Railway safety shielding method and system based on sensor

Abstract

The invention discloses a railway safety shielding method and system based on a sensor, and relates to the technical field of railway safety, and the method comprises the steps: carrying out the time-space correlation analysis of historical environment data, outputting a first analysis result, collecting real-time data, storing the real-time data to an edge node, calculating a statistical characteristic quantity, and building a safety baseline according to the statistical characteristic quantity. According to the method, data acquisition is optimized through space-time correlation analysis, an accurate safety baseline is established, dynamic risk assessment is achieved through the LSTM model, a response strategy is set according to the risk level, and the risk assessment accuracy is improved. According to the method, a digital twin platform is adopted, a full-link closed loop from risk identification to active protection is formed, meanwhile, the digital twin platform provides visual monitoring and evacuation optimization, daily sensor calibration ensures data accuracy, the intelligence, real-time performance and accuracy of railway safety monitoring are integrally improved, and railway operation safety is effectively guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of railway safety, and particularly to a railway safety shielding method and system based on sensors. Background Art

[0002] With the rapid development of railway transportation, railway safety has become a crucial issue. Traditional railway safety monitoring methods mainly rely on manual inspections and fixed sensor monitoring, which have problems such as slow response speed, limited monitoring range, and low data utilization rate. Especially in complex and changeable railway operation environments, such as urban rail transit and high-speed railways, where the train operation density is high and the passenger flow is large, higher requirements are put forward for the real-time performance, accuracy, and intelligence level of the safety monitoring system.

[0003] Currently, the Chinese invention patent with the application number CN202410947050.4 discloses a railway personal safety protection and early warning system. This platform collaborates through multiple logical layers to achieve real-time monitoring and early warning of railway safety risks. The positioning layer is responsible for receiving the positioning information of trains and operating personnel; the data processing layer analyzes and compares the positioning information to predict safety risks, including collision risks, curve or turnout overspeed risks, illegal entry risks, and multi-train meeting risks, etc.; the early warning layer triggers the early warning mechanism based on the risk prediction results to generate early warning information; the communication layer establishes a communication link to ensure smooth information; the control layer receives dispatching instructions and adjusts the platform control and early warning settings.

[0004] The above technology is difficult to build a full-link closed loop from data collection to active protection and generate an optimized evacuation path plan, which affects the protection effect. Summary of the Invention

[0005] The technical problem solved by the present invention is that the prior art is difficult to build a full-link closed loop from data collection to active protection and generate an optimized evacuation path plan, which affects the protection effect.

[0006] To solve the above technical problem, the present invention provides the following technical solutions: A railway safety shielding method based on sensors, comprising the following steps: Step S1, perform spatio-temporal correlation analysis on abnormal events in historical environmental data, and output the first analysis result; Step S2, dynamically adjust the sensor sampling frequency according to the first analysis result, collect real-time data and store it in the edge node; Step S3, calculate statistical feature quantities based on the first analysis result, and establish a safety baseline according to the statistical feature quantities; Step S4, construct a dynamic risk assessment model, match the corresponding sub-risk assessment models, calculate the deviation degree and analyze it to generate the second analysis result; Step S5, set the response strategy grading according to the second analysis result, and output a response instruction.

[0007] Preferably, the step S1 includes the following sub-steps: Step S101, retrieve historical environmental data from the distributed sensor network, where the historical environmental data includes historical vibration spectra, train stopping time series, abnormal events, and passenger flow distribution; Step S102, perform spatio-temporal correlation analysis on the abnormal events in the historical environmental data, and use the Moran's spatial autocorrelation index to quantify the event clustering: ; where N is the total number of abnormal events, is the Moran's index, is the spatial weight matrix, is the severity level at the th time, is the severity level at the th time, is the average severity level, and are event indices; If is greater than the preset first autocorrelation index, it is determined as a high-risk aggregation area, and the output is the first analysis result.

[0008] Preferably, the step S2 includes the following sub-steps: Step S201, if the Moran's index of the high-risk aggregation area is greater than or equal to the second autocorrelation index, output a signal to increase the sampling frequency of the vibration sensor; If the Moran's index is between the first autocorrelation index and the second autocorrelation index, maintain the initial sampling frequency; Step S202, collect the track vibration acceleration signal, GPS positioning, and passenger heat map, and calculate the vibration energy spectral density: ; where, is the sampling window, is the frequency, is the vibration energy spectral density, is the historical vibration spectrum; Step S203, compress and transmit the vibration energy spectral density, GPS positioning, and passenger heat map to the edge node for storage.

[0009] Preferably, the step S3 includes the following sub-steps: Step S301, segment the historical vibration spectrum by time window, and calculate the statistical characteristic quantities of each segment of data: ; ; ; Among them, is the divided difference, is the variance, is the peak factor, is the maximum amplitude of the historical vibration spectrum within the time window, is the historical vibration spectrum at the th sampling point within the time window; Step S302, taking as the range, establish the safety baseline of vibration energy, train braking distance and passenger density.

[0010] Preferably, the step S4 includes the following sub-steps: Step S401, construct a dynamic risk assessment model based on LSTM: ; Among them, is the hidden state, is the current input, is the environmental weight coefficient, is the real-time risk score, is the input vector at the current moment, is the contribution weight of a certain dimension in the input feature or hidden state to the risk score, is the current moment; Step S402, according to the train operation stage matching sub-model, calculate the deviation degree between the real-time data and the baseline: ; Among them, is the deviation degree, is the th feature in the input vector at the current moment, is the standard deviation corresponding to the th feature in the baseline; Step S403, if the deviation degree continuously exceeds the preset deviation threshold for a preset first continuous duration, mark it as an emergency risk, otherwise mark it as a potential risk, and output the second analysis result.

[0011] Preferably, the step S5 includes the following sub-steps: Step S501, set the response strategy classification: Emergency risk: Activate the sound and light alarm + close the physical barrier, and the closing time is less than or equal to the preset first closing duration; Potential risk: Start the warning projection + local air pressure buffering, and the response delay is less than or equal to the preset first delay duration; Output response instruction; Step S502: Send the response instruction to the platform edge controller through the 5G-Uu interface to synchronously update the status of the digital twin platform.

[0012] Preferably, the logic for closing the physical barrier is as follows: Generate a flexible barrier deformation path based on the train car body contour, and the path function satisfies: ; where is the path function, is the amplitude, is the attenuation coefficient, is the phase shift, is the wave number, is the position; Adjust the actuator displacement of the shape memory alloy through a PID controller: ; where is the actuator displacement, is the contour matching error, is the PID control ratio, is the PID control integral, is the PID control differential coefficient.

[0013] Preferably, the update of the digital twin platform includes: Map the real-time sensor data to the virtual platform model and render a 3D risk heat map.

[0014] Generate an evacuation path optimization plan based on the 3D risk heat map. The path cost function of the evacuation path optimization plan is: ; where is the path length, is the risk value of the area passed by the path, and are weight coefficients, and is the comprehensive cost of the evacuation path.

[0015] Preferably, it further includes step S6: Trigger the sensor calibration process during the daily off-peak period, output the calibration error data. If the calibration error data is less than the preset calibration error threshold, correct the infrared passenger count error through a laser rangefinder. If the calibration error data is greater than the preset calibration error threshold, switch to the backup sensor and report it to the operation and maintenance system.

[0016] A sensor-based railway safety shielding system, which is applied to the sensor-based railway safety shielding method described above, includes an association analysis module, a sensor adjustment module, a baseline establishment module, a risk assessment module, and a hierarchical response module.

[0017] The association analysis module is used to perform spatio-temporal association analysis on abnormal events in historical environmental data and output a first analysis result.

[0018] The sensor adjustment module is used to dynamically adjust the sensor sampling frequency according to the first analysis result, collect real-time data and store it in the edge node.

[0019] The baseline establishment module is used to calculate statistical feature quantities based on the first analysis result and establish a safety baseline according to the statistical feature quantities.

[0020] The risk assessment module is used to construct a dynamic risk assessment model, match the corresponding sub-risk assessment models, calculate and analyze the deviation degree, and generate a second analysis result.

[0021] The hierarchical response module is used to set the response strategy level according to the second analysis result and output a response instruction.

[0022] Advantages of the present invention: The present invention optimizes data collection through spatio-temporal association analysis, establishes an accurate safety baseline, realizes dynamic risk assessment using the LSTM model, and sets response strategies according to risk levels, forming a full-link closed loop from risk identification to active protection. At the same time, the digital twin platform provides intuitive monitoring and evacuation optimization, and daily sensor calibration ensures accurate data, overall improving the intelligence, real-time performance, and accuracy of railway safety monitoring, and effectively ensuring railway operation safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a step flowchart of a sensor-based railway safety shielding method provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of the basic process of a sensor-based railway safety shielding system provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention is provided in conjunction with the accompanying drawings of the specification. Obviously, the described embodiments are some embodiments of the present invention, rather than all embodiments.

[0025] Embodiment 1, referring to Figure 1 , provides a sensor-based railway safety shielding method, including the following steps: Step S1, perform spatio-temporal correlation analysis on abnormal events in historical environmental data, and output the first analysis result.

[0026] Step S2, dynamically adjust the sensor sampling frequency according to the first analysis result, collect real-time data and store it in the edge node.

[0027] Step S3, calculate statistical characteristic quantities based on the first analysis result, and establish a safety baseline according to the statistical characteristic quantities.

[0028] Step S4, construct a dynamic risk assessment model, match the corresponding sub-risk assessment models, calculate the deviation degree and analyze it to generate the second analysis result.

[0029] Step S5, set the response strategy grading according to the second analysis result and output the response instruction.

[0030] Step S1 includes the following sub-steps: Step S101, retrieve historical environmental data from the distributed sensor network. The historical environmental data includes historical vibration spectra, train stop time series, abnormal events, and passenger flow distribution.

[0031] Step S101 retrieves historical environmental data from the distributed sensor network. This data covers multiple aspects such as historical vibration spectra, train stop time series, abnormal events, and passenger flow distribution. The effect of this step is that by retrieving various types of historical data, it can comprehensively reflect the operating environment and safety status along the railway. The data provided by the distributed sensor network is highly accurate and reliable, providing strong support for subsequent analysis. The historical data contains information over a long time span, which helps to discover potential safety hazards and trends.

[0032] Step S102, perform spatio-temporal correlation analysis on abnormal events in historical environmental data, and use the Moran's spatial autocorrelation index to quantify event clustering: ; where N is the total number of abnormal events, is the Moran's index, is the spatial weight matrix, is the severity level at the th time, is the severity level at the th time, is the average severity level, and are the event indices.

[0033] If If it is greater than a preset first autocorrelation index, it is determined as a high-risk aggregation area and the output is the first analysis result. Among them, the severity level is calculated by summing up the values according to the preset assignment indicators based on the historical vibration spectrum, train stopping time series, abnormal events, and passenger flow distribution.

[0034] Step S102 performs spatio-temporal correlation analysis on abnormal events in the historical environmental data and uses the Moran spatial autocorrelation index to quantify the event aggregation. The effect of this step is that through spatio-temporal correlation analysis, it can reveal the distribution law and correlation of abnormal events in time and space, which helps to identify high-risk areas. As an index to quantify event aggregation, the Moran index can objectively and accurately reflect the aggregation degree of abnormal events, providing a scientific basis for the determination of high-risk areas. According to the comparison result between the Moran index and the preset first autocorrelation index, high-risk aggregation areas can be accurately determined, providing key information for subsequent safety monitoring and early warning. At the same time, the first analysis result output by this step is also an important basis for the implementation of subsequent steps.

[0035] Step S1 aims to identify high-risk aggregation areas along the railway through in-depth analysis of historical environmental data, providing key information for subsequent safety monitoring and early warning. By retrieving diverse historical data from the distributed sensor network and performing spatio-temporal correlation analysis, this step can comprehensively and accurately evaluate the safety status of railway operations, laying a solid foundation for the implementation of subsequent steps.

[0036] Step S2 includes the following sub-steps: Step S201, if the Moran index of the high-risk aggregation area is greater than or equal to the second autocorrelation index, output a signal to increase the sampling frequency of the vibration sensor.

[0037] If the Moran index is between the first autocorrelation index and the second autocorrelation index, maintain the initial sampling frequency.

[0038] Step S201 compares the Moran index of the high-risk aggregation area with the preset second autocorrelation index and dynamically adjusts the sampling frequency of the vibration sensor. The effect of this step is that by comparing the Moran index with the second autocorrelation index, the sampling frequency can be flexibly adjusted to meet the data acquisition requirements of different risk areas. A higher sampling frequency is used in high-risk areas, which helps to capture more detailed information and improve data quality. Maintaining the initial sampling frequency in low-risk or medium-risk areas not only ensures the integrity of the data but also avoids unnecessary resource waste. Dynamically adjusting the sampling frequency can quickly respond to changes in safety risks and improve the overall safety and reliability of the system.

[0039] Step S202, collect the track vibration acceleration signal, GPS positioning, and passenger heat map, and calculate the vibration energy spectral density: ; Among them, is the sampling window, is the frequency, is the vibration energy spectral density, is the historical vibration spectrum.

[0040] In step S202, the track vibration acceleration signal, GPS positioning, and passenger heat map are collected, and the vibration energy spectral density is calculated. The effect of this step is that the types of collected data cover multiple aspects such as track vibration, geographical location, and passenger distribution, which can comprehensively reflect the safety status of railway operations. By calculating the vibration energy spectral density, the intensity and frequency distribution of track vibration can be quantified, providing an important basis for subsequent safety assessments. The collected data is professionally processed and analyzed, with high accuracy and reliability, providing strong support for the implementation of subsequent steps.

[0041] In step S203, the vibration energy spectral density, GPS positioning, and passenger heat map are compressed and then transmitted to the edge node for storage.

[0042] In step S203, the vibration energy spectral density, GPS positioning, and passenger heat map are compressed and then transmitted to the edge node for storage. The effect of this step is that through data compression technology, the occupied space for data transmission and storage can be reduced, improving the efficiency of data processing. Storing the data in the edge node can shorten the latency time for data access and improve the response speed of the system. At the same time, edge storage also enhances the reliability and security of the data, providing strong guarantee for subsequent data analysis and applications.

[0043] Step S2 aims to dynamically adjust the sampling frequency of the vibration sensor according to the determination result of the high-risk aggregation area, and collect key data for subsequent analysis. By optimizing the data collection strategy, this step can ensure that sufficient and effective data is obtained in key areas and critical moments, providing strong support for subsequent safety assessments and early warnings. At the same time, data compression and edge storage also improve the efficiency and reliability of data processing.

[0044] Step S3 includes the following sub-steps: In step S301, the historical vibration spectrum is segmented by time window, and the statistical characteristic quantities of each segment of data are calculated: ; ; ; Among them, is the mean difference, is the variance, is the peak factor, is the maximum amplitude of the historical vibration spectrum within the time window, is the vibration amplitude of the historical vibration spectrum at the -th sampling point within the time window.

[0045] In step S301, the historical vibration spectrum is segmented by time window, and the statistical characteristic quantities of each segment of data are calculated. The effect of this step is mainly reflected in that by segmenting the historical vibration spectrum by time window, the changes in the vibration spectrum in different time periods can be analyzed more carefully. This helps to capture potential safety hazards, improve the accuracy and timeliness of risk assessment. Calculating the statistical characteristic quantities of each segment of data can quantify the characteristics and change trends of the vibration spectrum. These statistical characteristic quantities will be used as important reference bases for establishing safety baselines in the follow-up. By preprocessing historical data, such as denoising and filtering, the accuracy and reliability of the data can be improved, providing a high-quality data source for subsequent analysis.

[0046] In step S302, with as the range, establish safety baselines for vibration energy, train braking distance, and passenger density.

[0047] In step S302, safety baselines for vibration energy, train braking distance, and passenger density are established. The effect of this step is mainly reflected in that based on the statistical characteristic quantities of historical data, safety baselines for vibration energy, train braking distance, and passenger density are established. These baselines will be used as important criteria for evaluating the safety status of railway operations, helping to detect and warn of potential safety risks in a timely manner. By establishing safety baselines, safety risks can be managed quantitatively, providing a more scientific and objective basis for railway safety monitoring. This helps to improve the safety and reliability of railway operations, reduce the probability of accidents. The establishment of safety baselines can also provide strong support for railway operation decision-making. When the real-time monitoring data deviates from the safety baseline, corresponding measures can be taken in a timely manner for intervention and adjustment to ensure the safety and stability of railway operations. At the same time, the safety baseline can also be used as an important reference basis for railway safety training and drills, improving the safety awareness and emergency handling capabilities of employees.

[0048] The core purpose of step S3 is to establish a set of scientific and reasonable safety baselines by analyzing historical vibration spectrum data. This set of baselines will be used as an important basis for subsequent risk assessment and early warning to ensure the safety and reliability of railway operations. Through the implementation of this step, quantitative management of key safety parameters such as vibration energy, train braking distance, and passenger density along the railway can be achieved, providing strong data support for railway safety monitoring.

[0049] Step S4 includes the following sub-steps: In step S401, construct a dynamic risk assessment model based on LSTM: ; Among them, is in the hidden state, is the current input, is the environmental weight coefficient, is the real-time risk score, is the input vector at the current moment, is the contribution weight of a certain dimension in the input feature or hidden state to the risk score, is the current moment.

[0050] Step S401 constructs a dynamic risk assessment model based on LSTM. The effect of this step is mainly reflected in that through the LSTM network, the time series information in historical data can be fully utilized to capture the long-term dependence relationship between data, thereby constructing a more accurate risk assessment model. The LSTM model has the ability of dynamic assessment and can continuously update and adjust the risk assessment results according to the input of real-time data to ensure the accuracy and timeliness of the assessment. The LSTM model can adapt to the risk assessment requirements of different train operation stages and scenarios. Through training and optimization, accurate prediction and assessment of various complex situations can be achieved.

[0051] Step S402, match the sub-model according to the train operation stage, and calculate the deviation degree between the real-time data and the baseline: ; Among them, is the deviation degree, is the th feature in the input vector at the current moment, is the standard deviation corresponding to the th feature in the baseline.

[0052] Step S402 matches the sub-model according to the train operation stage and calculates the deviation degree between the real-time data and the baseline. The effect of this step is mainly reflected in that for different train operation stages, the corresponding sub-models are matched for evaluation. This helps to more accurately reflect the risk characteristics of different stages and improve the accuracy of the evaluation. By calculating the deviation degree between the real-time data and the baseline, the size and degree of the risk can be quantified. This step provides an important basis for subsequent risk marking and early warning. The input of real-time data and the calculation of the deviation degree can realize the instant monitoring and evaluation of railway operation risks, ensuring the timeliness and effectiveness of early warning.

[0053] Step S403, if the deviation degree continuously exceeds the preset deviation threshold for the preset first continuous duration, mark it as an emergency risk, otherwise mark it as a potential risk, and the output is the second analysis result.

[0054] Step S403 marks it as an emergency risk or a potential risk according to the duration of the deviation degree and a preset deviation degree threshold, and outputs it as the second analysis result. The effect of this step is mainly reflected in that by setting the deviation degree threshold and the first duration, the risks can be classified into two categories: emergency risks and potential risks. This step helps to conduct hierarchical management of risks, improve the pertinence and effectiveness of responses. When the deviation degree continuously exceeds the threshold, it is promptly marked as an emergency risk and a warning message is output. This helps the railway operation department to quickly take measures for intervention and adjustment, reducing the probability of accidents. The output form of the second analysis result is clear and definite, facilitating the understanding and application by the railway operation department. This step provides strong support for subsequent risk management and decision-making.

[0055] Step S4 aims to achieve accurate assessment of real-time risks in railway operation by constructing a dynamic risk assessment model based on LSTM. This step can comprehensively consider historical data and real-time data. By matching sub-models for different train operation stages, it calculates the deviation degree between real-time data and the baseline, thereby accurately identifying emergency risks and potential risks. The implementation of this step provides timely and effective risk warnings for railway operation, helps to reduce the probability of accidents, and enhances the safety and reliability of railway operation.

[0056] Step S5 includes the following sub-steps: Step S501, set the classification of response strategies: Emergency risk: Activate the audible and visual alarm + close the physical barrier, and the closing time is less than or equal to the preset first closing duration.

[0057] Potential risk: Start the warning projection + local air pressure buffering, and the response delay is less than or equal to the preset first delay duration.

[0058] Output the response instruction.

[0059] Step S501 sets the classification of response strategies and outputs the response instruction. The effect of this step is mainly reflected in that according to the risk assessment results, the risks are divided into emergency risks and potential risks, and corresponding response strategies are set respectively. For emergency risks, measures such as activating the audible and visual alarm and closing the physical barrier are taken to ensure that the risks are quickly controlled; for potential risks, warning measures such as starting the warning projection and local air pressure buffering are initiated to provide a time buffer for risk prevention and control. According to the hierarchical response strategy, the corresponding response instruction is output. These instructions will serve as the basis for subsequent execution to ensure that the response measures can be accurately and timely implemented. By setting the hierarchical response strategy and outputting the response instruction, this step improves the efficiency and accuracy of emergency response, helps to reduce the probability of accidents and mitigate the losses caused by accidents.

[0060] Step S502 sends the response instruction to the platform edge controller through the 5G-Uu interface and synchronously updates the status of the digital twin platform.

[0061] The logic of the physical barrier closing is as follows: Generate a flexible barrier deformation path according to the train car body contour, and the path function satisfies: ; Among them, is the path function, is the amplitude, is the attenuation coefficient, is the phase shift, is the wave number, is the position.

[0062] Adjust the actuator displacement of the shape memory alloy through a PID controller: ; Among them, is the actuator displacement, is the contour matching error, is the PID control ratio, is the PID control integral, is the PID control differential coefficient.

[0063] The update of the digital twin platform includes: Map the real-time sensor data to the virtual platform model and render a 3D risk heat map.

[0064] Generate an evacuation path optimization plan based on the 3D risk heat map, and the path cost function of the evacuation path optimization plan is: ; Among them, is the path length, is the risk value of the area passed by the path, and are weight coefficients, and is the comprehensive cost of the evacuation path.

[0065] Step S502 issues the response instruction to the platform edge controller via the 5G-Uu interface and synchronously updates the status of the digital twin platform. The effect of this step is mainly reflected in leveraging the high-speed transmission characteristics of the 5G-Uu interface to quickly issue the response instruction to the platform edge controller. This ensures that the response measures can be implemented promptly, improving the timeliness of emergency response. Map the real-time sensor data to the virtual platform model, render the 3D risk heat map, and generate an optimized evacuation route plan based on this heat map. This step realizes the real-time monitoring and dynamic management of risks, providing more intuitive and accurate information support for emergency response. Through the synchronous update of the digital twin platform, the risk situation can be grasped in real time, and emergency response measures such as optimizing the evacuation route can be carried out. This helps to improve the overall operation efficiency and reduce the emergency response cost. At the same time, the digital twin platform can also provide an important reference basis for subsequent emergency drills and training.

[0066] Step S5 aims to quickly and effectively implement the response strategy according to the risk assessment results to ensure the safety of railway operations. By setting a hierarchical response strategy, this step can take different countermeasures for emergency risks and potential risks to ensure that the risks are promptly controlled. At the same time, the response instruction is quickly issued to the platform edge controller via the 5G-Uu interface, and the status of the digital twin platform is synchronously updated, realizing the real-time monitoring and dynamic management of risks. The implementation of this step not only improves the safety of railway operations but also optimizes the emergency response process and enhances the overall operation efficiency.

[0067] In step S6, the sensor calibration process is triggered during the daily low peak period, and the calibration error data is output. If the calibration error data is less than the preset calibration error threshold, the infrared passenger counting error is corrected by the laser rangefinder. If the calibration error data is greater than the preset calibration error threshold, the backup sensor is switched to and reported to the operation and maintenance system.

[0068] The core purpose of step S6 is to ensure the accuracy and reliability of the passenger counting system at railway platforms. By triggering the sensor calibration process regularly during off-peak hours, errors can be detected and corrected in a timely manner to ensure the normal operation of the system and the accuracy of data. Step S6 is set to trigger the sensor calibration process during the daily off-peak hours. This time period is selected to avoid operational interference during peak hours and ensure the smooth progress of the calibration work. The output of the calibration process is calibration error data, which directly reflects the current working state and accuracy level of the sensor. If the calibration error data is less than the preset calibration error threshold, it indicates that although there are certain errors in the sensor, they are still within an acceptable range. At this time, the infrared passenger counting error can be corrected by a laser rangefinder, which can further improve the accuracy of counting. As a high-precision measurement tool, the application of the laser rangefinder can compensate for the errors of the infrared sensor under specific conditions and ensure the reliability of the counting results. If the calibration error data is greater than the preset calibration error threshold, it indicates that the sensor may have serious faults or performance degradation and cannot meet the normal counting requirements. In this case, step S6 will automatically switch to a backup sensor to ensure the continuous operation of the passenger counting system. This switching mechanism avoids system paralysis caused by a single sensor failure and ensures the operational safety of railway platforms.

[0069] Meanwhile, the information of the faulty sensor will be reported to the operation and maintenance system so that the operation and maintenance personnel can be informed in a timely manner and take repair or replacement measures to prevent the fault from expanding further. The implementation of step S6 not only improves the accuracy and reliability of the passenger counting system, but also reduces the risk of operational interruption caused by sensor failures through regular calibration and fault switching mechanisms. Through the reporting mechanism of the operation and maintenance system, the operation and maintenance personnel can more accurately locate the fault point, improve the repair efficiency, and avoid unnecessary sensor replacement costs.

[0070] This method can accurately identify high-risk aggregation areas by performing spatio-temporal correlation analysis on abnormal events in historical environmental data and dynamically adjust the sensor sampling frequency according to the risk level. This can not only reduce unnecessary data acquisition and storage costs, but also ensure sufficient data support in key areas and critical moments. Real-time collect data such as track vibration acceleration signals, GPS positioning, and passenger heat maps, and compress and transmit them to the edge node for storage, achieving fast data processing and efficient utilization.

[0071] Based on historical vibration spectrum data, calculate statistical characteristic quantities and establish a safety baseline, which provides a reliable reference basis for subsequent risk assessment. The safety baseline takes into account multiple factors such as vibration energy, train braking distance, and passenger density, and can comprehensively reflect the safety status of railway operations.

[0072] Build a dynamic risk assessment model based on LSTM, which can match sub-models according to the train operation stage, calculate the deviation degree between real-time data and the baseline, realize the dynamic assessment of risks, set the response strategy grading according to the risk assessment results, output response instructions, and achieve a full-link closed loop from risk identification to active protection. Different response strategies are adopted for emergency risks and potential risks, ensuring the pertinence and effectiveness of safety measures.

[0073] Map real-time sensor data to the virtual platform model through the digital twin platform, render a 3D risk heat map, provide an intuitive safety monitoring interface for operation personnel, generate an evacuation route optimization plan based on the 3D risk heat map, and improve the efficiency and accuracy of emergency response.

[0074] Trigger the sensor calibration process during the daily low-peak period, output calibration error data, and take corresponding correction measures according to the calibration results. This can not only ensure the accuracy and reliability of sensor data, but also detect and handle sensor failures in a timely manner, improving the stability and reliability of safety monitoring.

[0075] Example 2, refer to Figure 2 and provides a sensor-based railway safety shielding system, including a correlation analysis module, a sensor adjustment module, a baseline establishment module, a risk assessment module, and a hierarchical response module.

[0076] The correlation analysis module is used to perform spatio-temporal correlation analysis on abnormal events in historical environmental data and output the first analysis result.

[0077] The correlation analysis module is one of the core components of the railway safety shielding system. Its main function is to perform in-depth spatio-temporal correlation analysis on abnormal events in historical environmental data. Through this analysis, the module can identify potential connections between abnormal events, such as temporal continuity, spatial aggregation, etc., and thus output the first analysis result. This result has important guiding significance for the operations of subsequent modules, can help the system more accurately understand the safety situation of the railway operation environment, and provide strong support for subsequent risk assessment and response strategy formulation.

[0078] The sensor adjustment module is used to dynamically adjust the sensor sampling frequency according to the first analysis result, collect real-time data and store it in the edge node.

[0079] The sensor adjustment module dynamically adjusts the sampling frequency of sensors according to the first analysis result output by the correlation analysis module. This means that in the face of different security postures, the system can flexibly adjust the density of data collection to more efficiently capture key information. At the same time, this module is also responsible for collecting real-time data and storing it in the edge node, providing a solid data foundation for the real-time monitoring and rapid response of the system. By intelligently adjusting the sampling frequency, the sensor adjustment module not only improves the pertinence of data collection, but also effectively reduces data redundancy and enhances the overall efficiency of the system.

[0080] The baseline establishment module is used to calculate statistical feature quantities based on the first analysis result and establish a security baseline according to the statistical feature quantities.

[0081] Based on the first analysis result output by the correlation analysis module, the baseline establishment module further calculates statistical feature quantities, which can quantitatively describe the normal state of the railway operation environment. Subsequently, the module establishes a security baseline according to these statistical feature quantities, serving as an important reference for evaluating whether subsequent real-time data deviates from the normal state. The establishment of the security baseline enables the system to more accurately identify potential security risks and provides a scientific evaluation benchmark for the risk assessment module. At the same time, the regular update of the baseline also ensures that the system can adapt to the changing security environment and maintain the accuracy and effectiveness of its evaluation.

[0082] The risk assessment module is used to construct a dynamic risk assessment model, match the corresponding sub-risk assessment models, calculate the deviation degree and analyze it, and generate a second analysis result.

[0083] The risk assessment module is a crucial link in the railway safety shielding system. It uses a dynamic risk assessment model to calculate according to real-time data and the established security baseline by matching the corresponding sub-risk assessment models. By calculating the deviation degree and analyzing the results, the module can generate a second analysis result, which directly reflects the current security risk level of the railway operation environment. The dynamic nature and accuracy of the risk assessment module enable the system to quickly respond to environmental changes, timely adjust security strategies, and effectively reduce the probability of safety accidents.

[0084] The hierarchical response module is used to set the response strategy levels according to the second analysis result and output response instructions.

[0085] The hierarchical response module sets the hierarchical response strategy according to the second analysis result output by the risk assessment module. This means that in the face of security risks at different levels, the system can take corresponding response measures, such as activating early warnings, adjusting operation strategies, or conducting emergency evacuations. By outputting response instructions, the module guides on-site personnel or automated equipment to execute corresponding security measures, ensuring the safety and stability of railway operations. The introduction of the hierarchical response mechanism not only improves the system's ability to respond to risks but also ensures the pertinence and effectiveness of response measures, providing a solid guarantee for railway safety.

[0086] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. Among them, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM for short), electrically erasable programmable read-only memory (EEPROM for short), erasable programmable read-only memory (EPROM for short), programmable read-only memory (PROM for short), read-only memory (ROM for short), magnetic memory, flash memory, a magnetic disk, or an optical disc. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction device that implements the functions specified in one process Figure 1 one process or multiple processes and / or boxes Figure 1 specified in one box or multiple boxes.

[0087] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not restrictive. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A sensor-based railway safety shielding method, characterized in that, It includes the following steps: Step S1, perform spatio-temporal correlation analysis on abnormal events in historical environmental data, and output the first analysis result; Step S2, dynamically adjust the sensor sampling frequency according to the first analysis result, collect real-time data and store it in the edge node; Step S3, calculate statistical feature quantities based on the first analysis result, and establish a safety baseline according to the statistical feature quantities; Step S4, construct a dynamic risk assessment model, match the corresponding sub-risk assessment models, calculate the deviation degree and analyze it, and generate the second analysis result; Step S5, set the response strategy grading according to the second analysis result, and output a response instruction.

2. The railway safety shielding method based on a sensor according to claim 1, characterized in that The said Step S1 includes the following sub-steps: Step S101, retrieve historical environmental data from the distributed sensor network, where the historical environmental data includes historical vibration spectra, train stop time series, abnormal events, and passenger flow distribution; Step S102, perform spatio-temporal correlation analysis on abnormal events in historical environmental data, and use the Moran spatial autocorrelation index to quantify event clustering: ； where N is the total number of abnormal events, is the Moran's index, is the spatial weight matrix, is the severity level at the -th time, is the severity level at the -th time, is the average severity level, and are event indices; If is greater than a preset first autocorrelation index, it is determined as a high-risk aggregation area and the output is the first analysis result.

3. The method for railway safety shielding based on sensors according to claim 2, characterized in that, The said Step S2 includes the following sub-steps: Step S201, if the Moran index of the high-risk aggregation area is greater than or equal to the second autocorrelation index, output a signal to increase the sampling frequency of the vibration sensor; If the Moran index is between the first autocorrelation index and the second autocorrelation index, maintain the initial sampling frequency; Step S202, collect track vibration acceleration signals, GPS positioning, and passenger heat maps, and calculate the vibration energy spectral density: ； Among them, is the sampling window, is the frequency, is the vibration energy spectral density, is the historical vibration spectrum; Step S203, compress and transmit the vibration energy spectral density, GPS positioning, and passenger heat maps to the edge node for storage.

4. The method for railway safety shielding based on sensors according to claim 3, characterized in that, The said Step S3 includes the following sub-steps: Step S301, segment the historical vibration spectrum by time window, and calculate the statistical feature quantities of each segment of data: ； ； ； Among them, is the divided difference, is the variance, is the peak factor, is the maximum amplitude of the historical vibration spectrum within the time window, is the vibration amplitude of the th sampling point of the historical vibration spectrum within the time window; Step S302, taking as the range, establish the safety baseline of vibration energy, train braking distance and passenger density.

5. The method for railway safety shielding based on sensors according to claim 4, characterized in that, The said Step S4 includes the following sub-steps: Step S401, construct a dynamic risk assessment model based on LSTM: ； Among them, is in the hidden state, is the current input, is the environmental weight coefficient, is the real-time risk score, is the input vector at the current moment, is the contribution weight of a certain dimension in the input feature or hidden state to the risk score, is the current moment; Step S402, match the sub-model according to the train operation stage, and calculate the deviation degree between the real-time data and the baseline: ； Among them, is the deviation degree, is the th feature in the input vector at the current moment, is the standard deviation corresponding to the th feature in the baseline; Step S403, if the deviation degree continuously exceeds the preset deviation threshold for the preset first continuous duration, mark it as an emergency risk, otherwise mark it as a potential risk, and output it as the second analysis result.

6. The method for railway safety shielding based on sensors according to claim 5, wherein, The said Step S5 includes the following sub-steps: Step S501, set the response strategy grading: Emergency risk: Activate the sound and light alarm + close the physical barrier, and the closing time is less than or equal to the preset first closing duration; Potential risk: Start the warning projection + local air pressure buffering, and the response delay is less than or equal to the preset first delay duration; Output the response instruction; Step S502, send the response instruction to the platform edge controller through the 5G-Uu interface, and synchronously update the status of the digital twin platform.

7. The method for railway safety shielding based on sensors according to claim 6, characterized in that, The logic of the closing of the physical barrier is: Generate a flexible barrier deformation path according to the train car body contour, and the path function satisfies: ； Among them, is a path function, is the amplitude, is the attenuation coefficient, is the phase shift, is the wave number, is the position; Adjust the actuator displacement of the shape memory alloy through a PID controller: ； Among them, is the actuator displacement, is the contour matching error, is the PID control proportionality, is the PID control integral, is the PID control differential coefficient.

8. The method for railway safety shielding based on sensors according to claim 7, characterized in that The update of the digital twin platform includes: Map the real-time sensor data to the virtual platform model, and render a three-dimensional risk heat map; Generate an evacuation path optimization plan based on the three-dimensional risk heat map, and the path cost function of the evacuation path optimization plan is: ； Among them, is the path length, is the risk value of the area passed by the path, and are the weight coefficients, and is the comprehensive cost of the evacuation path.

9. The method for railway safety shielding based on sensors according to claim 8, wherein It further includes step S6, triggering the sensor calibration process during the daily low peak period, outputting calibration error data. If the calibration error data is less than the preset calibration error threshold, the infrared passenger counting error is corrected by the laser rangefinder. If the calibration error data is greater than the preset calibration error threshold, switch to the backup sensor and report it to the operation and maintenance system.

10. A sensor-based railway safety shielding system, which is applied to a sensor-based railway safety shielding method as described in any one of claims 1-8, characterized in that, It includes an association analysis module, a sensor adjustment module, a baseline establishment module, a risk assessment module, and a hierarchical response module; The association analysis module is used to perform spatio-temporal association analysis on abnormal events in historical environmental data and output the first analysis result; The sensor adjustment module is used to dynamically adjust the sensor sampling frequency according to the first analysis result, collect real-time data and store it in the edge node; The baseline establishment module is used to calculate statistical feature quantities based on the first analysis result and establish a safety baseline according to the statistical feature quantities; The risk assessment module is used to construct a dynamic risk assessment model, match the corresponding sub-risk assessment models, calculate and analyze the deviation degree, and generate the second analysis result; The hierarchical response module is used to set the response strategy level according to the second analysis result and output a response instruction.

Images