Urban Tunnel Intelligent Safety Operation and Maintenance Method and System Combining Multiple Data

Abstract

The present invention discloses an intelligent safety operation and maintenance method and system for urban tunnels integrating multiple data, specifically related to the field of intelligent transportation operation and maintenance, which is used to solve the problems of prediction deviation and congestion management caused by tunnel traffic fluctuations. By constructing a fluctuation feature mapping matrix, the entrance peak features are systematically expressed, and combined with the traffic parameters of multi-point detectors, the wave similarity analysis model is used to accurately evaluate the propagation path and target section of the fluctuations. At the same time, with the help of the time series superposition correction mechanism, the historical fluctuation law is fused with real-time data to generate a dynamically corrected vehicle flow transfer prediction curve, achieving a balance between physical constraints and prediction accuracy. Finally, by dynamically updating the traffic flow prediction model, key indicators are output to guide the efficient linkage of the signal coordination and induction control modules, thereby being able to minimize the congestion accumulation and traffic risks caused by the propagation of fluctuations to the greatest extent, optimize the tunnel operation and maintenance efficiency, and enhance the traffic safety guarantee ability.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation operation and maintenance, and more specifically, to an intelligent safety operation and maintenance method and system for urban tunnels combining multi-data. Background Art

[0002] In the urban traffic network, tunnels, as important transportation hubs, carry a large volume of vehicle traffic and connect complex ground road networks and highway systems. To improve traffic efficiency and ensure safe operation, short-cycle signal lights are usually installed at tunnel entrances to dynamically regulate the incoming vehicle flow. However, this frequent signal switching causes vehicles to flood into the tunnel in waves, forming periodic traffic flow pulses, which significantly increases the complexity of traffic management inside the tunnel. The non-linear propagation characteristics of wave-like traffic flows may lead to problems such as local congestion, uneven traffic flow, and accident risks inside the tunnel, posing great challenges to intelligent safety operation and maintenance.

[0003] Under short-cycle signal control, the vehicle flow at the tunnel entrance shows periodic fluctuations and propagates deep into the tunnel in waves, forming a dynamic and complex traffic situation. However, the existing technologies lack accurate identification of the entrance fluctuation characteristics and modeling of the propagation laws, and are unable to combine the real-time traffic data provided by multi-point detectors inside the tunnel to dynamically evaluate the fluctuation propagation path and its impact on the congestion risk of the target section. As a result, the traffic flow prediction model cannot effectively correct the dynamic impact of entrance fluctuations on the internal traffic flow of the tunnel. Therefore, there is an urgent need for an intelligent method based on multi-data fusion to improve the accuracy of traffic flow prediction and congestion risk assessment inside the tunnel by accurately identifying entrance fluctuations and dynamically modeling the fluctuation propagation laws, providing support for intelligent safety operation and maintenance.

[0004] To solve the above problems, a technical solution is provided as follows. Summary of the Invention

[0005] To overcome the above-mentioned defects of the prior art, embodiments of the present invention provide an intelligent safety operation and maintenance method and system for urban tunnels that combines multi-data. By constructing a fluctuation feature mapping matrix, the entrance wave peak features are systematically expressed for the first time, and combined with the traffic parameters of multi-point detectors, a wave similarity analysis model is used to accurately evaluate the propagation path of fluctuations and the target section; at the same time, with the help of a time series superposition correction mechanism, the historical fluctuation law is fused with real-time data to generate a dynamically corrected vehicle flow transfer prediction curve, achieving a balance between physical constraints and prediction accuracy; finally, through the dynamic update of the flow prediction model, key indicators such as the maximum flow upper limit and the critical queuing diffusion duration are accurately output, guiding the efficient linkage of the signal coordination and induction control modules, which can significantly improve the timeliness and accuracy of traffic flow prediction in tunnels under complex fluctuation scenarios, minimize congestion accumulation and traffic risks caused by the propagation of fluctuations to the greatest extent, optimize the tunnel operation and maintenance efficiency, and enhance the traffic safety guarantee ability to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An intelligent safety operation and maintenance method for urban tunnels that combines multi-data, including the steps:

[0008] S1. Use the short-period signal switching log and vehicle detection data, and use the wave amplitude extraction algorithm to determine the time window and peak scale of the rapid influx of vehicles in each cycle, and generate a fluctuation feature mapping matrix;

[0009] S2. Combine the mapping matrix with the traffic detection data in the tunnel, and predict the congestion propagation path of fluctuations in the tunnel through wave similarity analysis, and identify the target section and the corresponding time window;

[0010] S3. Iteratively correct the traffic density, queuing duration, and wave amplitude attenuation of the target section, use the historical fluctuation case library for time series superposition calculation, and output the vehicle flow transfer prediction curve of real-time fluctuation feedback;

[0011] S4. Update the parameters of the traffic flow prediction model of the tunnel operation and maintenance system according to the vehicle flow transfer prediction curve, and issue dynamic control instructions in a timely manner through the external signal light coordination and lane induction linkage mechanism.

[0012] In a preferred embodiment, step S1 includes the following contents:

[0013] First, obtain the start time and end time of each signal cycle from the short-period signal switching log at the tunnel entrance, where represents the th signal cycle; at the same time, obtain the real-time vehicle entry data within each signal cycle from the vehicle detection device;

[0014] For each signal cycle , segment and statistically analyze the vehicle entry data within the corresponding cycle at fixed time intervals to generate a vehicle entry rate sequence , where represents the th time period; subsequently, define the wave amplitude as the cumulative increment of the vehicle entry rate within the corresponding signal cycle;

[0015] After determining the wave amplitude, if the wave amplitude value exceeds the wave peak threshold , it is considered that there is a significant surge in vehicle flow within the corresponding signal cycle; next, calculate the wave peak moment within the signal cycle , which is defined as the specific time point at which the cumulative increment reaches the maximum value within the signal cycle

[0016] For each signal cycle , if , add the corresponding record item to the fluctuation feature mapping matrix.

[0017] In a preferred embodiment, step S2 includes the following contents:

[0018] In the multi-point detector, first define a traffic parameter set based on time series , where represents the detector position label denotes the discretized time index;

[0019] Associate data retrieval and alignment: Traverse the traffic parameter set of the multi-point detector, and filter out the detector positions adjacent to the wave peak moment or covering within the time period and their corresponding time series ; Align the discretized time index with the wave peak moment. When there is a time difference between the sampling interval of the detector and the entrance data, use the same discretization scale to intercept and store the detector sequence in the intermediate data table;

[0020] Associate kernel function and wave similarity calculation: Define an association kernel function between each wave peak record and its aligned intermediate data table to measure the similarity between the wave peak and the actual traffic flow. Let represent the traffic parameter mapping vector of detector at time , represent the incremental feature mapping of the wave peak entry in the fluctuation feature mapping matrix; The association kernel function is in the form of: ; where and is a positive constant value used to regulate the attenuation rate of the difference in traffic characteristics in the exponential term.

[0021] In a preferred embodiment, step S2 further includes the following:

[0022] Based on the correlation kernel function, calculate the similarity mean or maximum value at several moments to quantify the degree of coincidence between the current wave peak and the actually measured traffic flow of the detector, and obtain the fluctuation pattern matching degree;

[0023] Congestion propagation path evaluation: After obtaining the fluctuation pattern matching degree, based on the tunnel geographical topology order and the spatial distribution of multi-point detectors, evaluate the potential traffic queue extension caused by the wave peak section by section; for each detector position , if the fluctuation pattern matching degree exceeds a certain critical threshold, it is considered that the vehicle speed at this detector position drops significantly or the occupancy rate continues to rise under the influence of the current wave peak, and is marked as a potential congestion trigger point; then, according to the sequential numbering of the detectors in the tunnel, recursively from the entrance to the deep part of the tunnel, and record the specific time interval when the congestion trigger point appears at each position , and finally obtain a set of target sections forming congestion accumulation and their corresponding time windows ;

[0024] Write all the obtained target sections, congestion trigger points, and corresponding time windows into the congestion propagation matrix in tabular form.

[0025] In a preferred embodiment, step S3 includes the following:

[0026] Historical case retrieval and mapping: According to the , target sections and congestion trigger time intervals recorded in the congestion propagation matrix, select from the historical fluctuation case library a reference entry with a similar entrance fluctuation pattern and similar section length and queue evolution characteristics corresponding to , denoted as ; then, retrieve the corresponding historical traffic flow delay function ; where represents the queue growth rate of the reference entry within the duration after entering the tunnel;

[0027] Fluctuation correction based on the principle of time series superposition: For the traffic density detected within the target section​ and queueing duration , define a set of fluctuation correction equations based on time series superposition; at the time , let ; where ; among them represents the correction coefficient at the position at the moment , is a pre-designed fluctuation superposition function; through recurrence in a similar form, calculate the fluctuation correction amount point by point along the to direction, and perform the following progression between adjacent positions: if and are separated by , then ; where is the index number of the detectors arranged in sequence in the tunnel, is the section length correction function.

[0028] In a preferred embodiment, step S3 further includes the following content:

[0029] Comprehensive consideration of the attenuation factor and traffic capacity: When performing superposition, if the wave amplitude is less than the wave peak threshold or the congestion formation time window exceeds a specific range, then introduce the attenuation factor to reflect the degree of influence of the entrance wave peak gradually decreasing with the travel distance of the vehicle flow, and add the upper limit of the traffic capacity in the middle and rear sections when including the large flow bottleneck section, to constrain the cumulative value of the superposition function not to exceed the physical capacity limit;

[0030] Generation of the vehicle flow transfer prediction curve: After completing the time series superposition and fluctuation correction, for each position within the target section at consecutive moments , obtain the corrected density and queueing duration, and fit the vehicle flow transfer prediction curve ; store each and the sequence of vehicle flow transfer prediction curves corresponding to the target section in the real-time fluctuation prediction table.

[0031] In a preferred embodiment, step S4 includes the following content:

[0032] Dynamic extraction of key parameters of the traffic flow prediction model: According to the target section in the real-time fluctuation prediction table, the corresponding consecutive moments and the corrected vehicle flow transfer prediction curve , first extract the core indicators that can reflect the tunnel accommodation capacity and queue evolution intensity under the entrance fluctuation state, including the maximum flow upper limit , Queuing Diffusion Critical Duration .

[0033] In a preferred embodiment, step S4 further includes the following:

[0034] Numerical Adjustment and Parameter Mapping: For the core indicators recorded for the key parameters of the traffic flow model, combined with the vehicle flow transfer prediction curve sequence, the original traffic flow prediction model is numerically mapped and improved in the following form: ; where represents the core state of the current prediction model at time and is a non-linear correction function.

[0035] In a preferred embodiment, step S4 further includes the following:

[0036] External Module Linkage Strategy: When the updated traffic flow prediction model is obtained, two types of information are immediately injected into the corresponding modules according to the tunnel operation plan and the external signal control requirements:

[0037] 1). Entrance Signal Coordination Module: Based on the newly added maximum flow limit and queuing diffusion critical duration indicators in the traffic flow prediction model, it is determined whether to execute a tighter or looser signal switching cycle in the future according to the preset judgment criteria, and the signal coordination data object is entered in a structured form;

[0038] 2). Tunnel Internal Induction Control Module: According to the potential congestion accumulation period estimated by the new model, corresponding speed limit instructions and lane induction configurations are added to the lane induction rule object;

[0039] Finally, the updated traffic flow prediction model and the signal coordination data and lane induction rules derived therefrom are written into the data table named the updated traffic flow model together.

[0040] The urban tunnel intelligent safety operation and maintenance system combining multiple data includes: a pulse extraction module, a path calculation module, a timing correction module, and a dynamic update module;

[0041] Pulse Extraction Module: Using the short-cycle signal switching log and vehicle detection data, the wave amplitude extraction algorithm is applied to identify the surge time window and peak scale of vehicles in each cycle, generate a fluctuation feature mapping matrix, and transfer the fluctuation feature mapping matrix to the path calculation module;

[0042] Path Calculation Module: Combining the mapping matrix with the traffic detection data in the tunnel, predicting the congestion propagation path of the fluctuation in the tunnel through wave similarity analysis, and identifying the target section and the corresponding time window, and transferring the identification result to the timing correction module;

[0043] Timing correction module: Iteratively correct the target section and time window marked in the congestion propagation matrix. Combine the historical fluctuation case library, use timing superposition calculation to analyze the traffic density, queuing duration, and amplitude attenuation law, generate a traffic flow transfer prediction curve with real-time fluctuation feedback, and transfer the traffic flow transfer prediction curve to the dynamic update module;

[0044] Dynamic update module: Update the tunnel operation and maintenance model parameters according to the traffic flow transfer prediction curve, and generate a signal light cycle adjustment strategy and a lane guidance instruction.

[0045] Technical effects and advantages of the intelligent and safe operation and maintenance method and system for urban tunnels combining multiple data of the present invention:

[0046] The present invention systematically expresses the inlet wave peak characteristics for the first time by constructing a fluctuation feature mapping matrix, and combines the traffic parameters of multiple point detectors. Using the wave similarity analysis model, it accurately evaluates the propagation path and target section of the fluctuation; at the same time, with the help of the timing superposition correction mechanism, it fuses the historical fluctuation law and real-time data to generate a dynamically corrected traffic flow transfer prediction curve, achieving a balance between physical constraints and prediction accuracy; finally, through the dynamic update of the traffic flow prediction model, it accurately outputs key indicators such as the maximum traffic flow upper limit and the critical queuing diffusion duration, guides the efficient linkage of the signal coordination and induction control modules, can significantly improve the timeliness and accuracy of traffic flow prediction in the tunnel under complex fluctuation scenarios, minimize congestion accumulation and traffic risks caused by fluctuation propagation, optimize the tunnel operation and maintenance efficiency, and enhance the traffic safety guarantee ability. Brief description of the drawings

[0047] Figure 1 It is a schematic flow chart of the intelligent and safe operation and maintenance method for urban tunnels combining multiple data of the present invention;

[0048] Figure 2 It is a schematic structural diagram of the intelligent and safe operation and maintenance system for urban tunnels combining multiple data of the present invention. Detailed implementation manners

[0049] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0050] Embodiment 1: Figure 1 The intelligent and safe operation and maintenance method for urban tunnels combining multiple data of the present invention is given, including:

[0051] S1. Switch the log of short - cycle signals and vehicle detection data, use the wave amplitude extraction algorithm to determine the time window and peak scale of the rapid influx of vehicles in each cycle, and generate a fluctuation feature mapping matrix;

[0052] S2. Combine the mapping matrix with the traffic detection data in the tunnel, predict the congestion propagation path of the fluctuation in the tunnel through wave similarity analysis, and identify the target section and the corresponding time window;

[0053] S3. Iteratively correct the traffic density, queue length, and wave amplitude attenuation in the target section, perform time - series superposition calculation using the historical fluctuation case library, and output the vehicle flow transfer prediction curve of real - time fluctuation feedback;

[0054] S4. Update the parameter of the traffic flow prediction model of the tunnel operation and maintenance system according to the vehicle flow transfer prediction curve, and issue dynamic control instructions in a timely manner through the external signal light coordination and lane induction linkage mechanism.

[0055] With the increasing complexity of the urban traffic network, as a key traffic hub, the operation efficiency and safety of traffic tunnels directly affect the overall traffic flow and urban operation. However, the frequent switching of short - cycle signal lights at the tunnel entrance causes vehicles to pour into the tunnel in waves, forming a non - linear and highly fluctuating traffic flow pattern. This volatility not only poses significant challenges to traditional traffic flow prediction models but also may lead to prediction errors, thus affecting the real - time traffic management and safety warning in the tunnel. Therefore, there is an urgent need for a method that can accurately identify and quantify the vehicle flow fluctuations caused by short - cycle signal control to optimize the traffic flow prediction model in the tunnel, improve the response speed of the traffic management system, and enhance the decision - making accuracy.

[0056] Step S1 is based on the short - cycle signal switching log at the tunnel entrance and the real - time output of vehicle detection equipment. By constructing a wave amplitude extraction algorithm, it statistically segments the vehicle entry times in each signal cycle, marks the sudden increase in vehicle flow in each time window using the wave peak threshold, and then stores the obtained short - cycle fluctuation sequence in the fluctuation feature mapping matrix. It specifically includes the following content:

[0057] First, obtain the start time and end time of each signal cycle from the short - cycle signal switching log at the tunnel entrance, where represents the th signal cycle. At the same time, obtain the real - time vehicle entry data in each signal cycle from the vehicle detection equipment, where Represents a specific point in time. To ensure data consistency and accuracy, it is necessary to perform time synchronization and format standardization on the original data to eliminate possible clock deviations and data format differences between different devices.

[0058] For each signal cycle , segment and statistically analyze the vehicle entry data within the corresponding cycle at fixed time intervals (for example, once per second) to generate a vehicle entry rate sequence , where represents the th time period. Subsequently, define the wave amplitude as the cumulative increment of the vehicle entry rate within the corresponding signal cycle, that is ; where is the total number of time periods within the signal cycle . This metric is used to quantify the degree of surge in the vehicle entry rate for each signal cycle.

[0059] After determining the wave amplitude, if the wave amplitude value exceeds the peak threshold , it is considered that there is a significant surge in vehicle flow within the corresponding signal cycle. Next, calculate the peak time within the signal cycle , which is defined as the specific point in time within the signal cycle at which the cumulative increment reaches the maximum:

[0060] This peak time represents the strongest moment of the surge in the vehicle entry rate and has crucial identification significance. For each signal cycle , if , add the corresponding record item to the fluctuation feature mapping matrix : where is the signal cycle number, is the wave amplitude, and are the start and end times of the signal cycle respectively. Through this process, the fluctuation feature mapping matrix will systematically record all significant vehicle flow surge events to form a complete short - cycle fluctuation sequence.

[0061] To ensure efficient invocation and reference of the fluctuation feature information in subsequent steps, store the fluctuation feature mapping matrix in a dedicated database in the form of a structured data table.

[0062] Step S1 constructs a fluctuation feature mapping matrix by combining short - cycle signal switching logs with real - time vehicle detection data, accurately identifying the traffic flow surges and their peak moments within each signal cycle. This method not only realizes the detailed quantification of vehicle flow fluctuations but also ensures the structured storage of data, facilitating the efficient invocation and analysis in subsequent steps. The specific benefits include: significantly improving the accuracy of the flow prediction model under short - cycle signal regulation, reducing misjudgments and delays caused by prediction deviations; providing a reliable data basis for the accurate assessment of congestion propagation paths, thus optimizing the traffic control strategies in the tunnel; and enhancing the system's scalability and data processing efficiency through structured data management. These advantages jointly promote the overall effectiveness of the urban tunnel intelligent safety operation and maintenance system, ensuring the efficient and safe operation of the tunnel.

[0063] During the creation stage of the fluctuation feature mapping matrix, the traffic flow surges caused by short - cycle signal switching have been carefully quantified and stored in the unified data structure, the fluctuation feature mapping matrix. Next, it is necessary to combine the fluctuation feature mapping matrix with the multi - dimensional traffic parameters provided by multi - point detectors in the tunnel. Using the correlation kernel function configured in the wave similarity analysis module, identify and evaluate the congestion propagation paths that may be triggered by these significant peaks, thus laying a foundation for subsequent time - series superposition correction and flow prediction. To enable subsequent steps to accurately obtain the congestion evolution information of each target section, this process will produce a data structure containing the congestion propagation time window and section identification, preparing for the execution of fluctuation correction and transfer prediction in the next stage.

[0064] After obtaining the fluctuation feature mapping matrix in Step S2, call the actual traffic parameters uploaded by multi - point detectors in the tunnel. Using the wave similarity analysis module relying on the built - in correlation kernel function, evaluate the congestion propagation paths triggered under the current fluctuation pattern, and perform congestion transfer estimation for each peak in the mapping matrix, extracting the target sections and corresponding time windows where congestion accumulates. The specific contents include the following:

[0065] In the multi - point detector, first define a traffic parameter set based on time series , where represents the detector position label, represents the discretized time index. At the same time, the fluctuation feature mapping matrix retains several entries, where the peak moment is used to locate the central moment of the vehicle flow surge. In this step, for each peak moment, the following operations need to be carried out:

[0066] Associated data retrieval and alignment: Traverse the traffic parameter set of the multi - point detector , filter out the detector positions adjacent to or covering the time period corresponding to the peak moment and their corresponding time series To ensure data consistency, it is necessary to align the discretized time index with the peak time. When there is a time difference between the sampling interval of the detector and the inlet data, the detector sequence is intercepted using the same discretization scale and stored in the intermediate data table to provide directly callable aligned traffic flow features for subsequent analysis.

[0067] Association kernel function and wave similarity calculation: Define an association kernel function between each peak record and its aligned intermediate data table to measure the similarity between the peak and the actual traffic flow. Let represent the traffic parameter mapping vector of the detector at time , and represent the incremental feature mapping of the peak entry in the fluctuation feature mapping matrix. The association kernel function is in the form of: ; where and are positive constants of , which are used to regulate the attenuation rate of the difference in traffic characteristics in the exponential term.

[0068] Based on the association kernel function, calculate the mean or maximum similarity at several times within to quantify the degree of coincidence between the current peak and the measured traffic flow of the detector, obtain the fluctuation pattern matching degree, and record the result as .

[0069] Congestion propagation path evaluation: After obtaining the fluctuation pattern matching degree, based on the tunnel geographical topology order and the spatial distribution of multi-point detectors, evaluate the potential traffic queue extension caused by the peak segment by segment. For each detector position , if the fluctuation pattern matching degree exceeds a certain critical threshold , it is considered that there is a high probability of a significant speed drop or a continuous increase in occupancy at this detector position under the influence of the current peak. Then is marked as a potential congestion trigger point. Subsequently, according to the sequential numbering of the detectors in the tunnel, recursively push from the entrance to the deep part of the tunnel, and record the specific time interval when the congestion trigger point appears at each position , and finally obtain a set of target sections forming congestion accumulation and their corresponding time windows .

[0070] Write all the above obtained target sections, congestion trigger points, and corresponding time windows into the congestion propagation matrix , whose structure includes , where Corresponding to the signal period identifier in the fluctuation feature mapping matrix, used to determine the detector section that may be covered by congestion, characterizing the formation of the congestion time window. After this table object is associated with the fluctuation feature mapping matrix, it will be read and used in the subsequent time series superposition correction and prediction steps. In this way, it can be ensured that in the next stage, when correcting according to the fluctuation features, the quantitative information of the congestion propagation path can be accurately called.

[0071] Step S2 establishes a close mapping relationship between the fluctuation features at the entrance and the spatio-temporal traffic flow data of the multi-point detector. It not only calculates the wave similarity but also mines the congestion accumulation sections and time windows that the wave peaks may trigger inside the tunnel based on the correlation kernel function. Corresponding one-to-one with the fluctuation feature mapping matrix in Step S1, the congestion propagation matrix generated in Step S2 stores the congestion probability information and target section positioning matching each wave peak entry, providing a key quantitative basis for the subsequent time series superposition correction and transfer prediction. In the face of the scenario of frequent switching of short-period signals, the results of this step can more accurately capture the impact degree of traffic fluctuations on the internal traffic situation of the tunnel, thus laying a solid data foundation for subsequent fluctuation correction and dynamic induction decision-making.

[0072] The multi-point detector is a network of traffic parameter acquisition devices deployed at multiple key positions inside the tunnel, aiming to capture real-time multi-dimensional traffic operation state data including vehicle flow, vehicle speed, lane occupancy rate, queue length, etc. Its core function is to accurately depict the spatio-temporal traffic state inside the tunnel through a distributed sensing layout, providing a dynamic data sequence for each monitoring point at different time periods. These detectors usually include various sensing technologies such as inductive loops, microwave radars, and video monitoring devices, which can record traffic parameter changes in a continuous or discrete form, providing basic support for congestion propagation path evaluation and traffic management. Through the collaborative work of multi-point detectors, a full-coverage data acquisition network can be formed in space, thus realizing the real-time monitoring and multi-level analysis of the traffic situation inside the tunnel.

[0073] In Step S1, the vehicle fluctuation information caused by the short-period signal switching is recorded in the fluctuation feature mapping matrix; in Step S2, the fluctuation feature mapping matrix and the traffic data of the multi-point detector are matched through the correlation kernel function to generate a congestion propagation matrix, clarifying the target sections and time windows where congestion may accumulate. At this stage, based on the above data structure and the typical fluctuation evolution curves stored in the historical fluctuation case library, the traffic density and queue length of the target section are iteratively corrected using the time series superposition principle, and dynamic factors such as the entrance wave amplitude attenuation rate and the mid- and rear-section traffic capacity are comprehensively considered, and finally a vehicle flow transfer prediction curve with real-time fluctuation feedback is formed.

[0074] Step S3 compares the obtained congestion propagation path and the target section results with the historical fluctuation case library, invokes the fluctuation correction mechanism based on the time series superposition principle, iteratively corrects the traffic density and vehicle queuing duration of each target section, and comprehensively considers the dynamic changes of the entrance wave amplitude attenuation rate and the mid- and rear-section traffic capacity to form a predicted curve of vehicle flow transmission with real-time fluctuation feedback, so as to reflect how the entrance wave peak deforms or fades at different positions in the tunnel depth. The specific contents are as follows:

[0075] Historical case retrieval and mapping: According to the , target section and congestion trigger time interval recorded in the congestion propagation matrix, select from the historical fluctuation case library a reference entry with an entrance fluctuation pattern similar to that of and having a similar section length and queuing evolution characteristics, denoted as . Subsequently, retrieve the corresponding historical vehicle flow delay function ; where represents the queuing growth rate of the reference entry within duration after entering the tunnel, which is used to reflect the accumulation or attenuation of the entrance wave amplitude deep in the tunnel. Temporarily store this delay function in an intermediate data object named historicalDelayMap.

[0076] Fluctuation correction based on the time series superposition principle: For the detected traffic density and queuing duration within the target section, define a set of fluctuation correction equations based on time series superposition. For example, at time at , let ; where represents the correction coefficient at position at time is a pre-designed fluctuation superposition function used to quantify the superposition coupling between the cumulative impact revealed by the historical delay function and the currently observed occupancy characteristics to calculate a comprehensive correction amount, reflecting the superposition impact of historical fluctuation laws on the current traffic state. The aim is to capture the complex relationship between wave peak accumulation, attenuation, and queuing growth rate to perform a more practical correction on the traffic flow prediction model. The specific form can be defined as: ; where represents the delay function extracted from historical fluctuation cases, quantifying the cumulative queuing growth rate caused by the wave peak on the propagation path, the current target section at time The traffic density, represents the maximum traffic density of the section and is used to standardize the current density level. is the queuing decay factor, which controls the rate at which the amount of superimposed fluctuations gradually weakens as the queuing length increases. The density amplification factor is used to quantify the degree of influence of traffic density on the superimposed correction.

[0077] Through a recurrence of a similar form, along to the direction, the fluctuation correction amount is calculated point by point, and the following progression is performed between adjacent positions: If and are apart, then ; where is the index number of the detectors arranged in sequence in the tunnel. Here, is a section length correction function that transmits the fluctuation influence of the upstream position at time to through the spatial distance , simulating the propagation attenuation of the wave crest in the tunnel depth, so as to reflect the transmission law of the wave crest in the tunnel depth. The specific form can be defined as: ; where, is the distance attenuation factor, which is used to control the decreasing rate of the wave crest intensity as the distance increases, and its value is determined by the tunnel structure and traffic flow characteristics.

[0078] Comprehensive consideration of the attenuation factor and traffic capacity: When performing the above superposition, if the wave amplitude is less than the wave crest threshold or the congestion formation time window exceeds a specific range, an additional attenuation factor needs to be introduced to reflect the degree of influence of the entrance wave crest gradually decreasing with the travel distance of the vehicle flow, and the upper limit of the mid-to-late section traffic capacity is added when including the large flow bottleneck section, restricting the cumulative value of the superposition function not to exceed the physical capacity limit. For example, when calculating , if the fluctuation has started to significantly attenuate at , then can be set; thus correcting or reducing the over-accumulated fluctuation effect.

[0079] Generation of the vehicle flow transmission prediction curve: After completing the time series superposition and fluctuation correction, for each position within the target section at consecutive times , the corrected density and queuing duration are obtained, and then the vehicle flow transmission prediction curve is fitted; used to show how the entrance wave crest is deferred or dissipated in the tunnel depth. Each and the target section The corresponding traffic flow transfer prediction curve sequence It is stored in the real-time fluctuation prediction table realTimeWavePredict, providing directly referable transfer prediction results for the implementation of key parameter updates and the linkage of external control strategies in the next step.

[0080] Through the retrieval and mapping based on historical fluctuation cases, combined with the principle of time series superposition, multiple iterative corrections of the traffic density and queuing duration in the target section are realized. In step S3, a traffic flow transfer prediction curve covering the attenuation law of the cumulative effect of the entrance peak is finally generated. This curve intuitively reveals how congestion evolves or attenuates in different sections after vehicles pour in pulsively from the entrance. The structured storage method of this result ensures that relevant prediction data can be conveniently called in the next step, so as to dynamically update the key parameters of the traffic flow prediction model and link the external signal lights and the induction strategy in the tunnel in step S4, achieving more accurate congestion prevention and control and safe operation and maintenance in the environment of frequent short-cycle signal switching.

[0081] In steps S1 and S2, the construction from the fluctuation feature mapping matrix to the congestion propagation matrix has been completed, and the target section and its corresponding time window have been identified; in step S3, using the historical fluctuation case library and the principle of time series superposition, multiple rounds of fluctuation correction have been carried out on the traffic density and queuing duration of the above target section, and finally a real-time fluctuation prediction table recording the traffic flow transfer prediction results is generated. Next, based on the prediction curve described in the real-time fluctuation prediction table, the key parameters of the tunnel traffic flow prediction model need to be updated in real time, and the updated model information is distributed to the external signal light coordination module and the tunnel internal induction control module, so as to take effective traffic flow control measures before predicting the upcoming traffic flow accumulation.

[0082] Based on the obtained traffic flow transfer prediction curve in step S4, the key parameters of the traffic flow prediction model of the tunnel intelligent operation and maintenance system are updated in real time, and this updated information is simultaneously distributed to the external signal light coordination module and the tunnel internal induction control module, enabling the operation and maintenance platform to quickly and dynamically adjust the signal light timing and the lane guidance measures in the tunnel before predicting the upcoming traffic flow accumulation, so as to continuously optimize the overall traffic state and maximize the alleviation of the congestion or safety risks brought by the traffic flow fluctuation. Specifically, it includes the following contents:

[0083] Dynamic extraction of the key parameters of the traffic flow prediction model: According to the target section in the real-time fluctuation prediction table , the corresponding consecutive moments and the corrected traffic flow transfer prediction curve , first extract several core indicators that can reflect the tunnel accommodation capacity and the intensity of queuing evolution under the entrance fluctuation state, including the maximum flow upper limit , the critical queuing diffusion duration etc. And file them in the key parameters of the defined data structure traffic model. This ensures that the input in the subsequent numerical adjustment stage is consistent with the prediction results of the third step.

[0084] Numerical adjustment and parameter mapping: For each index recorded in the key parameters of the traffic model, such as the maximum traffic flow upper limit and the critical queuing diffusion duration, combined with the sequence of vehicle flow transfer prediction curves generated in step S3, the following form is used to perform numerical mapping and improvement on the original traffic flow prediction model: ; where represents the core state of the current prediction model at time and is a non-linear correction function used to absorb the congestion accumulation and fluctuation attenuation information calculated in step S3 and superimpose it on the original model, enabling the model to more sensitively respond to the dynamic impact of the entrance wave on traffic operation. The specific form can be defined as: ; where is the fluctuation response intensity coefficient used to adjust the overall amplitude of the correction amount, determined according to historical data statistics, and

[0085] External module linkage strategy: When the updated traffic flow prediction model is obtained, immediately inject the following two types of information into the corresponding modules according to the tunnel operation plan and external signal control requirements:

[0086] 1). Entrance signal coordination module: Based on the new maximum traffic flow upper limit, critical queuing diffusion duration and other indexes in the traffic flow prediction model, judge whether to execute a tighter or looser signal switching cycle in the future according to the preset judgment criteria, and record the signal coordination data object in a structured form; the signal coordination data is used to record and transfer the adjustment information of the signal control module, such as the recommended signal cycle or switching strategy.

[0087] 2). Tunnel internal induction control module: According to the potential congestion accumulation period estimated by the new model, add the corresponding speed limit instructions and lane induction configurations to the lane induction rule object to complete the pre-deployment before the specific time interval arrives.

[0088] Finally, write the updated traffic flow prediction model and the signal coordination data and lane induction rules derived from it into the data table named updated traffic flow model. This table is bidirectionally associated with the real-time fluctuation prediction table to enable quick retrieval and further iterative modification of the traffic flow prediction model in case of new wave surges in the future.

[0089] By deeply analyzing the traffic flow transfer information contained in the real-time fluctuation prediction table, step S4 completes the real-time calibration of the key parameters of the tunnel traffic flow prediction model, generates an updated model that is tightly coupled with the entrance fluctuation state, and immediately distributes the corresponding parameters to the external signal light coordination module and the tunnel internal induction control module. This can effectively improve the accuracy of traffic flow prediction in the scenario of frequent short-cycle signal switching, provide a precise basis for formulating reasonable speed limit, diversion, and signal cycle plans before the peak impact arrives, thereby preventing or alleviating potential congestion accumulation, and ensuring traffic efficiency and safety. The output updated traffic flow model is bidirectionally connected to the real-time fluctuation prediction table in a structured manner, and can be further corrected and optimized according to subsequent newly added fluctuation information to optimize the overall traffic flow prediction framework, enabling the tunnel operation and maintenance platform to maintain an agile response and continuous optimization ability to fluctuation risks in a highly dynamic traffic environment.

[0090] Embodiment 2: Figure 2 The urban tunnel intelligent safety operation and maintenance system combining multiple data according to the present invention is provided, including: a pulse extraction module, a path calculation module, a timing correction module, and a dynamic update module;

[0091] Pulse extraction module: Using the short-cycle signal switching log and vehicle detection data, applying the wave amplitude extraction algorithm to identify the surge time window and peak scale of vehicles in each cycle, generating a fluctuation feature mapping matrix, and transmitting the fluctuation feature mapping matrix to the path calculation module;

[0092] Path calculation module: Combining the mapping matrix with the traffic detection data in the tunnel, predicting the congestion propagation path of the fluctuation in the tunnel through wave similarity analysis, and identifying the target section and the corresponding time window, and transmitting the identification result to the timing correction module;

[0093] Timing correction module: Iteratively correct the target section and time window marked in the congestion propagation matrix, combine the historical fluctuation case library, use the timing superposition calculation to analyze the traffic density, queuing duration, and wave amplitude attenuation law, generate a traffic flow transfer prediction curve for real-time fluctuation feedback, and transmit the traffic flow transfer prediction curve to the dynamic update module;

[0094] Dynamic update module: Update the tunnel operation and maintenance model parameters according to the traffic flow transfer prediction curve, and generate a signal light cycle adjustment strategy and a lane induction instruction.

[0095] The above formulas are all dimensionless and take their numerical calculations. The formulas are obtained by collecting a large amount of data for software simulation to obtain a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0096] Only some exemplary embodiments of the present invention have been described by way of illustration. Without doubt, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0097] It should be noted that in this text, if there are relational terms such as first and second, they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0098] The above are only the specific embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered by the scope of protection of the present application. Therefore, the scope of protection of the present application shall be subject to the scope of protection of the claims.

Claims

1. An intelligent and safe operation and maintenance method for urban tunnels combining multiple data, characterized in that: Includes steps: S1. Use the short-period signal switching log and vehicle detection data, use the wave amplitude extraction algorithm to determine the time window and peak scale of the rapid influx of vehicles in each cycle, and generate a fluctuation feature mapping matrix; S2. Combine the mapping matrix with the traffic detection data in the tunnel, predict the congestion propagation path of the wave in the tunnel through wave similarity analysis, and identify the target section and the corresponding time window; S3. Iterate and correct the traffic density, queue length and amplitude attenuation of the target section, use the historical fluctuation case library to perform time series superposition calculation, and output the traffic flow prediction curve with real-time fluctuation feedback; Step S3 includes the following contents: Historical case retrieval and mapping: Based on the records in the congestion propagation matrix , Target Segment and congestion triggering time interval , select from the historical fluctuation case library The corresponding reference items with similar entrance fluctuation shapes and similar segment lengths and queue evolution characteristics are denoted as ; Then, retrieve the corresponding historical traffic delay function ;in Indicates a reference item After entering the tunnel The growth rate of the queue within the duration; Fluctuation correction based on the principle of time series superposition: The traffic density detected in the target section and waiting time , define a set of fluctuation correction equations based on time series superposition; in Time On, order ;in express Position at time The correction factor of is a pre-designed wave superposition function; through a similar form of recursion, along to The fluctuation correction amount is calculated point by point in the direction, and the following progression is performed between adjacent positions: and Distance ,but ;in The detectors are indexed sequentially in the tunnel. is the segment length correction function; Comprehensive consideration of attenuation factor and traffic capacity: When superimposing, if the wave amplitude is less than the peak threshold or the congestion time window exceeds a specific range, the attenuation factor is introduced. , to reflect the impact of the entrance peak gradually decreasing as the traffic flow travels further, and to add the upper limit of the capacity of the middle and rear sections when including the large flow bottleneck section , constrain the cumulative value of the superposition function not to exceed the upper limit of the physical capacity; Generation of traffic flow prediction curve: After completing time series superposition and fluctuation correction, the target section Each position In continuous moments The corrected density and queue duration are obtained, and the traffic flow transfer prediction curve is fitted. ; Each The traffic flow transfer prediction curve sequence corresponding to the target section is stored in the real-time fluctuation prediction table; S4. Update the traffic prediction model parameters of the tunnel operation and maintenance system according to the traffic flow prediction curve, and issue dynamic control instructions in a timely manner through external signal light coordination and lane induction linkage mechanism; Step S4 includes the following contents: Dynamic extraction of key parameters of the traffic prediction model: predicting the target segment in the table based on real-time fluctuations , corresponding to the continuous moments And the revised traffic flow prediction curve First, we extract the core indicators that can reflect the tunnel capacity and queue evolution intensity under the entrance fluctuation state, including the maximum flow limit , Queue diffusion critical time ; Numerical adjustment and parameter mapping: Based on the core indicators recorded by the key parameters of the traffic model and combined with the traffic flow prediction curve sequence, the original traffic flow prediction model is numerically mapped and improved in the following ways: ;in Indicates that the current prediction model is at time The core status is a nonlinear correction function.

2. The intelligent safety operation and maintenance method for urban tunnels combined with multiple data according to claim 1 is characterized in that: Step S1 includes the following contents: First, obtain the start time of each signal cycle from the short-period signal switching log at the tunnel entrance and end time ,in Indicates signal cycle; at the same time, real-time vehicle entry data in each signal cycle is obtained from the vehicle detection equipment; For each signal cycle , the vehicle entry data in the corresponding period is segmented and counted at fixed time intervals to generate a vehicle entry rate sequence ,in Indicates time periods; then, define the wave amplitude is the cumulative increment of vehicle entry rate in the corresponding signal period; After the wave amplitude is determined, if the wave amplitude value exceeds the peak threshold , it is considered that there is a significant increase in vehicle traffic in the corresponding signal period; Next, calculate the peak time within the corresponding signal period , which is defined as the period of the signal The specific time point within which the cumulative increment reaches the maximum value; For each signal cycle ,like , then add the corresponding record item to the fluctuation feature mapping matrix.

3. The intelligent safety operation and maintenance method for urban tunnels combined with multiple data according to claim 2 is characterized in that: Step S2 includes the following contents: In the multi-point detector, first define the traffic parameter set based on time series ,in represents the detector position number, Represents a discretized time index; Related data retrieval and alignment: traverse the traffic parameter set of multiple point detectors and filter out the traffic parameters that are adjacent to or covered by the peak time. Detector position within the time period The corresponding time series ; Align the discretized time index with the peak time. When there is a time difference between the sampling interval of the detector and the input data, the detector sequence is intercepted using the same discretization scale and stored in the intermediate data table. Correlation kernel function and wave similarity calculation: Define the correlation kernel function between each peak record and its aligned intermediate data table To measure the similarity between the peak and the actual traffic flow, let Indicates the detector At the moment The traffic parameter mapping vector at Indicates peak entry The incremental feature map in the fluctuation feature map matrix; the associated kernel function is as follows: ;in and is a positive constant used to regulate the decay rate of the difference in traffic characteristics in the exponential term.

4. The intelligent safety operation and maintenance method for urban tunnels combined with multiple data according to claim 3 is characterized in that: Step S2 also includes the following contents: Based on the correlation kernel function, Calculate several moments The similarity mean or maximum value is used to quantify the degree of agreement between the current peak and the flow rate measured by the detector, and obtain the fluctuation shape matching degree; Congestion propagation path assessment: After obtaining the wave shape matching degree, the potential traffic queue extension caused by the wave crest is evaluated section by section based on the tunnel geographical topology order and the spatial distribution of multi-point detectors; For each detector position If the fluctuation shape matching degree exceeds a certain critical threshold, it is considered that the vehicle speed at this detector position has dropped significantly or the occupancy rate has continued to rise under the influence of the current peak. Mark as potential congestion trigger points; Then, the detectors are numbered in the order in which they are located in the tunnel. , recursively from the entrance to the depth of the tunnel and records it at each location The specific time interval of the congestion trigger point , and finally a set of target sections for congestion accumulation is obtained and its corresponding time window ; All the obtained target sections, congestion triggering points and corresponding time windows are written into the congestion propagation matrix in a tabular form.

5. The intelligent safety operation and maintenance method for urban tunnels combined with multiple data according to claim 4 is characterized in that: Step S4 also includes the following contents: External module linkage strategy: When the updated traffic prediction model is obtained After that, the following two types of information are immediately injected into the corresponding modules according to the tunnel operation plan and external signal light control requirements: 1) Entrance signal light coordination module: Based on the newly added maximum flow upper limit and queue diffusion critical time indicators in the flow prediction model, it is determined whether a tighter or looser signal switching cycle needs to be implemented in the future according to the preset judgment criteria, and the signal coordination data object is entered in a structured form; 2) Tunnel internal guidance control module: According to the potential congestion accumulation period estimated by the new model, add corresponding speed limit instructions and lane guidance configurations in the lane guidance rule object; Finally, the updated traffic prediction model and the signal coordination data and lane induction rules derived therefrom are written into a data table named updated traffic model.

6. An intelligent safety operation and maintenance system for urban tunnels combined with multiple data, used to implement the intelligent safety operation and maintenance method for urban tunnels combined with multiple data as described in any one of claims 1 to 5, characterized in that: include: Pulse extraction module, path calculation module, timing correction module and dynamic update module; Pulse extraction module: uses short-period signal switching logs and vehicle detection data, applies wave amplitude extraction algorithm to identify the surge time window and peak size of vehicles in each cycle, generates fluctuation feature mapping matrix, and passes the fluctuation feature mapping matrix to the path calculation module; Path calculation module: combines the mapping matrix with the traffic detection data in the tunnel, predicts the congestion propagation path of the wave in the tunnel through wave similarity analysis, identifies the target section and the corresponding time window, and transmits the identification result to the timing correction module; Time series correction module: It iteratively corrects the target sections and time windows marked in the congestion propagation matrix, combines the historical fluctuation case library, uses time series superposition calculation to analyze the traffic density, queue length and amplitude attenuation law, generates a traffic flow transfer prediction curve with real-time fluctuation feedback, and transmits the traffic flow transfer prediction curve to the dynamic update module; Dynamic update module: Updates tunnel operation and maintenance model parameters according to the traffic flow prediction curve, and generates traffic light cycle adjustment strategies and lane induction instructions.

Images