Intelligent command and dispatch system for emergency rescue in highway tunnels

Abstract

The present invention relates to the technical field of highway tunnel rescue, and specifically to an intelligent command and dispatch system for emergency rescue in highway tunnels, including a data acquisition module, a data processing module, a real-time situation adaptation module, a decision support module, and a resource scheduling module, wherein: the data acquisition module is responsible for collecting real-time data inside and outside the tunnel; the data processing module analyzes and processes the collected real-time data; the real-time situation adaptation module identifies abnormal situations in the data; the decision support module provides rescue decision suggestions; and the resource scheduling module allocates and optimizes rescue resources according to the suggestions of the decision support module. The present invention makes rescue decisions faster and more accurate, significantly improves the response speed to emergency situations, and ensures that rescue operations can be performed within the most critical time.

Description

Technical Field

[0001] The invention relates to the technical field of highway tunnel rescue, and in particular to an intelligent command and dispatch system for emergency rescue in highway tunnels. Background Art

[0002] Tunnel accidents often result in serious casualties and property losses due to their special environment and conditions. In traditional rescue systems, rescue responses are often limited by the lack of understanding of the accident site and environmental conditions, resulting in delayed rescue operations and improper resource allocation. In addition, rescue operations inside tunnels usually require fast and accurate decision support to respond to sudden and changing situations, such as traffic accidents, fires, or leaks of hazardous substances.

[0003] Existing rescue systems often lack effective data integration capabilities and are unable to obtain comprehensive tunnel status information in real time, which limits the ability of the rescue command center to respond quickly in emergency situations. At the same time, the evaluation models in existing technologies often cannot be updated in real time and cannot accurately reflect the current tunnel environment and traffic conditions, further reducing rescue efficiency and resource utilization.

[0004] In addition, the deployment of rescue resources often fails to effectively take into account real-time traffic and environmental changes, resulting in the delay of rescue resources or waste of resources. These problems seriously affect the overall performance and response speed of the tunnel rescue system, thereby increasing the risks and consequences of tunnel accidents. Summary of the invention

[0005] Based on the above purpose, the present invention provides an intelligent command and dispatch system for emergency rescue in highway tunnels.

[0006] The intelligent command and dispatch system for emergency rescue in highway tunnels includes a data acquisition module, a data processing module, a real-time situation adaptation module, a decision support module and a resource dispatch module, among which;

[0007] The data acquisition module is responsible for collecting real-time data inside and outside the tunnel, including environmental data, vehicle operation data and emergency event information;

[0008] The data processing module analyzes and processes the collected real-time data to generate real-time rescue information, and uses the evaluation model in combination with historical data to perform predictive analysis;

[0009] The real-time situation adaptation module analyzes the real-time data and historical data, identifies abnormalities in the data, and dynamically adjusts the evaluation model in the data processing module;

[0010] The decision support module provides rescue decision suggestions based on the analysis results of the data processing module and the real-time situation adaptation module;

[0011] The resource scheduling module allocates and optimizes rescue resources according to the suggestions of the decision support module.

[0012] Furthermore, the data acquisition module includes:

[0013] Environmental monitoring unit: Environmental sensors are deployed inside the tunnel and at the tunnel entrance and exit to collect real-time temperature, humidity, smoke concentration and harmful gas levels;

[0014] Vehicle Monitoring Unit: Continuously monitors the speed, density and behavior patterns of vehicles in the tunnel through cameras and traffic sensors installed on the tunnel ceiling and walls on both sides;

[0015] Emergency Detection Unit: Utilizes vibration sensors, infrared sensors and additional video analysis software to monitor emergency situations in the tunnel, such as accidents, fires or safety obstacles.

[0016] Furthermore, the data processing module includes:

[0017] Real-time data analysis: Use data analysis algorithms to analyze collected environmental data, vehicle operation data, and emergency event information in real time to identify and assess the current safety status and risks inside the tunnel;

[0018] Data integration: Combine historical data records with real-time data analysis to analyze data trends and behavior patterns in the tunnel;

[0019] Application of assessment models: Combining historical and real-time data, the assessment model is used to conduct predictive analysis to predict rescue needs and event development trends.

[0020] Furthermore, the real-time data analysis includes:

[0021] Feature selection: Select the features that have an impact on predicting the safety status and risks inside the tunnel from the collected real-time data. Use the information gain method to select features. The calculation formula is:

[0022] ;

[0023] in, It is a dataset The entropy of It is a feature, It is a feature The value of The data subset at that time;

[0024] Risk assessment analysis: Evaluate the current safety status inside the tunnel and the risk of emergency events. Use a logistic regression model to estimate the probability of an event (such as a traffic accident or fire). The logistic regression model is expressed as:

[0025] ;

[0026] in, is the probability of the event occurring, is the influencing factor from feature selection;

[0027] Time series analysis: To predict the environment and traffic trends inside the tunnel in the future, the autoregressive integrated moving average model (ARIMA) is used to predict time series data. The autoregressive integrated moving average model is expressed as:

[0028] ;

[0029] in, It's time point The predicted value of and are the model parameters of the autoregressive integrated moving average model, is the error term.

[0030] Furthermore, the data integration includes:

[0031] Data fusion: Integrate real-time data and historical data to create a unified data view and provide a basis for analysis;

[0032] Trend analysis: Use moving average and exponential smoothing for trend analysis. The calculation formula is:

[0033] Moving Average : ;

[0034] in, It's time point data points, is the time window size;

[0035] Exponential Smoothing: ;

[0036] in, It's time point The smoothed value of It's time point data points, is the smoothing coefficient;

[0037] Behavior pattern recognition: The K-means clustering model is used to identify the behavior patterns of vehicles and environmental data in the tunnel. The K-means clustering model is expressed as:

[0038] ;

[0039] in, is the number of clusters, is a data point, It is clustering center.

[0040] Furthermore, the evaluation model adopts a time convolutional network (TCN) model, and the time convolutional network (TCN) model includes:

[0041] Data preparation: collect and preprocess real-time data and historical data. Preprocessing includes data standardization, missing value processing and feature extraction.

[0042] Constructing TCN model: Use temporal convolutional network to capture the temporal dependencies in the data and predict future rescue needs and event developments. Use dilated convolutional layers to process sequence data. For each dilated convolution operation, the calculation formula is:

[0043] ;

[0044] in, It is The convolution kernel of the layer, is the expansion rate, is the size of the convolution kernel, is the bias term, is the output of the previous layer;

[0045] Introducing adaptive feature fusion: Dynamically fuse features of different time scales to adapt to different prediction needs and enhance the prediction accuracy of the model. The calculation formula is:

[0046] ;

[0047] in, and denote the features extracted from fine-grained and coarse-grained time windows, respectively. It is The learnable fusion parameters of the layer, is the sigmoid function, represents element-wise multiplication;

[0048] Introducing weighted residual connections: Enhance the learning and information transmission capabilities of the network, and adjust the information flow between different layers through weighting factors. The calculation formula is:

[0049] ;

[0050] in, It is The input of the layer, is the output after dilated convolution and activation function, It is The weighted residual connection weights of the layer;

[0051] Training model: Use historical data to train the TCN model, optimize the TCN model parameters to minimize the prediction error, select mean square error (MSE) as the loss function, and perform model training;

[0052] Prediction and application: The trained TCN model is used to predict the rescue needs and event development trends in the tunnel in real time. By inputting real-time data into the model, rescue demand predictions and probability assessments of event development are obtained to provide data support for emergency response.

[0053] Furthermore, the real-time situation adaptation module includes:

[0054] Dynamic data fusion: Fusion of real-time data with historical data to provide an understanding of current tunnel conditions;

[0055] Abnormal situation detection and adaptation: Analyze the fused real-time data and historical data through dynamic time window analysis (DTWA), automatically detect abnormal patterns, identify abnormalities or behaviors that deviate from the normal in the current data, and adjust countermeasures in time;

[0056] Real-time update of evaluation model: Dynamically adjust the parameters of the evaluation model through online learning methods to adapt to new data patterns or emergencies, and automatically update the evaluation model when changes or data trend deviations are identified.

[0057] Furthermore, the dynamic time window analysis (DTWA) method includes:

[0058] Data Window Construction: Construct a fixed-length data window to capture behaviors and patterns in time series data, for each data point Create a window with a size of , and calculate the average value within the window and standard deviation , the calculation formula is:

[0059] ;

[0060] ;

[0061] Anomaly score calculation: By comparing the real-time data point with the statistical features within the window, the anomaly score is calculated to identify anomalies or deviations from normal behavior in the data. The window-based Z-score is used to evaluate the anomaly score of the current point. The calculation formula is:

[0062] ;

[0063] like The absolute value of exceeds the predetermined threshold , then it is believed that It is an outlier;

[0064] Dynamic window adjustment: dynamically adjust the window size based on past anomaly detection results , through the most recent time interval The number of anomalies detected within Adjust the window size, the calculation formula is:

[0065] ;

[0066] in, is The number of anomalies detected during the time period, is the threshold of the number of anomalies, is the adjustment sensitivity parameter.

[0067] Furthermore, the real-time updating of the evaluation model includes:

[0068] Real-time model update mechanism: Use online learning methods to dynamically adjust the parameters of the evaluation model, allowing the evaluation model to be updated in real time when new data is received, thereby adapting to new data patterns and emergencies. The calculation formula is:

[0069] ;

[0070] in, and are the evaluation model parameters before and after the update, is the evaluation model, is the learning rate, is the mean squared error loss function for evaluating model parameters The gradient of is the newly received data point, is the corresponding true value or target value;

[0071] Change and trend deviation detection: Monitor and identify changes or deviations in data trends. By implementing sliding window technology, it is used to calculate the real-time prediction accuracy of the evaluation model. The calculation formula is:

[0072] ;

[0073] in, is the number of data points in the window, is the indicator function, It is The actual value of the observation point, Is the evaluation model for The predicted value of each observation point.

[0074] Furthermore, the decision support module includes:

[0075] Integrated data analysis: The decision support module receives analysis results from the data processing module, including traffic flow, environmental monitoring data and timely updated event reports. At the same time, the real-time situation adaptation module provides anomaly detection results and evaluation model update feedback to obtain the current tunnel status and rescue needs;

[0076] Rescue decision suggestion generation: Based on the integrated data analysis results received, rescue decision suggestions are generated, including dispatching rescue teams and medical support to the scene, starting ventilation equipment and closing tunnel entrances, adjusting traffic lights and optimizing rescue routes;

[0077] Dynamic decision adjustment: Dynamically adjust rescue decisions based on updated real-time data and scenario analysis, including updating navigation and guiding the use of alternative routes, increasing rescue forces, and updating safety guidance.

[0078] Beneficial effects of the present invention:

[0079] The present invention, by integrating a data acquisition module, a data processing module and a real-time situation adaptation module, can accurately monitor and analyze the environment, traffic and emergency event information inside and outside the tunnel in real time. This highly integrated information processing capability makes rescue decisions more rapid and accurate, significantly improves the response speed to emergency situations, and ensures that rescue operations can be executed within the most critical time.

[0080] The present invention provides scientific rescue decision suggestions and resource allocation plans based on the collected and processed data through the decision support module and the resource scheduling module. This dynamic prediction adjustment and resource optimization function not only reduces the waste of resources caused by inaccurate predictions, but also ensures the efficient use of rescue resources and enhances the overall efficiency and effectiveness of the rescue operation.

[0081] The present invention can adapt to complex and changeable tunnel environments and emergencies through the dynamic data fusion and abnormal situation detection functions of the real-time situation adaptation module and the real-time update mechanism of the evaluation model. This high degree of adaptability and real-time update capability enables the system to adjust rescue strategies and action plans in a timely manner, thereby improving the stability and reliability of the system and ensuring that efficient operation can be maintained under various circumstances. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0083] Figure 1A schematic diagram of system function modules according to an embodiment of the present invention;

[0084] Figure 2 Schematic diagram of a real-time context adaptation module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0085] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0086] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0087] like Figure 1-Figure 2 As shown, the intelligent command and dispatch system for emergency rescue in highway tunnels includes a data acquisition module, a data processing module, a real-time situation adaptation module, a decision support module and a resource dispatch module, wherein;

[0088] The data acquisition module is responsible for collecting real-time data inside and outside the tunnel, including environmental data, vehicle operation data and emergency event information;

[0089] The data processing module analyzes and processes the collected real-time data, generates real-time rescue information, and uses the evaluation model combined with historical data for predictive analysis;

[0090] The real-time situation adaptation module analyzes real-time and historical data to identify anomalies in the data and dynamically adjusts the evaluation model in the data processing module to improve the adaptability and accuracy of the prediction;

[0091] The decision support module provides rescue decision suggestions based on the analysis results of the data processing module and the real-time situation adaptation module;

[0092] The resource scheduling module allocates and optimizes rescue resources based on the recommendations of the decision support module;

[0093] The highway tunnel emergency rescue intelligent command and dispatch system of the present invention significantly improves the accuracy and response speed of rescue decisions through integrated data processing and real-time situation adaptation capabilities. The system's dynamic prediction adjustment and resource optimization functions ensure the efficient use of rescue resources and reduce resource waste caused by inaccurate predictions.

[0094] The data acquisition module includes:

[0095] Environmental monitoring unit: Environmental sensors are deployed inside the tunnel and at the tunnel entrance and exit to collect real-time data on temperature, humidity, smoke concentration, and harmful gas levels. These data help assess the safety conditions in the tunnel and the impact of the external environment;

[0096] Vehicle Monitoring Unit: Through cameras and traffic sensors installed on the tunnel ceiling and both sides of the wall, the speed, density and behavior patterns of vehicles in the tunnel are continuously monitored. This information facilitates understanding of traffic flow and identification of possible traffic accidents;

[0097] Emergency Detection Unit: uses vibration sensors, infrared sensors and additional video analysis software to monitor emergency situations in the tunnel, such as accidents, fires or safety obstacles;

[0098] Through the organic integration of the above units, the data acquisition module collects key data inside and outside the tunnel in real time and accurately, providing necessary input for other modules of the system, ensuring that rescue operations can be efficiently dispatched and decision-made based on the latest and most comprehensive information.

[0099] The data processing module includes:

[0100] Real-time data analysis: Use data analysis algorithms to analyze collected environmental data, vehicle operation data, and emergency event information in real time to identify and assess the current safety status and risks inside the tunnel;

[0101] Data integration: Combine historical data records with real-time data analysis to analyze data trends and behavior patterns in the tunnel, and improve the depth and accuracy of data analysis;

[0102] Evaluation model application: Combine historical and real-time data to conduct predictive analysis through evaluation models to predict rescue needs and event development trends;

[0103] In this way, the data processing module can provide accurate data analysis and prediction for the system, support decision-making and resource optimization of the decision support module and resource scheduling module. This module ensures that rescue operations can be based on scientific analysis and data-driven foresight, thereby improving the efficiency and effectiveness of rescue operations.

[0104] Real-time data analysis includes:

[0105] Feature selection: Select the features that have an impact on predicting the safety status and risks inside the tunnel from the collected real-time data. Use the information gain method to select features. The calculation formula is:

[0106] ;

[0107] in, It is a dataset The entropy of It is a feature, It is a feature The value of The data subset at that time;

[0108] Risk assessment analysis: Evaluate the current safety status inside the tunnel and the risk of emergency events. Use a logistic regression model to estimate the probability of an event (such as a traffic accident or fire). The logistic regression model is expressed as:

[0109] ;

[0110] in, is the probability of the event occurring, is the influencing factor from feature selection;

[0111] Time series analysis: To predict the environment and traffic trends inside the tunnel in the future, the autoregressive integrated moving average model (ARIMA) is used to predict time series data. The autoregressive integrated moving average model is expressed as:

[0112] ;

[0113] in, It's time point The predicted value of and are the model parameters of the autoregressive integrated moving average model, is the error term;

[0114] Through the above content, the highway tunnel emergency rescue intelligent command and dispatch system can effectively monitor and predict the safety status inside the tunnel in real time, and provide accurate data support to optimize the command and resource dispatch of rescue operations. The logistic regression model provides insights into the probability of emergency events, and the ARIMA model helps predict the future direction of the environment and traffic, so that the rescue team can prepare in advance and reduce potential risks and impacts.

[0115] Data integration includes:

[0116] Data fusion: Integrate real-time data and historical data to create a unified data view and provide a basis for analysis;

[0117] Trend analysis: Use moving average and exponential smoothing for trend analysis. The calculation formula is:

[0118] Moving Average : ;

[0119] in, It's time point data points, is the time window size;

[0120] Exponential Smoothing: ;

[0121] in, It's time point The smoothed value of It's time point data points, is the smoothing coefficient;

[0122] Behavior pattern recognition: The K-means clustering model is used to identify the behavior patterns of vehicles and environmental data in the tunnel. The K-means clustering model is expressed as:

[0123] ;

[0124] in, is the number of clusters, is a data point, It is clustering the center of

[0125] Through this historical data integration function, the highway tunnel emergency rescue intelligent command and dispatch system can effectively combine historical and real-time data for comprehensive analysis. Time series analysis provides an in-depth understanding of data trends, while cluster analysis reveals behavioral patterns in tunnel use, such as traffic peaks or environmental risk patterns. These analysis results not only improve the efficiency of data utilization, but also enhance the ability to predict events and manage emergency responses, making rescue operations more precise and timely.

[0126] The evaluation model uses the temporal convolutional network (TCN) model, which includes:

[0127] Data preparation: collect and preprocess real-time data and historical data. Preprocessing includes data standardization, missing value processing and feature extraction.

[0128] Constructing TCN model: Use temporal convolutional network to capture temporal dependencies in data and predict future rescue needs and event developments. Use dilated convolutional layers to process sequence data and expand the receptive field to cover dependencies over a longer time range. For each dilated convolution operation, the calculation formula is:

[0129] ;

[0130] in, It is The convolution kernel of the layer, is the expansion rate, is the size of the convolution kernel, is the bias term, is the output of the previous layer;

[0131] Introducing adaptive feature fusion: Dynamically fuse features of different time scales to adapt to different prediction needs and enhance the prediction accuracy of the model. The calculation formula is:

[0132] ;

[0133] in, and denote the features extracted from fine-grained and coarse-grained time windows, respectively. It is The learnable fusion parameters of the layer, is the sigmoid function, represents element-wise multiplication;

[0134] The calculation formula of the sigmoid function is:

[0135] ;

[0136] in, is the output of the sigmoid function, is the input value, is the base of natural logarithms;

[0137] Introducing weighted residual connections: Enhance the learning and information transmission capabilities of the network, and adjust the information flow between different layers through weighting factors. The calculation formula is:

[0138] ;

[0139] in, It is The input of the layer, is the output after dilated convolution and activation function, It is The weighted residual connection weights of the layer;

[0140] Training model: Use historical data to train the TCN model, optimize the TCN model parameters to minimize the prediction error, select mean square error (MSE) as the loss function, and perform model training;

[0141] The mean square error (MSE) is calculated as:

[0142] ;

[0143] in, It is The actual observed value, It is Prediction values, is the total number of data points;

[0144] Prediction and application: Use the trained TCN model to predict the rescue needs and event development trends in the tunnel in real time. By inputting real-time data into the model, the rescue demand prediction and event development probability assessment are obtained to provide data support for emergency response;

[0145] The temporal convolutional network (TCN) model can effectively process and predict complex time series data in highway tunnels. The model improves the utilization efficiency of historical and real-time data through adaptive feature fusion and weighted residual connection, enhances the accuracy of prediction and the stability of the network, which is extremely useful for decision-making in dynamic and complex highway tunnel emergency rescue scenarios, and can help rescue teams allocate resources and plan responses more effectively.

[0146] The real-time situation adaptation module includes:

[0147] Dynamic data fusion: Fusion of real-time data with historical data to provide an understanding of current tunnel conditions;

[0148] Abnormal situation detection and adaptation: Analyze the fused real-time data and historical data through dynamic time window analysis (DTWA), automatically detect abnormal patterns, identify abnormalities or behaviors that deviate from the normal in the current data, and adjust countermeasures in time;

[0149] Real-time update of evaluation model: Dynamically adjust the parameters of the evaluation model through online learning methods to adapt to new data patterns or emergencies, and automatically update the evaluation model when changes or data trend deviations are identified;

[0150] Through the real-time situation adaptation module, the highway tunnel emergency rescue intelligent command and dispatch system can more accurately predict and respond to various situations in the tunnel. The dynamic data fusion and real-time update functions of the evaluation model of this module ensure that the system can maintain efficient operation under various sudden or changing environmental conditions.

[0151] Dynamic Time Window Analysis (DTWA) includes:

[0152] Data Window Construction: Construct a fixed-length data window to capture behaviors and patterns in time series data, for each data point Create a window with a size of , and calculate the average value within the window and standard deviation , the calculation formula is:

[0153] ;

[0154] ;

[0155] Anomaly score calculation: By comparing the real-time data point with the statistical features within the window, the anomaly score is calculated to identify anomalies or deviations from normal behavior in the data. The window-based Z-score is used to evaluate the anomaly score of the current point. The calculation formula is:

[0156] ;

[0157] like The absolute value of exceeds the predetermined threshold (for example, 3.0), then it is considered It is an outlier;

[0158] Dynamic window adjustment: dynamically adjust the window size based on past anomaly detection results , which allows the system to better adapt to changes in the data, thereby improving detection accuracy by The number of anomalies detected within Adjust the window size, the calculation formula is:

[0159] ;

[0160] in, is The number of anomalies detected during the time period, is the threshold of the number of anomalies, It is a sensitivity adjustment parameter used to adjust the change range of the window size;

[0161] The present invention dynamically adjusts the data window size through the DTWA algorithm, and can flexibly respond to rapid changes and complexity in the data. It is particularly suitable for monitoring and responding to traffic and environmental monitoring data in complex environments such as highway tunnels, supporting decision makers to respond quickly based on the latest and most accurate information, and improving the efficiency and effectiveness of emergency response.

[0162] Real-time updates of the evaluation model include:

[0163] Real-time model update mechanism: Use online learning methods to dynamically adjust the parameters of the evaluation model, allowing the evaluation model to be updated in real time when new data is received, thereby adapting to new data patterns and emergencies. The calculation formula is:

[0164] ;

[0165] in, and are the evaluation model parameters before and after the update, is the evaluation model, is the learning rate, which controls the step size of parameter updates, is the mean squared error loss function for evaluating model parameters The gradient of is the newly received data point, is the corresponding true value or target value;

[0166] Change and trend deviation detection: Monitor and identify changes or deviations in data trends. By implementing sliding window technology, it is used to calculate the real-time prediction accuracy of the evaluation model. The calculation formula is:

[0167] ;

[0168] in, is the number of data points in the window, is the indicator function, It is The actual value of the observation point, Is the evaluation model for The predicted value of the observation point, when the predicted value Equal to the true value The value is 1 when it is, otherwise it is 0;

[0169] Indicator function The calculation formula is:

[0170] ;

[0171] Through the real-time update mechanism, the evaluation model can continuously optimize itself to more accurately predict emergency situations and normal behavior patterns in the tunnel. Online learning allows the model to respond quickly to new data inputs, especially when responding to emergencies, and can immediately adjust the prediction strategy. Since there is no need for offline retraining, model updates will not interrupt the normal operation of the system, which is crucial to maintaining continuous monitoring of tunnel transportation and safety.

[0172] The decision support modules include:

[0173] Integrated data analysis: The decision support module receives analysis results from the data processing module, including traffic flow, environmental monitoring data and timely updated event reports. At the same time, the real-time situation adaptation module provides anomaly detection results and evaluation model update feedback to obtain the current tunnel status and rescue needs;

[0174] Rescue decision suggestion generation: Based on the integrated data analysis results received, rescue decision suggestions are generated, including dispatching rescue teams and medical support to the scene, starting ventilation equipment and closing tunnel entrances, adjusting traffic lights and optimizing rescue routes;

[0175] Dynamic decision adjustment: Dynamically adjust rescue decisions based on updated real-time data and scenario analysis, including updating navigation and directing the use of alternative routes, increasing rescue forces, and updating safety guidance;

[0176] The decision support module enables the highway tunnel emergency rescue intelligent command and dispatch system to handle emergencies more intelligently and effectively, ensuring the best use of rescue resources while improving the safety and efficiency of the rescue team.

[0177] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention is limited to these examples. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity.

[0178] The present invention is intended to cover all such substitutions, modifications and variations that fall within the broad scope of the claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. Highway tunnel emergency rescue intelligent command and dispatch system, characterized by: It includes data collection module, data processing module, real-time situation adaptation module, decision support module and resource scheduling module, among which; The data acquisition module is responsible for collecting real-time data inside and outside the tunnel, including environmental data, vehicle operation data and emergency event information; The data processing module analyzes and processes the collected real-time data to generate real-time rescue information, and uses the evaluation model in combination with historical data to perform predictive analysis; The real-time situation adaptation module analyzes the real-time data and historical data, identifies abnormalities in the data, and dynamically adjusts the evaluation model in the data processing module; The decision support module provides rescue decision suggestions based on the analysis results of the data processing module and the real-time situation adaptation module; The resource scheduling module allocates and optimizes rescue resources according to the suggestions of the decision support module; Wherein, the data processing module includes: Real-time data analysis: Use data analysis algorithms to analyze collected environmental data, vehicle operation data, and emergency event information in real time to identify and assess the current safety status and risks inside the tunnel; Data integration: Combine historical data records with real-time data analysis to analyze data trends and behavior patterns in the tunnel; Evaluation model application: Combine historical and real-time data to conduct predictive analysis through evaluation models to predict rescue needs and event development trends; The evaluation model adopts a time convolutional network (TCN) model, and the time convolutional network (TCN) model includes: Data preparation: collect and preprocess real-time data and historical data. Preprocessing includes data standardization, missing value processing and feature extraction. Construct TCN model: Use temporal convolutional network to capture the temporal dependencies in the data and predict future rescue needs and event developments. For each dilated convolution operation, the calculation formula is: ； in, It is The convolution kernel of the layer, is the expansion rate, is the size of the convolution kernel, is the bias term, is the output of the previous layer; Introducing adaptive feature fusion: Dynamically fuse features of different time scales to adapt to different prediction needs and enhance the prediction accuracy of the model. The calculation formula is: ； in, and denote the features extracted from fine-grained and coarse-grained time windows, respectively. It is The learnable fusion parameters of the layer, is the sigmoid function, represents element-wise multiplication; Introducing weighted residual connections: Enhance the learning and information transmission capabilities of the network, and adjust the information flow between different layers through weighting factors. The calculation formula is: ； in, It is The input of the layer, is the output after dilated convolution and activation function, It is The weighted residual connection weights of the layer; Training model: Use historical data to train the TCN model, optimize the TCN model parameters to minimize the prediction error, select mean square error as the loss function, and perform model training; Prediction and application: The trained TCN model is used to predict the rescue needs and event development trends in the tunnel in real time. By inputting real-time data into the model, rescue demand predictions and probability assessments of event development are obtained to provide data support for emergency response.

2. The highway tunnel emergency rescue intelligent command and dispatch system according to claim 1 is characterized in that: The data acquisition module comprises: Environmental monitoring unit: Environmental sensors are deployed inside the tunnel and at the tunnel entrance and exit to collect real-time information on temperature, humidity, smoke concentration, illumination, and harmful gas levels; Vehicle Monitoring Unit: Through cameras installed at the tunnel entrance and on both sides of the wall, the speed, density and behavior patterns of vehicles in the tunnel are continuously monitored; Emergency Detection Unit: Utilizes vibration sensors, infrared sensors and additional video analysis software to monitor emergency situations in the tunnel, such as accidents, fires or safety obstacles.

3. The intelligent command and dispatch system for emergency rescue in highway tunnels according to claim 1 is characterized in that: The real-time data analysis includes: Feature selection: Select the features that have an impact on predicting the safety status and risks inside the tunnel from the collected real-time data. Use the information gain method to select features. The calculation formula is: ； in, It is a dataset The entropy of It is a feature, It is a feature The value of The data subset at that time; Risk assessment analysis: Evaluate the current safety status inside the tunnel and the risk of emergency events. Use a logistic regression model to estimate the probability of an event. The logistic regression model is expressed as: ； in, is the probability of the event occurring, is the influencing factor from feature selection; Time series analysis: To predict the environment and traffic trends inside the tunnel in the future, the autoregressive integrated moving average model is used to predict time series data. The autoregressive integrated moving average model is expressed as: ； in, It's time point The predicted value of and are the model parameters of the autoregressive integrated moving average model, is the error term.

4. The highway tunnel emergency rescue intelligent command and dispatch system according to claim 1 is characterized in that: The data integration includes: Data fusion: Integrate real-time data and historical data to create a unified data view and provide a basis for analysis; Trend analysis: Use moving average and exponential smoothing for trend analysis. The calculation formula is: Moving Average : ; in, It's time point data points, is the time window size; Exponential Smoothing: ; in, It's time point The smoothed value of It's time point data points, is the smoothing coefficient; Behavior pattern recognition: The K-means clustering model is used to identify the behavior patterns of vehicles and environmental data in the tunnel. The K-means clustering model is expressed as: ； in, is the number of clusters, is a data point, It is clustering center.

5. The intelligent command and dispatch system for emergency rescue in highway tunnels according to claim 4 is characterized in that: The real-time situation adaptation module comprises: Dynamic data fusion: Fusion of real-time data with historical data to provide an understanding of current tunnel conditions; Abnormal situation detection and adaptation: Analyze the fused real-time data and historical data through dynamic time window analysis, automatically detect abnormal patterns, identify abnormalities or behaviors that deviate from the normal in the current data, so as to adjust countermeasures in time; Real-time update of evaluation model: Dynamically adjust the parameters of the evaluation model through online learning methods to adapt to new data patterns or emergencies, and automatically update the evaluation model when changes or data trend deviations are identified; Wherein, the dynamic time window analysis method includes: Data Window Construction: Construct a fixed-length data window to capture behaviors and patterns in time series data, for each data point Create a window with a size of , and calculate the average value within the window and standard deviation , the calculation formula is: ； ； Anomaly score calculation: By comparing the real-time data point with the statistical features within the window, the anomaly score is calculated to identify anomalies or deviations from normal behavior in the data. The window-based Z-score is used to evaluate the anomaly score of the current point. The calculation formula is: ； like The absolute value of exceeds the predetermined threshold , then it is believed that It is an outlier; Dynamic window adjustment: dynamically adjust the window size based on past anomaly detection results , through the most recent time interval The number of anomalies detected within Adjust the window size, the calculation formula is: ； in, is The number of anomalies detected during the time period, is the threshold of the number of anomalies, is to adjust the sensitivity parameter; The real-time updating of the evaluation model includes: Real-time model update mechanism: Use online learning methods to dynamically adjust the parameters of the evaluation model, allowing the evaluation model to be updated in real time when new data is received, thereby adapting to new data patterns and emergencies. The calculation formula is: ； in, and are the evaluation model parameters before and after the update, is the evaluation model, is the learning rate, is the mean squared error loss function for evaluating model parameters The gradient of is the newly received data point, is the corresponding true value or target value; Change and trend deviation detection: Monitor and identify changes or deviations in data trends. By implementing sliding window technology, it is used to calculate the real-time prediction accuracy of the evaluation model. The calculation formula is: ； in, is the number of data points in the window, is the indicator function, It is The actual value of the observation point, Is the evaluation model for The predicted value of each observation point.

6. The intelligent command and dispatch system for emergency rescue in highway tunnels according to claim 5 is characterized in that: The decision support module includes: Integrated data analysis: The decision support module receives analysis results from the data processing module, including traffic flow, environmental monitoring data and timely updated event reports. At the same time, the real-time situation adaptation module provides anomaly detection results and evaluation model update feedback to obtain the current tunnel status and rescue needs; Rescue decision suggestion generation: Based on the integrated data analysis results received, rescue decision suggestions are generated, including dispatching rescue teams and medical support to the scene, starting ventilation equipment and closing tunnel entrances, adjusting traffic lights and optimizing rescue routes; Dynamic decision adjustment: Dynamically adjust rescue decisions based on updated real-time data and scenario analysis, including updating navigation and guiding the use of alternative routes, increasing rescue forces, and updating safety guidance.

Images