagnetic field in the tunnel. These changes captured by the geomagnetic field sensor provide key information for train positioning.

[0096] In practical applications, environmental monitoring data is collected in the tunnel through fiber optic sensor networks, acoustic fingerprint sensors and geomagnetic field sensors, and necessary preprocessing is performed, such as denoising and normalization, to improve data quality. Key features that are crucial to train positioning are extracted from the preprocessed data, such as the frequency, amplitude, and delay of the acoustic signal, the temperature change trend, and the intensity change of the geomagnetic field. The actual position data of the labeled train is used, and the extracted key features are used as input. The first model is trained through a machine learning algorithm to learn the mapping relationship between the features and the train position.

[0097] The first model can use the following machine learning techniques and algorithms when integrating acoustic wave, temperature and geomagnetic field data:

[0098] Deep learning (especially convolutional neural networks, CNNs): suitable for processing and identifying complex acoustic fingerprint features and their relationship with environmental factors (such as temperature and geomagnetic field changes);

[0099] Ensemble learning (such as random forest, RF): Integrate multiple models to improve the stability and accuracy of prediction, especially when integrating multi-source data, which can effectively improve the model's ability to recognize the characteristics of different data sources;

[0100] Long Short-Term Memory Network (LSTM): Suitable for processing time series data, such as geomagnetic field change data, which can capture patterns that change over time and help improve the dynamic accuracy of positioning.

[0101] In the actual training of the first model, other strategies that can be applied include:

[0102] Artificially increase data diversity in the training set and simulate different environmental conditions (such as changes in sound waves under extreme temperatures) to enhance the generalization ability of the model;

[0103] Use advanced feature fusion technology, such as a multimodal deep learning framework, to effectively integrate data from different sensors and improve the model's adaptability to complex environments and positioning accuracy;

[0104] Using the model pre-trained in a similar environment as a basis, fine-tuning is performed for the specific environment of high-altitude and ultra-long tunnels to reduce training time and improve model performance.

[0105] Implement an adaptive algorithm adjustment mechanism to automatically adjust model parameters based on real-time feedback (such as positioning error) to ensure that the model maintains optimal performance in a changing environment.

[0106] In practical applications, through the above strategies and technologies, the first model can effectively integrate data from acoustic fingerprint sensors, fiber optic sensor networks and geomagnetic field sensors, and achieve accurate and reliable train positioning even under extreme temperature conditions at high altitudes and ultra-long tunnels, providing solid technical support for real-time monitoring of tunnel safety.

[0107] By utilizing comprehensive environmental monitoring data, the first model can achieve accurate train positioning in special environments such as high altitudes and ultra-long tunnels; the model can process monitoring data collected in real time to provide instant positioning information for train operation and tunnel safety; through training with diverse environmental monitoring data, the model has good adaptability and robustness to extreme climate changes and different interference signals.

[0108] Preferably, the first model predicts the position of the train based on new environmental monitoring data by learning the relationship between environmental monitoring data and the train position; in the prediction stage, the first model receives new environmental monitoring data input, and outputs corresponding train positioning information through the rules and relationships learned in the training stage, and the train positioning information includes: the position coordinates and travel speed of the train.

[0109] The core principle of the first model is to learn and understand the relationship between environmental monitoring data and train location through machine learning technology, and then use new environmental monitoring data to accurately predict the location of the train. This method is particularly suitable for high-altitude and ultra-long tunnel environments, where traditional positioning technologies (such as GPS) may be ineffective or unreliable.

[0110] During the model training process, the system uses machine learning algorithms to identify the patterns and regularities contained in the data by analyzing historical environmental monitoring data and its corresponding actual train positions (position coordinates and travel speeds). These regularities are manifested in the reflection characteristics of sound waves, changes in temperature distribution, and changes in the strength of the geomagnetic field, which are closely related to the specific location and driving status of the train. Through training, the model can understand the train location information represented by different environmental monitoring data. In the prediction stage, the model receives new environmental monitoring data and calculates the current train's position coordinates and travel speed based on the relationship between the data pattern learned in the training stage and the train position.

[0111] Based on the current train positioning information and environmental monitoring data, the pre-trained second model is used to predict the train position and evaluate the tunnel safety status, generating real-time tunnel monitoring data and safety alerts.

[0112] The core principle of the second model is to use machine learning technology, especially predictive analysis and pattern recognition, to further analyze the current train positioning information and environmental monitoring data to achieve future predictions of train positions and real-time assessments of tunnel safety status. The model not only considers the data at the current moment, but also integrates the trends and patterns of historical data, making the predictions more accurate and reliable.

[0113] Train position prediction includes real-time train positioning information (such as position coordinates and travel speed) and environmental monitoring data (sound waves, temperature, geomagnetic field, etc.). The model predicts the train's position at a certain point in the future by analyzing historical and current data patterns. This prediction takes into account the historical trajectory of the train's operation, the speed change trend, and the impact of environmental factors on the train's operating status.

[0114] Tunnel safety status assessment uses environmental monitoring data, such as abnormal sound wave reflection signals that may indicate problems with the tunnel structure, and abnormal temperature changes that may indicate fire risks. The model evaluates the safety status of the tunnel, identifies potential safety hazards, and generates safety alerts based on real-time analysis of data.

[0115] In one embodiment, the second model is trained with a large amount of historical environmental monitoring data and train position data, including the annotation of safety conditions (e.g., safe, slight risk, high risk status). During the model operation phase, the current train positioning and environmental monitoring data are collected and processed in real time, and the model analyzes these data and makes predictions and assessments based on the learned historical patterns. The model output includes the future position prediction of the train and the safety status assessment results of the tunnel, providing real-time monitoring data and safety alarms for the tunnel monitoring system.

[0116] Combining historical and real-time data analysis, the second model can accurately predict the future location of the train, providing a scientific basis for train scheduling and management in the tunnel; by real-time assessment of the tunnel's safety status, the model can promptly detect and alert potential safety risks, effectively reducing the possibility of accidents; the model can process real-time data and quickly respond to environmental changes and changes in the train's operating status to ensure the safety and efficiency of tunnel operations.

[0117] In practical applications, when the second model deals with data fluctuations under extreme climate conditions, such as the impact of low temperatures caused by high altitudes on environmental monitoring data, the optimization includes:

[0118] Standardize and normalize environmental monitoring data to reduce the impact of data fluctuations under extreme climate conditions on model performance, which helps the model to more stably identify and learn patterns in the data rather than being disturbed by extreme values; use outlier detection algorithms to identify and process abnormal data points that may arise under extreme climate conditions, such as using the IQR (interquartile range) method or deep learning-based anomaly detection technology.

[0119] Introducing environmental factors (such as temperature, air pressure, etc.) into the model as additional features helps the model understand the context of data changes under extreme climate conditions and make more accurate predictions; using time series analysis techniques (such as ARIMA, LSTM) to capture the trends and periodicity of data changes over time, which is particularly important for dealing with data fluctuations that may occur under extreme climate conditions.

[0120] The use of ensemble learning methods such as random forests and gradient boosting machines (GBM) can integrate the prediction results of multiple models to improve the stability and accuracy of the overall prediction; deep learning models, especially convolutional neural networks (CNN) and recurrent neural networks (RNN) or long short-term memory networks (LSTM), can effectively capture complex data patterns with their powerful feature extraction and sequence data processing capabilities.

[0121] In practical applications, it is necessary to select specific machine learning algorithms, which are used in the second model to achieve future predictions of train positions and real-time assessments of tunnel safety status, as well as how these algorithms cope with complex and changing environmental conditions in high-altitude and ultra-long tunnels. The machine learning algorithms that can be selected include:

[0122] Long Short-Term Memory Networks (LSTM): LSTMs are particularly suitable for processing time series data and can capture the long-term dependencies between train position data and environmental monitoring data over time. They are particularly effective in predicting the future position of trains and assessing the safety status of tunnels under complex and changing environmental conditions.

[0123] Convolutional Neural Network (CNN): CNN can effectively process spatial data, such as spatial distribution data collected by acoustic fingerprint sensors and geomagnetic field sensors. CNN helps to identify specific patterns in environmental monitoring data by learning spatial features, thereby improving the accuracy of positioning and safety assessment.

[0124] Random Forest (RF): Random Forest is a powerful ensemble learning algorithm that improves prediction accuracy and robustness by building multiple decision trees and aggregating their prediction results. RF performs well when processing data with high-dimensional features and is suitable for integrating and analyzing environmental monitoring data from different sensors.

[0125] By adopting the above optimization strategies and machine learning algorithms, the second model can effectively cope with the complex and changing environmental conditions in high-altitude and ultra-long tunnels, including the challenges brought by extreme climate. These methods improve the accuracy and robustness of the model in dealing with fluctuations in environmental monitoring data, predicting train positions, and assessing the safety status of tunnels.

[0126] Preferably, the training of the second model includes:

[0127] Through the current train positioning information and environmental monitoring data, historical train positioning information and environmental monitoring data, historical train operation mode and special event history records, each data instance is annotated with the predicted train position and tunnel safety status, and then organized as the training set of the second model, where:

[0128] The historical train operation mode is used to help the second model understand the typical operation trajectory and change rules of the train. This information helps the model understand the train behavior under normal operating conditions and provides a benchmark for prediction. The historical train operation mode includes:

[0129] The average speed and position changes of trains in different time periods;

[0130] The train’s behavior patterns in response to different environmental conditions;

[0131] The impact of specific times on train speed and route selection;

[0132] The special event history records are used to help the second model identify special situations that affect train safety and positioning accuracy, so that the model can consider these abnormal situations when evaluating the tunnel safety status. The special event history records include: historical tunnel safety accidents and maintenance records;

[0133] The second model uses the training set to learn to predict the future position of the train and assess the safety status of the tunnel.

[0134] In practical applications, a comprehensive training set is constructed by integrating current and historical train positioning information, environmental monitoring data, and historical records of train operation patterns and special events. This training set is used to train the second model, which learns how to predict the future position of the train and the safety status of the tunnel based on existing information by analyzing the relationship between different data sources. After training, the model can receive new data input in real time, predict the train position and evaluate the safety status of the tunnel, and output real-time monitoring data and safety alerts.

[0135] With in-depth analysis of historical and current data, the second model can accurately predict the future position of the train and assess the safety status of the tunnel; the model's real-time data processing and prediction capabilities provide strong support for tunnel management and can promptly detect and respond to potential safety threats; the model can handle complex and changeable environmental conditions from high altitudes and ultra-long tunnels, including data fluctuations under extreme climates, ensuring the accuracy of positioning and safety monitoring under various conditions.

[0136] In the present invention, the second model adopts a highly integrated machine learning method to predict the future position of the train and evaluate the safety status of the tunnel. The core principle of this model is to analyze the current and historical train positioning information, environmental monitoring data, train operation mode and special event history records. Through these multi-dimensional data, the model learns the typical operation trajectory of the train, the response mode to environmental conditions, and the impact of specific time on its operation speed and path selection.

[0137] The second model is designed to predict the train position and assess the tunnel safety status by comprehensively analyzing train positioning information and environmental monitoring data, which requires the algorithm to be able to handle complex data patterns, especially in special environments such as high altitudes and ultra-long tunnels. The following are several machine learning algorithms suitable for the second model:

[0138] Deep Learning: Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN), especially Long Short-Term Memory Networks (LSTM). Deep learning algorithms are particularly suitable for processing large amounts of unstructured data, such as sound waves and temperature change data. CNN can effectively process spatial data features, such as acoustic fingerprint images, while LSTM excels at processing time series data, such as continuous train positions and environmental monitoring data. In high-altitude ultra-long tunnel environments, deep learning models can improve their robustness by adding data augmentation techniques, such as simulating environmental monitoring data under extreme climate change, and using transfer learning to pre-train models from other tasks to reduce the need for large amounts of labeled data.

[0139] Time series analysis algorithms: Autoregressive integrated moving average model (ARIMA), seasonal decomposition time series prediction algorithm (Prophet). Time series analysis algorithms are good at processing and predicting data that changes over time, and are very suitable for predicting the future position of trains and trends in environmental monitoring data. The unique climate change effects in high-altitude areas can be addressed by seasonally adjusting time series data, while using external regressors to introduce additional information such as sunshine duration and temperature to enhance the model's adaptability to extreme environmental changes.

[0140] Ensemble Learning: Random Forest, Gradient Boosting Decision Trees (such as XGBoost). Ensemble learning improves the accuracy and stability of the overall prediction by combining the predictions of multiple models, and is suitable for processing high-dimensional and nonlinear data sets. In view of the instability of high-altitude and ultra-long tunnel environments, ensemble learning algorithms can improve the robustness of the system by optimizing the diversity of sub-models, for example, by training models with different parameters or using different types of machine learning algorithms as sub-models.

[0141] In summary, in view of the complex and changeable environmental conditions in high-altitude and ultra-long tunnels, the deep learning algorithm provides strong technical support for the second model through its powerful feature extraction capabilities, the time series analysis algorithm captures the trend of data changes over time, and the integrated learning algorithm combines the prediction power of multiple models. Through the reasonable selection and optimization of these algorithms, the second model can accurately predict the position of trains and timely evaluate the safety status of tunnels, significantly improving the accuracy and efficiency of train positioning and real-time monitoring of tunnel safety in high-altitude ultra-long tunnels.

[0142] Preferably, the second model predicts the position of the train at a specific time point in the future by analyzing the learned spatiotemporal relationship and train operation mode, using current train positioning information and environmental monitoring data and historical train operation mode;

[0143] The second model predicts safety hazards in the tunnel and evaluates the safety status of the tunnel by analyzing the current environmental monitoring data and the predicted position of the train, combined with the historical records of special events;

[0144] The generated tunnel real-time monitoring data and security alerts are output.

[0145] The second model is an advanced analytical tool designed to use complex machine learning algorithms to predict the future location of trains and assess the safety status of tunnels by analyzing spatiotemporal relationships and train operation patterns. This process covers the entire process from data collection and processing to model training, prediction execution and output, including:

[0146] Spatiotemporal relationship analysis, the second model identifies the key factors affecting train operation, such as speed changes, running time and environmental conditions at specific locations, by learning the historical data of train operation (including time and space variables).

[0147] Train operation pattern learning: The model analyzes historical train operation patterns, including the average speed and position changes of trains in different time periods, and the response patterns of trains to specific environmental conditions (such as high altitude and extreme weather).

[0148] Environmental monitoring and safety assessment: Using real-time environmental monitoring data (sound waves, temperature, geomagnetic field, etc.) and predicted train positions, the model can detect potential safety hazards, such as tunnel structural problems or potential fire risks, and assess the safety status of the tunnel.

[0149] The second model in the present invention can accurately predict the position of the train at a specific point in the future, providing important information for train scheduling and traffic management in the tunnel; timely identification and assessment of safety hazards: through comprehensive analysis of real-time environmental data and historical events, the model can timely identify and warn of potential safety hazards, thereby improving the safety of tunnel operations; adapt to the challenges of high altitudes and ultra-long tunnels: the model is specially designed to adapt to the complex environment and extreme climatic conditions in high altitudes and ultra-long tunnels, ensuring the accuracy and reliability of predictions and assessments.

[0150] In practical applications, the strategies and techniques adopted by the second model in processing the historical records of special events ensure that this data is effectively used in the prediction and assessment process, especially in identifying and preventing potential tunnel safety problems in advance. The following are some key implementation methods:

[0151] The special event history records are classified and labeled in detail, such as "tunnel internal damage", "fire incident", "water intrusion accident", etc. Such labeling helps the model learn the specific impact of different types of events on tunnel safety and train operation.

[0152] Feature engineering is performed on special event records to extract features directly related to tunnel safety and train positioning, such as the specific location, time, and duration of the event, as well as the impact of the event on train speed and route selection.

[0153] Using machine learning algorithms, such as deep learning or ensemble learning, the model can identify typical signal patterns before special events occur. For example, by learning changes in environmental monitoring data before historical events (such as abnormal sound waves, sudden temperature rise, etc.), the model can issue early warnings when similar patterns occur again.

[0154] Conduct time series analysis on special events and the environmental monitoring data before and after them to identify trends and patterns in data changes before and after the events, thereby improving the accuracy of predictions of potential future safety issues.

[0155] The real-time collected environmental monitoring data is compared and analyzed with the learned data patterns before special events occur. When a matching pattern is detected, the system automatically generates a security alert and assesses the potential security risk level.

[0156] By simulating the occurrence of special events, the effectiveness of different emergency plans in alleviating or solving problems can be evaluated. This helps tunnel management departments develop more effective safety measures and emergency response plans.

[0157] The second model can identify potential safety hazards in the tunnel, such as structural damage and fire, in advance by analyzing special event historical records and real-time environmental monitoring data, thus providing a time window for taking preventive measures.

[0158] The dynamic evaluation capability of the second model of the present invention makes tunnel safety monitoring more flexible and forward-looking, and can timely update the tunnel safety status assessment based on current data and historical experience; through the application of these strategies and technologies, the second model can make full use of the historical records of special events, effectively predict the train position and evaluate the safety status of the tunnel in high-altitude and ultra-long tunnel environments, and significantly improve the intelligence and early warning capabilities of tunnel safety management.

[0159] like Figure 4 As shown, a system for executing the train positioning and tunnel safety real-time monitoring method comprises:

[0160] A data acquisition module, wherein the data acquisition module comprises:

[0161] The fiber optic sensor network is deployed in the tunnel to capture the changes in optical signals caused by sound wave reflection and temperature, and provide real-time sound wave reflection and temperature distribution data; it provides important data support for train positioning and tunnel environment monitoring.

[0162] The acoustic fingerprint sensor is installed at a specific location in the tunnel to collect the unique characteristics of the acoustic echoes generated by the movement of the train and generate acoustic fingerprint data for identifying and locating the train's position; the acoustic fingerprint data is the key data for identifying and locating the train's position.

[0163] Geomagnetic field sensors are placed at various monitoring points in the tunnel to regularly measure the geomagnetic field change data caused by the geomagnetic field disturbance caused by the metal body of the train in the tunnel; the geomagnetic field sensors placed in the tunnel regularly measure the geomagnetic field changes caused by the movement of the metal body of the train. These data are crucial for assisting the acoustic fingerprint data to locate the train and monitor the abnormal changes in the magnetic field in the tunnel caused by the train or other metal objects.

[0164] The data processing center, as the operating carrier of the first model and the second model, is used to receive the environmental monitoring data collected from the optical fiber sensor network, the acoustic fingerprint sensor and the geomagnetic field sensor, process it through the first model, and output the train positioning information; based on the current train positioning information and environmental monitoring data, the second model is used to predict the train position and evaluate the tunnel safety status, and output the real-time monitoring data and safety alarm of the tunnel.

[0165] The system of the present invention can achieve high-precision positioning of trains in special environments such as high altitudes and ultra-long tunnels by comprehensively analyzing acoustic fingerprints, temperature and geomagnetic field data; using advanced data processing and prediction technologies, the system can not only monitor the environmental conditions in the tunnel in real time, but also predict potential safety issues and issue alarms in a timely manner, greatly improving the safety management efficiency of the tunnel.

[0166] A storage medium stores a computer program, which, when executed by a processor, implements the steps of the train positioning and tunnel safety real-time monitoring method.

[0167] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0168] The present application is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0169] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0170] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0171] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0172] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0173] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0174] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0175] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for real-time monitoring of train positioning and tunnel safety, characterized in that: The following steps are involved: Setting monitoring points in the tunnel and deploying optical fiber sensor networks, acoustic fingerprint sensors and geomagnetic field sensors, establishing the monitoring network layout, collecting environmental monitoring data through the initialized monitoring network, and pre-processing the environmental monitoring data, the environmental monitoring data including: acoustic reflection and temperature change data, acoustic fingerprint data and geomagnetic field change data; Using the pre-trained first model, the environmental monitoring data is used as input and the train positioning information is output; The first model predicts the train position according to the new environmental monitoring data by learning the relationship between the environmental monitoring data and the train position; in the prediction stage, the first model receives the new environmental monitoring data input, and outputs the corresponding train positioning information according to the rules and relationships learned in the training stage, and the train positioning information includes: the position coordinates and the travel speed of the train; Based on the current train positioning information and environmental monitoring data, the pre-trained second model is used to predict the train position and evaluate the tunnel safety status, generating real-time tunnel monitoring data and safety alarms. The second model predicts the location of a train at a specific time in the future by analyzing the learned spatiotemporal relationship and train operation mode, using current train positioning information and environmental monitoring data and historical train operation modes; The second model predicts safety hazards in the tunnel and evaluates the safety status of the tunnel by analyzing the current environmental monitoring data and the predicted position of the train, combined with the historical records of special events; The generated tunnel real-time monitoring data and security alerts are output.

2. The method for real-time monitoring of train positioning and tunnel safety according to claim 1, characterized in that: Through the fiber optic sensor network deployed in the tunnel, the changes in sound waves and temperature are monitored in real time, and the changes in light signals caused by sound wave reflection and temperature are captured respectively, to obtain sound wave reflection and temperature distribution data; the acoustic wave fingerprint sensor is used to collect the unique characteristics of the acoustic wave echoes generated by the movement of the train in the tunnel to generate acoustic wave fingerprint data; the geomagnetic field intensity changes in the tunnel are regularly measured through the geomagnetic field sensor, the geomagnetic field disturbance caused by the metal body of the train is captured, and the geomagnetic field change data is obtained.

3. The method for real-time monitoring of train positioning and tunnel safety according to claim 1, characterized in that: The preprocessing of the environmental monitoring data includes: The preprocessing of acoustic wave reflection and temperature change data includes: applying digital signal processing technology to remove background noise and extract effective acoustic wave reflection signals and temperature change signals; normalizing the acoustic wave reflection intensity and temperature data to eliminate the influence of differences between devices and measurement range; using signal enhancement technology to enhance the characteristics of acoustic wave signals; The preprocessing of acoustic fingerprint data includes: extracting acoustic features from the collected acoustic echo data, compressing the acoustic feature data and identifying and processing abnormal values, wherein the acoustic features include: frequency, amplitude and echo delay; The preprocessing of geomagnetic field change data includes: performing baseline calibration on geomagnetic field data based on the measured values ​​of geomagnetic field strength at monitoring points when no trains pass, so as to measure the geomagnetic field changes caused by trains; applying filtering technology to remove short-term fluctuations and electromagnetic interference in geomagnetic field data, and retaining the geomagnetic field change signals caused by trains; analyzing the time series trend of geomagnetic field data, and identifying the geomagnetic field change patterns caused by train movement.

4. The method for real-time monitoring of train positioning and tunnel safety according to claim 1, characterized in that: The training of the first model includes: Collect historical environmental monitoring data and corresponding actual train location data, and mark the relationship between the environmental monitoring data and the actual train location, wherein the historical environmental monitoring data includes: historical environmental monitoring data and pre-processed data thereof; Extract key features from the data, the key features including: temperature change trend; frequency, amplitude, and delay of the acoustic wave signal; and intensity change of the geomagnetic field; The first model is trained using the key features as a training set to learn the relationship between the data features and the train positions.

5. The method for real-time monitoring of train positioning and tunnel safety according to claim 4, characterized in that: The training of the second model includes: Through the current train positioning information and environmental monitoring data, historical train positioning information and environmental monitoring data, historical train operation mode and special event history records, each data instance is annotated with the predicted train position and tunnel safety status, and then organized as the training set of the second model, where: The historical train operation mode is used to help the second model understand the typical operation trajectory and change law of the train. The historical train operation mode includes: The average speed and position changes of trains in different time periods; The train’s behavior patterns in response to different environmental conditions; The impact of specific times on train speed and route selection; The special event history records are used to help the second model identify special situations that affect train safety and positioning accuracy, and the special event history records include: historical safety accidents and maintenance records in the tunnel; The second model uses the training set to learn to predict the future position of the train and assess the safety status of the tunnel.

6. A system for executing the train positioning and tunnel safety real-time monitoring method according to any one of claims 1 to 5, characterized in that: include: A data acquisition module, wherein the data acquisition module comprises: Fiber optic sensor networks are deployed in tunnels to capture changes in light signals caused by acoustic reflection and temperature, providing real-time acoustic reflection and temperature distribution data; Acoustic fingerprint sensors, installed at specific locations in the tunnel, are used to collect the unique characteristics of the acoustic echoes generated by the movement of the train and generate acoustic fingerprint data for identifying and locating the train's position; Geomagnetic field sensors are arranged at various monitoring points in the tunnel to regularly measure the geomagnetic field change data caused by the geomagnetic field disturbance caused by the metal body of the train in the tunnel; The data processing center, as the operating carrier of the first model and the second model, is used to receive the environmental monitoring data collected from the optical fiber sensor network, the acoustic fingerprint sensor and the geomagnetic field sensor, process it through the first model, and output the train positioning information; based on the current train positioning information and environmental monitoring data, the second model is used to predict the train position and evaluate the tunnel safety status, and output the real-time monitoring data and safety alarm of the tunnel.

7. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method for real-time monitoring of train positioning and tunnel safety described in any one of claims 1 to 5 are implemented.