Tunnel emergency parking strip safety guiding method and system

Abstract

The invention relates to the technical field of intelligent traffic, in particular to a tunnel emergency parking strip safety guiding method and system. The method comprises the following steps: the smart workstation supports a unified vehicle-road cooperation communication protocol, so that multi-dimensional data acquired by a radar, a range finder, a telephoto camera, a snapshot camera, a vehicle-road cooperation terminal and a vehicle-mounted OBU device in real time are all sent to the smart workstation; dynamically predicting the average vehicle speed, the traffic flow and the vehicle density of the emergency parking strip and the adjacent lanes of the parking strip through a machine learning algorithm; according to a preset driving habit model, the real-time multi-dimensional data and the historical traffic data, identifying a rear-end collision risk, a side collision risk and a lane change conflict risk; and generating an information board control strategy according to the dynamic prediction result and the risk identification result. According to the tunnel emergency parking strip safety guiding method and system, the information board control strategy is generated, and it is ensured that the vehicle can safely and efficiently drive away from the emergency parking strip.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, and in particular to a method and system for guiding emergency stopping lanes in tunnels. Background Technology

[0002] With the increasing number of highway tunnels, traffic safety issues within tunnels are becoming increasingly prominent. Emergency stopping lanes, as crucial facilities within tunnels, provide vehicles with a place to stop and avoid danger in emergencies. However, due to the enclosed space and limited visibility within tunnels, vehicles exiting emergency stopping lanes are highly susceptible to collisions with vehicles on the main road, causing traffic accidents. Therefore, effectively guiding vehicles safely out of emergency stopping lanes has become a pressing issue in tunnel safety management.

[0003] Existing tunnel emergency stopping lane safety guidance systems primarily rely on simple video surveillance and manual intervention, lacking the ability to comprehensively analyze multi-dimensional traffic data and make intelligent decisions. These systems typically only provide basic early warning information and cannot dynamically adjust guidance strategies based on real-time traffic conditions, resulting in limited safety guidance effectiveness.

[0004] For example, Chinese patent publication number CN115762191A, entitled "A Tunnel Emergency Parking Lane Status Warning System and Situation Response Method," includes a telephoto camera, flashing yellow guidance signs, flashing yellow light strips, a monitoring camera, parking lane indicator lights, tunnel broadcasting, parking lane information screens, and a smart workstation. This invention captures event images within a scene using a monitoring camera and classifies them using algorithms to achieve functions such as vehicle and pedestrian behavior prediction and intelligent event recognition. The smart workstation communicates with various devices in the system, controlling each device to issue corresponding warnings based on different situations. It also establishes a remote connection with a central management platform, enabling remote reporting of event conditions and provision of remote assistance. The drawback is that it can only provide basic early warning information and cannot dynamically adjust guidance strategies based on real-time traffic conditions, resulting in limited safety guidance effectiveness. Summary of the Invention

[0005] To address the problem that existing tunnel emergency stopping lane guidance systems can only provide basic warning information and cannot dynamically adjust guidance strategies according to real-time traffic conditions, resulting in limited safety guidance effectiveness, this invention provides a tunnel emergency stopping lane safety guidance method and system that generates information board control strategies to ensure that vehicles can safely and efficiently leave the emergency stopping lane.

[0006] To achieve the above-mentioned technical objectives, the present invention provides a technical solution: a method and system for safe guidance of emergency stopping lanes in tunnels, comprising the following steps: S1, the emergency stopping lane area is equipped with radar, rangefinder, telephoto camera, snapshot camera and smart workstation, the vehicle is equipped with vehicle-road cooperative terminal and on-board unit (OBU) device, the smart workstation supports unified vehicle-road cooperative communication protocol so that the multi-dimensional data acquired in real time by radar, rangefinder, telephoto camera, snapshot camera, vehicle-road cooperative terminal and on-board unit (OBU) device are all sent to the smart workstation; S2 uses machine learning algorithms to dynamically predict the average speed, traffic flow, and vehicle density of emergency stopping lanes and adjacent lanes based on real-time multi-dimensional data and historical traffic data. S3 identifies rear-end collision risks, side collision risks, and lane change conflict risks based on preset driving habit models, real-time multi-dimensional data, and historical traffic data. S4. Based on the dynamic prediction results and risk identification results, generate the intelligence board control strategy and dynamically adjust the intelligence board indication status. S5 builds a strategy library, dynamically updates strategies based on real-time data, and sets emergency trigger conditions.

[0007] In this technical solution, by fusing multi-source data from radar, rangefinders, telephoto cameras, snapshot cameras, vehicle-road cooperative terminals, and on-board units (OBUs), the system can capture real-time vehicle dynamics, environmental conditions, and vehicle health parameters. Combined with historical traffic data, it enables dynamic prediction of average vehicle speed, traffic flow, vehicle density, and potential conflict points, providing data support for risk assessment. Based on driving habit models and real-time data, the system can identify rear-end collision, side collision, and lane-change conflict risks under different driving styles, achieving personalized and precise risk assessment and providing a basis for differentiated control strategies. Based on the prediction results and risk levels, the system generates information board control strategies, dynamically adjusts traffic light states, optimizes the passage efficiency of emergency lane exits, and reduces traffic congestion and accident risks caused by unreasonable traffic light settings. By building a strategy library and dynamically updating control strategies, the system can adapt to different traffic scenario needs. Simultaneously, emergency trigger conditions are set to ensure rapid response in emergency situations, coordinate rescue resources, and improve emergency response efficiency and safety. This ensures that vehicles can safely and efficiently leave the emergency lane.

[0008] The present invention is further configured such that the multi-dimensional data includes vehicle dynamic parameters, vehicle spacing parameters, environmental parameters, and vehicle self-parameters. The vehicle dynamic parameters include the vehicle's position, speed, acceleration, and direction within the parking area; The vehicle spacing parameter refers to the distance to vehicles in the lane adjacent to the parking lane, including both near and far distances. The environmental parameters include visibility inside the tunnel and road surface slipperiness; The vehicle's own status parameters are the vehicle's own status information within the parking area, including tire condition and engine condition.

[0009] This technical solution integrates vehicle dynamic parameters, vehicle spacing parameters, environmental parameters, and vehicle status parameters to construct a comprehensive perception system covering vehicle operation, spatial relationships, environmental conditions, and vehicle health, providing full-element data support for risk identification and decision-making. The collaborative analysis of vehicle dynamic parameters and spacing parameters can accurately capture lane-changing intentions and conflict risks. The correlation modeling of environmental parameters and vehicle status parameters can provide early warnings of loss of control risks caused by slippery road surfaces or tire abnormalities, achieving an upgrade from single-risk identification to composite risk assessment. Multi-dimensional data provides dynamic input for information board control strategies: adjusting traffic light phases based on vehicle position and spacing to avoid conflicts between vehicles queuing at exits and vehicles approaching from adjacent lanes; optimizing waiting frequency or extending warning duration based on visibility and slipperiness to enhance safety guidance in adverse environments; and prioritizing right-of-way or triggering emergency rescue lanes for vehicles with abnormal tire pressure or engine malfunctions. By continuously monitoring the vehicle's own status parameters, the system can identify potential mechanical failures in advance, such as low tire pressure or engine overheating. Before the vehicle loses control, it can guide it to a safe stop via information board signals or link with the emergency response module to initiate a rescue process, turning passive response into active defense.

[0010] The present invention is further configured such that: in step S2, the machine learning algorithm includes a time series analysis algorithm, a regression analysis algorithm, and a spatial interpolation algorithm; the step of dynamically predicting the average vehicle speed, traffic flow, vehicle density, and potential conflict points of the emergency stopping lane and adjacent lanes using the machine learning algorithm includes: Based on historical vehicle speed data, current vehicle speed data, and adjacent lane vehicle speed data, the average vehicle speed over a certain period of time in the future is predicted using a time series analysis algorithm. Based on historical traffic data, current vehicle queue length, and entrance ramp traffic flow, a regression analysis algorithm is used to predict traffic flow changes over a certain period of time in the future. Based on vehicle location data and tunnel geometric parameters, a spatial interpolation algorithm is used to calculate the current and predicted vehicle density distribution over a certain future period.

[0011] This technical solution captures the periodic fluctuations in vehicle speed based on historical and real-time vehicle speed data, such as the decreasing trend of vehicle speed during weekday morning and evening rush hours. Combined with the correlation analysis of vehicle speeds in adjacent lanes, it predicts changes in average vehicle speed in advance, providing a forward-looking basis for traffic light timing adjustments. By integrating historical traffic flow, queue length, and entrance ramp traffic flow data, it extrapolates traffic flow trends over a future period, such as the risk of queue overflow due to a surge in ramp merging traffic, enabling proactive intervention in traffic flow fluctuations. Through spatial interpolation algorithms, it fuses vehicle positioning data with tunnel geometric parameters, such as lane width and radius of curvature, to generate a real-time vehicle density distribution heatmap and predict density diffusion paths over a future period. This accurately identifies high-density areas, such as blind spots on curves and ramp merging points, supporting the spatially differentiated configuration of traffic light control strategies.

[0012] The present invention is further configured such that, in step S3, the establishment of the preset driving habit model includes: Based on drivers' historical travel data, frequency and magnitude of accelerator / brake pedal operation, lane change frequency, and following distance, drivers are divided into aggressive and conservative types. The aggressive type includes drivers with high accelerator pedal operation frequency, large brake pedal operation magnitude, high lane change frequency, and short average following distance. The conservative type includes drivers with low accelerator pedal operation frequency, small brake pedal operation magnitude, low lane change frequency, and long average following distance.

[0013] In this technical solution, a driver behavior feature database is constructed based on multi-dimensional data such as historical trip data (e.g., trip duration, route selection), pedal operation characteristics (accelerator / brake frequency and amplitude), lane change behavior (frequency and timing), and following distance. The system can quantitatively analyze two types of driving styles: aggressive (high-frequency accelerator / sudden braking, high-frequency lane changes, short following distance) and conservative (low-frequency operation, long following distance), achieving standardized representation of driving behavior. By combining the driving habit model with real-time data (e.g., current following distance, lane change intention), the system can dynamically identify the matching degree between the driver's style and current behavior. For example, if a conservative driver suddenly makes high-frequency lane changes, the system will trigger a higher risk level warning; if an aggressive driver continuously accelerates rapidly near the exit of the parking lane, the permitted passage time will be extended in advance to avoid misjudgment or missed judgment of risks caused by sudden changes in driving style. Tiered control is implemented for different driving styles: For aggressive drivers, the system limits frequent lane changes by shortening the permitted passage time, triggers flashing warnings for waiting before potential conflict points, and prioritizes guiding them to use exit lanes with lower conflict risk; for conservative drivers, the permitted waiting time is extended to avoid the risk of delays due to slow operation, and right-of-way is prioritized during low-traffic periods to reduce their waiting anxiety. Driving habit models can predict potential driver behavior in emergency stopping lane scenarios: aggressive drivers are more prone to lane-changing conflicts due to impatience, so the system intervenes by activating traffic lights before they approach the exit; conservative drivers may cause exit congestion due to hesitation, so the system optimizes their passage efficiency through dynamic signal timing. Through deep coupling of human risk and signal control, accident prevention is transformed from passive response to proactive intervention.

[0014] The present invention is further configured such that the identification of rear-end collision risk, side collision risk, and lane change conflict risk includes: Identify rear-end collision risk, side collision risk, and lane change conflict risk; calculate the rear-end collision risk index; when the risk index exceeds the first preset threshold, it is determined that there is a rear-end collision risk. The side collision risk index is calculated based on the vehicle's speed, direction, distance from vehicles in adjacent lanes, and the speed and direction of vehicles in adjacent lanes. When the risk index exceeds the second preset threshold, it is determined that there is a risk of side collision. Based on the vehicle's lane change intention, lane change speed, distance to vehicles in front and behind in the target lane, and the speed and acceleration of vehicles in front and behind in the target lane, a lane change conflict risk index is calculated. When the risk index exceeds the third preset threshold, it is determined that there is a risk of lane change conflict.

[0015] In this technical solution, the system constructs independent assessment models for rear-end collision risk, side collision risk, and lane change conflict risk. It calculates risk indices using multi-dimensional parameters (speed, distance, acceleration, direction, etc.) and sets differentiated thresholds (first / second / third preset thresholds). For example, rear-end collision risk focuses on the relative speed difference and braking distance between the vehicle and the vehicle in front; side collision risk considers the lateral speed component and lane spacing; and lane change conflict risk integrates lane change intent and target lane dynamics. This scenario-based risk modeling avoids misjudgment based on a single indicator and improves the accuracy of risk identification.

[0016] The present invention is further configured such that the information board control strategy includes: When an increase in traffic flow is predicted at the emergency stopping lane exit, the permitted passage time is extended; when a decrease in traffic flow is predicted, the permitted passage time is shortened. In areas with a high risk of rear-end collisions, extend the waiting time; in areas with a high risk of lane change conflicts, install directional traffic lights. For aggressive drivers, shorten the permitted passage time; for conservative drivers, increase the permitted passage time.

[0017] In this technical solution, based on historical traffic data and a real-time prediction model, the system anticipates exit traffic flow trends in advance. When the predicted traffic flow increases, the allowed passage time is dynamically extended to accelerate vehicle evacuation; when the traffic flow decreases, the allowed passage time is shortened to avoid unnecessary waiting. This mechanism improves exit passage efficiency and significantly reduces the risk of queue overflow. When a high risk of rear-end collision is detected (such as insufficient braking distance of the following vehicle), the waiting time is extended to provide drivers with an additional reaction window, reducing the probability of rear-end collisions due to misjudgment. For behaviors such as frequent lane changes and rapid acceleration, the system shortens the allowed passage time, restricting their operating space and forcing them to decelerate or maintain their lane, while simultaneously triggering rear lane warning signals to reduce the possibility of conflicts. For characteristics such as low-frequency operations and long following distances, the allowed passage time is increased to provide sufficient passage windows, avoiding the risk of being delayed due to slow operation and reducing the anxiety of being urged by the following vehicle.

[0018] The present invention is further configured such that: S5 also includes: Store the information board control strategy, its corresponding triggering conditions, and execution instructions in the strategy library; Set emergency trigger conditions to open emergency escape routes, guide rescue vehicles to the scene, and issue emergency evacuation instructions when an emergency vehicle is detected to require rescue.

[0019] This technical solution structures and stores the information board control strategies (allowed passage duration, waiting warning, directional indication) and their triggering conditions (flow threshold, risk index, driving type) and execution instructions (traffic light sequence, lane permission) under different traffic conditions (flow changes, risk levels, driving styles) to form a reusable strategy knowledge base. The system monitors traffic conditions in real time and automatically matches the optimal control scheme through the strategy base, avoiding delays caused by manual intervention and improving the response speed of traffic lights.

[0020] One technical solution provided by this invention is a safety guidance system for emergency stopping lanes in tunnels, applied to a method for guiding safety in emergency stopping lanes in tunnels, comprising: The multi-source perception module deploys telephoto cameras, snapshot cameras, radar, and rangefinders in the emergency parking lane area, as well as vehicle-road cooperative terminals and on-board OBU devices on vehicles, forming a data acquisition network. It collects real-time data on vehicle position coordinates, velocity vectors, longitudinal acceleration, and heading angles in the parking lane area, obtains license plate recognition information through the snapshot cameras, and collects tire pressure and engine operating conditions through the on-board OBU devices. The edge computing module, configured with a smart workstation, is used for preprocessing the collected data, predicting traffic flow, assessing risks, and detecting conflicts. The control execution module dynamically adjusts the information board control strategy and displays parking lane information based on the output of the edge computing module. The emergency response module is used to respond to emergencies and coordinate rescue resources.

[0021] In this technical solution, the system constructs a three-dimensional perception network covering the entire scene of the tunnel parking lane. Through the collaborative work of radar, rangefinders, telephoto cameras, snapshot cameras, vehicle-road cooperative terminals, and on-board units (OBUs), it captures real-time motion status information such as vehicle position, speed, acceleration, and heading angle. It also integrates license plate recognition and vehicle operating condition data, eliminating the limitations of single sensors and significantly improving data integrity and real-time performance in complex tunnel environments. The edge computing module, based on a locally deployed smart workstation, performs real-time preprocessing, traffic flow prediction, risk assessment, and conflict detection on multi-source data, achieving low-latency response across the entire link from data acquisition to decision output. The system can dynamically identify potential risks such as rear-end collisions, side collisions, and lane change conflicts, and generate targeted information board control strategies, avoiding response delays caused by cloud transmission latency. This allows for proactive intervention in traffic flow before risks occur, reducing the accident rate. The control execution module adjusts the timing of traffic lights and the content displayed in the parking lane in real-time based on the output of the edge computing module, ensuring efficient traffic flow management and precise control of risk areas. The system supports dynamic optimization of traffic light timing based on traffic flow changes and risk levels. It also pushes risk warnings and evacuation instructions to drivers via variable message signs, improving the efficiency of human-vehicle coordination. The emergency response module automatically identifies emergencies such as fires and accidents through multi-source signal fusion analysis and triggers a tiered response mechanism. The system can quickly open escape routes, guide rescue vehicles directly to the scene, issue evacuation orders, and coordinate external resources such as fire and medical services, significantly shortening emergency response time and reducing the risk of secondary accidents.

[0022] The present invention is further configured such that the control execution module includes: The information board control terminal dynamically adjusts the permitted passage phase time based on traffic flow forecast results, and increases the waiting frequency in high-risk areas; Yellow flashing guidance signs are placed at a certain distance in front of the parking lane entrance. When an emergency vehicle is detected, the flashing warning mode is activated. The parking area features an information screen that displays the number of remaining parking spaces, estimated waiting time, and safety reminders in real time. For vehicles that exceed the waiting time limit, the screen displays the license plate number and a warning sign.

[0023] In this technical solution, the information board control terminal adjusts the permitted passage phase time in real time based on traffic flow prediction results (such as sudden changes in traffic flow or increased conflict probability). This adjustment includes extending the permitted passage time for high-demand lanes and shortening inefficient passage periods, ensuring a dynamic balance between exit efficiency and risk control at the parking lane. For areas prone to rear-end collisions or lane-changing conflicts, the system increases the frequency of waiting, enhancing visual warnings to drivers and forcing them to slow down or maintain a safe distance, reducing risky behaviors caused by misjudged signals. The system supports automatic switching of traffic light modes based on risk levels, preventing the failure of fixed timing strategies in complex scenarios. High-brightness LED guidance signs deployed before the parking lane entrance provide visual guidance to ordinary vehicles through multi-mode strobe control and trigger mandatory yield signals for emergency vehicles. When an ambulance, fire truck, or other emergency vehicle is detected approaching, the system immediately activates the guidance sign's strobe warning and, in conjunction with the information board control terminal, locks the conflict lane as a no-entry zone, creating an unobstructed passage path. Simultaneously, it pushes yield instructions to other vehicles through the vehicle-to-infrastructure (V2I) terminal, achieving a second-level response for emergency vehicles. Through real-time information transparency and dynamic guidance, drivers can anticipate the status of parking lanes in advance, reducing blind lane changes or sudden stops due to lack of information and lowering the risk of human error. Intelligent management of vehicles exceeding their permitted parking time and precise push notifications of parking space information improve the utilization rate of parking lane resources and alleviate pressure during peak hours. Priority passage for emergency vehicles and guidance signs shorten rescue response time and significantly improve tunnel emergency response capabilities.

[0024] The present invention is further configured such that the emergency response module includes: The warning unit triggers a yellow warning when a non-emergency vehicle stays for an extended period, and a red warning when an emergency vehicle approaches. The yellow warning is indicated by a warning message displayed on the information screen, while the red warning is indicated by the activation of an audible and visual alarm device and the display of an emergency rescue channel on the information screen. The vehicle control unit implements whitelist management for operating vehicles. Overtime records are automatically uploaded to the traffic management platform. For regular vehicles, if they do not leave after a three-minute timer, the camera will capture the license plate and preserve the chain of evidence. The rescue coordination unit includes a passive rescue mode and an active rescue mode. The passive rescue mode is automatically triggered when an accident is detected. The passive rescue mode includes activating a yellow light strip to guide the rescue route. The active rescue mode includes sending evacuation instructions to nearby vehicles, simultaneously reporting to the traffic police command center and recording routine handling logs, and coordinating rescue resources to arrive at the scene.

[0025] In this technical solution, the system automatically triggers yellow (non-emergency vehicle overstaying) or red (emergency vehicle approaching / accident occurring) warnings based on the risk type, enabling differentiated handling of risk levels. Yellow warnings display warning signs on the information screen, guiding drivers to proactively correct their behavior; red warnings activate audible and visual alarms and emergency lane guidance on the information screen, creating a dual visual and auditory warning to avoid secondary accident risks. A whitelist mechanism is implemented for commercial vehicles such as buses and hazardous materials transport vehicles, with overstaying records automatically synchronized to the traffic management platform, facilitating industry supervision and credit assessment, and improving the efficiency of key vehicle management. For non-commercial vehicles, the system sets a 3-minute overstay timer. After the time limit is exceeded, it automatically captures the license plate and generates a complete evidence chain including timestamps, location coordinates, and video clips of the violation, supporting post-event tracing and penalties, reducing manual evidence collection costs. The remaining stay time is dynamically displayed on the information screen ("2 minutes remaining"), combined with voice prompts ("Please leave as soon as possible"), guiding drivers to proactively avoid violations and reducing conflict risks. When an accident occurs, the system automatically activates the yellow light strip inside the tunnel, forming a continuous light strip along the optimal rescue route to guide rescue vehicles to the scene quickly. Simultaneously, it shuts off traffic lights in the conflict area to prevent other vehicles from entering. The system sends evacuation instructions ("Accident ahead, please change lanes immediately") to vehicles near the accident site, simultaneously reporting to the traffic police command center and recording the response log, coordinating external resources such as fire and medical personnel. Furthermore, it pushes a 3D map of the tunnel and real-time traffic conditions to rescue vehicles via vehicle-to-infrastructure (V2I) terminals, shortening response time.

[0026] The beneficial effects of this invention are: (1) generating information board control strategies to ensure that vehicles can safely and efficiently leave the emergency stopping lane; (2) through the fusion of multi-source data from radar, rangefinder, telephoto camera, snapshot camera, vehicle-road cooperative terminal and on-board OBU device, the system can capture vehicle dynamics, environmental status and vehicle health parameters in real time, and combined with historical traffic data, realize dynamic prediction of average vehicle speed, traffic flow, vehicle density and potential conflict points, providing data support for risk prediction; based on driving habit model and real-time data, the system can identify rear-end collisions, side collisions and lane change conflicts under different driving styles. The system enables personalized and precise risk assessment, providing a basis for differentiated control strategies. Based on prediction results and risk levels, the system generates information board control strategies, dynamically adjusts traffic light status, optimizes the passage efficiency of emergency parking lane exits, and reduces traffic congestion and accident risks caused by unreasonable traffic light settings. By building a strategy library and dynamically updating control strategies, the system can adapt to the needs of different traffic scenarios. At the same time, emergency trigger conditions are set to ensure rapid response in emergency situations, coordinate rescue resources, and improve emergency response efficiency and safety, thereby ensuring that vehicles can safely and efficiently leave the emergency parking lane. Attached Figure Description

[0027] Figure 1This is a flowchart illustrating the safe guidance method for emergency stopping lanes in tunnels according to the present invention. Figure 2 This is a schematic diagram of the information board strategy of the present invention; Figure 3 This is a schematic diagram of the parking lane warning system of the present invention. Figure 1 ; Figure 4 This is a schematic diagram of the parking lane warning system of the present invention. Figure 2 ; Figure 5 This is a flowchart of the parking lane warning system of the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only one preferred embodiment of this invention and are only used to explain this invention. They do not limit the scope of protection of this invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0029] like Figures 1 to 5 As shown in the figure, as an embodiment of the present invention, a method and system for safe guidance of emergency stopping lanes in tunnels includes the following steps: S1, the emergency stopping lane area is equipped with radar, rangefinder, telephoto camera, snapshot camera and smart workstation, the vehicle is equipped with vehicle-road cooperative terminal and on-board unit (OBU) device, the smart workstation supports unified vehicle-road cooperative communication protocol so that the multi-dimensional data acquired in real time by radar, rangefinder, telephoto camera, snapshot camera, vehicle-road cooperative terminal and on-board unit (OBU) device are all sent to the smart workstation; S2 uses machine learning algorithms to dynamically predict the average speed, traffic flow, and vehicle density of emergency stopping lanes and adjacent lanes based on real-time multi-dimensional data and historical traffic data. S3 identifies rear-end collision risks, side collision risks, and lane change conflict risks based on preset driving habit models, real-time multi-dimensional data, and historical traffic data. S4. Based on the dynamic prediction results and risk identification results, generate the intelligence board control strategy and dynamically adjust the intelligence board indication status. S5 builds a strategy library, dynamically updates strategies based on real-time data, and sets emergency trigger conditions.

[0030] In this embodiment, by fusing multi-source data from radar, rangefinders, telephoto cameras, snapshot cameras, vehicle-road cooperative terminals, and on-board unit (OBU) devices, the system can capture vehicle dynamics, environmental conditions, and vehicle health parameters in real time. Combined with historical traffic data, it enables dynamic prediction of average vehicle speed, traffic flow, vehicle density, and potential conflict points, providing data support for risk assessment. Based on driving habit models and real-time data, the system can identify rear-end collision, side-impact, and lane-change conflict risks under different driving styles, achieving personalized and precise risk assessment and providing a basis for differentiated control strategies. Based on the prediction results and risk levels, the system generates information board control strategies, dynamically adjusts traffic light states, optimizes the passage efficiency of emergency lane exits, and reduces traffic congestion and accident risks caused by unreasonable traffic light settings. By constructing a strategy library and dynamically updating control strategies, the system can adapt to different traffic scenario needs. Simultaneously, emergency trigger conditions are set to ensure rapid response in emergency situations, coordinate rescue resources, and improve emergency response efficiency and safety. This ensures that vehicles can safely and efficiently leave the emergency lane. The unified vehicle-road cooperative communication protocol is C-V2X (Cellular Vehicle-to-Everything), which is based on cellular networks and can achieve stable and efficient data transmission between tunnel sensors and vehicle-mounted sensors.

[0031] In one embodiment of the present invention, the multi-dimensional data includes vehicle dynamic parameters, vehicle spacing parameters, environmental parameters, and vehicle self-parameters. The vehicle dynamic parameters include the vehicle's position, speed, acceleration, and direction within the parking area; The vehicle spacing parameter refers to the distance to vehicles in the lane adjacent to the parking lane, including both near and far distances. The environmental parameters include visibility inside the tunnel and road surface slipperiness; The vehicle's own status parameters are the vehicle's own status information within the parking area, including tire condition and engine condition.

[0032] This technical solution integrates vehicle dynamic parameters, vehicle spacing parameters, environmental parameters, and vehicle status parameters to construct a comprehensive perception system covering vehicle operation, spatial relationships, environmental conditions, and vehicle health, providing full-element data support for risk identification and decision-making. The collaborative analysis of vehicle dynamic parameters and spacing parameters can accurately capture lane-changing intentions and conflict risks. The correlation modeling of environmental parameters and vehicle status parameters can provide early warnings of loss of control risks caused by slippery road surfaces or tire malfunctions, achieving an upgrade from single-risk identification to composite risk assessment. Multi-dimensional data provides dynamic input for information board control strategies: adjusting traffic light phases based on vehicle position and spacing to avoid conflicts between queuing vehicles at exits and vehicles approaching from adjacent lanes; optimizing waiting frequency or extending warning duration based on visibility and slipperiness to enhance safety guidance in adverse environments; and prioritizing right-of-way or triggering emergency rescue lanes for vehicles with abnormal tire pressure or engine malfunctions. By continuously monitoring the vehicle's own status parameters, the system can identify potential mechanical failures in advance, such as low tire pressure or engine overheating. Before the vehicle loses control, it can guide it to a safe stop via information board signals or link with the emergency response module to initiate a rescue process, turning passive response into active defense.

[0033] In step S2, the machine learning algorithm includes time series analysis, regression analysis, and spatial interpolation algorithms. The dynamic prediction of average vehicle speed, traffic flow, vehicle density, and potential conflict points in the emergency stopping lane and adjacent lanes using the machine learning algorithm includes: Based on historical vehicle speed data, current vehicle speed data, and adjacent lane vehicle speed data, the average vehicle speed over a certain period of time in the future is predicted using a time series analysis algorithm. Based on historical traffic data, current vehicle queue length, and entrance ramp traffic flow, a regression analysis algorithm is used to predict traffic flow changes over a certain period of time in the future. Based on vehicle location data and tunnel geometric parameters, a spatial interpolation algorithm is used to calculate the current and predicted vehicle density distribution over a certain future period.

[0034] This technical solution captures the periodic fluctuations in vehicle speed based on historical and real-time vehicle speed data, such as the decreasing trend of vehicle speed during weekday morning and evening rush hours. Combined with the correlation analysis of vehicle speeds in adjacent lanes, it predicts average speed changes 10-15 minutes in advance, providing a forward-looking basis for traffic light timing adjustments. By integrating historical traffic flow, queue length, and entrance ramp traffic flow data, it can predict traffic flow trends over the next 30 minutes, such as the risk of queue overflow due to a surge in ramp merging traffic, enabling proactive intervention in traffic fluctuations. Through spatial interpolation algorithms, it fuses vehicle positioning data with tunnel geometric parameters, such as lane width and radius of curvature, to generate a real-time vehicle density distribution heatmap and predict density diffusion paths over the next 5 minutes. This accurately identifies high-density areas, such as blind spots on curves and ramp merging points, supporting the spatially differentiated configuration of traffic light control strategies. The conflict detection algorithm combines vehicle trajectory data, such as longitudinal speed difference and lateral offset, with lane-changing intent recognition, such as turn signal signals and steering wheel angle, to identify potential conflict points 1.2-1.8 seconds in advance, such as the spatiotemporal intersection area between a lane-changing vehicle and an oncoming vehicle in an adjacent lane. By dynamically adjusting the traffic light phase or triggering a flashing yellow warning, conflicts are resolved at an early stage, reducing the rate of secondary accidents.

[0035] Understandably, time series analysis algorithms, regression analysis algorithms, and spatial interpolation algorithms form a closed-loop verification mechanism: time series and regression analysis provide trend predictions, while spatial interpolation locates risk areas.

[0036] Understandably, the time series analysis algorithm constructs a dynamic prediction model based on the periodic patterns of historical vehicle speed data, current vehicle speed, and vehicle speeds in adjacent lanes, including the following steps: Extract historical vehicle speeds (e.g., daily / weekly same time period), current vehicle speeds, adjacent lane vehicle speeds, and weather / visibility features, and remove abnormal vehicle speed data caused by equipment failure or sudden accidents; Using the Prophet algorithm, input vehicle speed data for the most recent 30 minutes and historical data for the same period over the past 7 days, and output the predicted average vehicle speed and confidence interval for the next 10 minutes; The model is retrained every 5 minutes to adapt to real-time traffic flow changes.

[0037] Understandably, regression analysis algorithms combine historical traffic flow, queue length, and entrance ramp traffic flow to predict future traffic flow trends, including the following steps: Input features: historical traffic flow (average over the past hour), current queue length (obtained via laser rangefinder), and entrance ramp traffic flow (detected via geomagnetic sensor or video). Using the XGBoost regression model, input the traffic data of the most recent 15 minutes and the historical traffic patterns of the same period, and output the traffic change trend within the next 5 minutes; Set a traffic threshold and trigger an alert when the predicted value exceeds 80% of the traffic threshold.

[0038] Understandably, spatial interpolation algorithms dynamically calculate vehicle density distribution based on vehicle location data and tunnel geometric parameters, including the following steps: The data input includes vehicle coordinates (x, y), timestamp, lane width, parking strip length, and curve radius provided by a millimeter-wave radar / laser rangefinder. A vehicle density heatmap was generated using the inverse distance weighting method. Density is categorized into low density (<10 vehicles / km), medium density (10-30 vehicles / km), and high density (>30 vehicles / km), corresponding to different risk levels; The density distribution is mapped onto the tunnel's 3D model and displayed in real time on the monitoring center's large screen.

[0039] In one embodiment of the present invention, step S3, establishing the preset driving habit model, includes: Based on drivers' historical travel data, frequency and magnitude of accelerator / brake pedal operation, lane change frequency, and following distance, drivers are divided into aggressive and conservative types. The aggressive type includes drivers with high accelerator pedal operation frequency, large brake pedal operation magnitude, high lane change frequency, and short average following distance. The conservative type includes drivers with low accelerator pedal operation frequency, small brake pedal operation magnitude, low lane change frequency, and long average following distance.

[0040] This technical solution constructs a driver behavior feature database based on multi-dimensional data, including historical trip data (such as trip duration and route selection), pedal operation characteristics (accelerator / brake frequency and amplitude), lane change behavior (frequency and timing), and following distance. The system can quantitatively analyze two driving styles: aggressive (high-frequency accelerator / sudden braking, high-frequency lane changes, short following distance) and conservative (low-frequency operation, long following distance), achieving standardized representation of driving behavior. By combining the driving habit model with real-time data (such as current following distance and lane change intention), the system can dynamically identify the match between the driver's style and current behavior. For example, if a conservative driver suddenly makes high-frequency lane changes, the system will trigger a higher-risk warning; if an aggressive driver continuously accelerates rapidly near the exit of the parking lane, the permitted passage time will be extended in advance to avoid misjudgment or missed judgment of risks caused by sudden changes in driving style. Tiered control is implemented for different driving styles: For aggressive drivers, the system limits frequent lane changes by shortening the permitted passage time, triggers flashing warnings for waiting before potential conflict points, and prioritizes guiding them to use exit lanes with lower conflict risk; for conservative drivers, the permitted waiting time is extended to avoid the risk of delays due to slow operation, and right-of-way is prioritized during low-traffic periods to reduce their waiting anxiety. Driving habit models can predict potential driver behavior in emergency stopping lane scenarios: aggressive drivers are more prone to lane-changing conflicts due to impatience, so the system intervenes by activating traffic lights before they approach the exit; conservative drivers may cause exit congestion due to hesitation, so the system optimizes their passage efficiency through dynamic signal timing. Through deep coupling of human risk and signal control, accident prevention is transformed from passive response to proactive intervention.

[0041] The identification of rear-end collision risk, side collision risk, and lane change conflict risk includes: Identify rear-end collision risk, side collision risk, and lane change conflict risk; calculate the rear-end collision risk index; when the risk index exceeds the first preset threshold, it is determined that there is a rear-end collision risk. The side collision risk index is calculated based on the vehicle's speed, direction, distance from vehicles in adjacent lanes, and the speed and direction of vehicles in adjacent lanes. When the risk index exceeds the second preset threshold, it is determined that there is a risk of side collision. Based on the vehicle's lane change intention, lane change speed, distance to vehicles in front and behind in the target lane, and the speed and acceleration of vehicles in front and behind in the target lane, a lane change conflict risk index is calculated. When the risk index exceeds the third preset threshold, it is determined that there is a risk of lane change conflict.

[0042] In this technical solution, the system constructs independent assessment models for rear-end collision risk, side collision risk, and lane change conflict risk. It calculates risk indices using multi-dimensional parameters (speed, distance, acceleration, direction, etc.) and sets differentiated thresholds (first / second / third preset thresholds). For example, rear-end collision risk focuses on the relative speed difference and braking distance between the vehicle and the vehicle in front; side collision risk considers the lateral speed component and lane spacing; and lane change conflict risk integrates lane change intent and target lane dynamics. This scenario-based risk modeling avoids misjudgment based on a single indicator and improves the accuracy of risk identification.

[0043] Understandably, the preset thresholds can be dynamically adjusted based on real-time traffic conditions: during peak hours or in severe weather, the system automatically lowers the risk threshold (e.g., reducing the side-collision risk threshold from 0.7 to 0.5) to trigger early warnings in response to high-density traffic flow; during low-traffic periods, the thresholds can be appropriately relaxed to improve traffic efficiency. This flexible threshold mechanism balances safety and efficiency needs, avoiding excessive intervention or delayed response.

[0044] Understandably, when the risk index exceeds the threshold, the system immediately generates targeted traffic light control instructions: For example, if there is a risk of rear-end collision, the system triggers a yellow flashing warning light in the rear lane and simultaneously extends the permitted passage time in this lane to accelerate vehicle evacuation; if there is a risk of side collision, the system prohibits oncoming traffic in the adjacent lane and guides the vehicle to prioritize using the lane with lower conflict risk; if there is a risk of lane change conflict, the system prohibits lane change signals for the vehicle and triggers a deceleration signal for vehicles behind in the target lane.

[0045] The formula for calculating the rear-end collision risk index is as follows: ; If RRI > the first preset threshold, it is determined that there is a risk of rear-end collision; RRI stands for Rear-end Collision Risk Index. The instantaneous speed of the vehicle behind; The instantaneous speed of the vehicle in front and behind; The longitudinal distance between the two vehicles; To slow down the vehicle behind; This is the minimum safe distance threshold; , , The weighting coefficient is determined from historical accident data.

[0046] The formula for calculating the side-impact risk index is as follows: ; If SRI > second preset threshold, it is determined that there is a risk of side collision; SRI stands for Side Collision Risk Index; This is the lateral velocity component of the vehicle; This refers to the lateral distance between this vehicle and vehicles in the adjacent lane. This is the longitudinal speed difference between this vehicle and vehicles in the adjacent lane; This is the longitudinal speed of the vehicle.

[0047] The formula for calculating the lane change conflict risk index is as follows: ; If LCRI > the third preset threshold, it is determined that there is a risk of lane change conflict; Among them, LCRI is the lane change conflict risk index; The available time for the gap between vehicles in front and behind the target lane; This is the time required for this vehicle to complete the lane change; For safety time margin; Acceleration of the vehicle following in the target lane; To provide maximum comfort deceleration for the following vehicle.

[0048] Preferably, the information board control strategy includes: When an increase in traffic flow is predicted at the emergency stopping lane exit, the permitted passage time is extended; when a decrease in traffic flow is predicted, the permitted passage time is shortened. In areas with a high risk of rear-end collisions, extend the waiting time; in areas with a high risk of lane change conflicts, install directional traffic lights. For aggressive drivers, shorten the permitted passage time; for conservative drivers, increase the permitted passage time.

[0049] In this technical solution, based on historical traffic data and real-time prediction models (such as ARIMA time series analysis), the system predicts exit traffic flow trends 10-15 minutes in advance. When the predicted traffic flow increases, the allowed passage time is dynamically extended (e.g., from 30 seconds to 45 seconds) to accelerate vehicle evacuation; when the traffic flow decreases, the allowed passage time is shortened (e.g., from 30 seconds to 20 seconds) to avoid unnecessary waiting. This mechanism improves exit traffic efficiency by 25%-35% and significantly reduces the risk of queue overflow. When a high risk of rear-end collision is detected (e.g., insufficient braking distance of following vehicles), the waiting time is extended (from 3 seconds to 5 seconds) to provide drivers with an additional reaction window and reduce the probability of rear-end collisions due to misjudgment. For behaviors such as frequent lane changes and rapid acceleration, the system shortens the allowed passage time (e.g., to 20 seconds) to restrict their operating space, forcing them to decelerate or maintain their lane, while simultaneously triggering rear lane warning signals to reduce the possibility of conflicts. In response to the characteristics of low-frequency operation and long following distance, the allowable passage time is increased (e.g., extended to 40 seconds) to provide sufficient passage window, avoid the risk of being stuck due to slow operation, and reduce the anxiety of being urged by the vehicle behind.

[0050] Understandably, when the traffic prediction error exceeds 10%, the system automatically triggers safety redundancy adjustments (such as allowing a buffer of ±5 seconds for passage time) to avoid exacerbating congestion due to prediction deviations.

[0051] Understandably, by setting directional traffic lights (such as arrow lights to guide lane changes), vehicles are forced to follow a preset path, thus preventing potential conflict routes. For example, in areas with high rates of lane change conflicts, only vehicles behind the target lane are allowed to slow down before proceeding, while vehicles in the target lane are prohibited from changing lanes, reducing the conflict rate by more than 60%.

[0052] In one embodiment of the present invention, S5 further includes: Store the information board control strategy, its corresponding triggering conditions, and execution instructions in the strategy library; Set emergency trigger conditions to open emergency escape routes, guide rescue vehicles to the scene, and issue emergency evacuation instructions when an emergency vehicle is detected to require rescue.

[0053] This technical solution structures and stores the information board control strategies (permissible passage duration, waiting time warning, directional indication) and their triggering conditions (flow threshold, risk index, driving type) and execution instructions (traffic light sequence, lane permission) under different traffic conditions (flow changes, risk levels, driving styles) to form a reusable strategy knowledge base. The system monitors traffic conditions in real time and automatically matches the optimal control scheme through the strategy base, avoiding delays caused by manual intervention and improving the response speed of traffic lights. Understandably, for scenarios of "aggressive drivers + high rear-end collision risk," a combined strategy of "shortening the permitted passage to 20 seconds + extending the waiting time to 5 seconds" is stored, supporting rapid retrieval.

[0054] Understandably, by fusing multiple sensor sources (such as radar, cameras, and emergency buttons), the location, speed, and rescue needs of emergency vehicles (ambulances, fire trucks) can be detected in real time, triggering a three-level response mechanism. Level 1 Response: Open emergency escape routes (e.g., open emergency doors, activate escape indicator lights), and simultaneously turn off conflict lane traffic lights to create an unobstructed passageway; Level 2 Response: Using dynamic path planning algorithms, rescue vehicles are guided to the scene via the shortest route (such as a dedicated lane for vehicles traveling in the wrong direction), and the traffic lights in the relevant lanes are simultaneously adjusted to be all green. Level 3 Response: Issue emergency evacuation instructions to vehicles inside the tunnel (such as voice broadcasts and LED screen prompts), and coordinate with exit traffic lights to give priority to evacuating traffic.

[0055] Understandably, during the passage of rescue vehicles, the system forcibly locks the traffic lights in the 100-meter area before and after the target lane to prohibit passage, in order to prevent other vehicles from accidentally entering and causing secondary accidents.

[0056] like Figures 3 to 5 As shown in Embodiment 2 of the present invention, a tunnel emergency stopping lane safety guidance system includes: The multi-source perception module deploys telephoto cameras, snapshot cameras, radar, and rangefinders in the emergency parking lane area, as well as vehicle-road cooperative terminals and on-board OBU devices on vehicles, forming a data acquisition network. It collects real-time data on vehicle position coordinates, velocity vectors, longitudinal acceleration, and heading angles in the parking lane area, obtains license plate recognition information through the snapshot cameras, and collects tire pressure and engine operating conditions through the on-board OBU devices. The edge computing module, configured with a smart workstation, is used for preprocessing the collected data, predicting traffic flow, assessing risks, and detecting conflicts. The control execution module dynamically adjusts the information board control strategy and displays parking lane information based on the output of the edge computing module. The emergency response module is used to respond to emergencies and coordinate rescue resources.

[0057] In this embodiment, the system constructs a three-dimensional perception network covering the entire tunnel parking area. Through the collaborative work of radar, rangefinders, telephoto cameras, snapshot cameras, vehicle-to-infrastructure (V2I) terminals, and on-board unit (OBU) devices, it captures real-time motion status information such as vehicle position, speed, acceleration, and heading angle. It also integrates license plate recognition and vehicle operating condition data, eliminating the limitations of single sensors and significantly improving data integrity and real-time performance in complex tunnel environments. The edge computing module, based on a locally deployed smart workstation, performs real-time preprocessing, traffic flow prediction, risk assessment, and conflict detection on multi-source data, achieving low-latency response across the entire link from data acquisition to decision output. The system can dynamically identify potential risks such as rear-end collisions, side collisions, and lane change conflicts, and generate targeted information board control strategies. This avoids response delays caused by cloud transmission latency, thus proactively intervening in traffic flow before risks occur and reducing the accident rate. The control execution module adjusts the timing of traffic lights and the content displayed in the parking area in real-time based on the output of the edge computing module, ensuring efficient traffic flow management and precise control of risk areas. The system supports dynamic optimization of traffic light timing based on traffic flow changes and risk levels. It also pushes risk warnings and evacuation instructions to drivers via variable message signs, improving the efficiency of human-vehicle coordination. The emergency response module automatically identifies emergencies such as fires and accidents through multi-source signal fusion analysis and triggers a tiered response mechanism. The system can quickly open escape routes, guide rescue vehicles directly to the scene, issue evacuation orders, and coordinate external resources such as fire and medical services, significantly shortening emergency response time and reducing the risk of secondary accidents.

[0058] Preferably, the control execution module includes: The information board control terminal dynamically adjusts the permitted passage phase time based on traffic flow forecast results, and increases the waiting frequency in high-risk areas; Yellow flashing guidance signs are placed at a certain distance in front of the parking lane entrance. When an emergency vehicle is detected, the flashing warning mode is activated. The parking area features an information screen that displays the number of remaining parking spaces, estimated waiting time, and safety reminders in real time. For vehicles that exceed the waiting time limit, the screen displays the license plate number and a warning sign.

[0059] In this technical solution, the information board control terminal adjusts the permitted passage phase time in real time based on traffic flow prediction results (such as sudden changes in traffic flow or increased conflict probability). This adjustment includes extending the permitted passage time for high-demand lanes and shortening inefficient passage periods, ensuring a dynamic balance between exit efficiency and risk control at the parking lane. For areas prone to rear-end collisions or lane-changing conflicts, the system increases the frequency of waiting, enhancing visual warnings to drivers and forcing them to slow down or maintain a safe distance, reducing risky behaviors caused by misjudged signals. The system supports automatic switching of traffic light modes based on risk levels, preventing the failure of fixed timing strategies in complex scenarios. High-brightness LED guidance signs deployed before the parking lane entrance provide visual guidance to ordinary vehicles through multi-mode strobe control and trigger mandatory yield signals for emergency vehicles. When an ambulance, fire truck, or other emergency vehicle is detected approaching, the system immediately activates the guidance sign's strobe warning and, in conjunction with the information board control terminal, locks the conflict lane as a no-entry zone, creating an unobstructed passage path. Simultaneously, it pushes yield instructions to other vehicles through the vehicle-to-infrastructure (V2I) terminal, achieving a second-level response for emergency vehicles. Through real-time information transparency and dynamic guidance, drivers can anticipate the status of parking lanes in advance, reducing blind lane changes or sudden stops due to lack of information and lowering the risk of human error. Intelligent management of vehicles exceeding their permitted parking time and precise push notifications of parking space information improve the utilization rate of parking lane resources and alleviate pressure during peak hours. Priority passage for emergency vehicles and guidance signs shorten rescue response time and significantly improve tunnel emergency response capabilities.

[0060] Understandably, the guidance signs use dual-color temperature LEDs (yellow light + red light) and anti-glare lenses to ensure clear identification even in strong light or smoke in tunnels, thus avoiding misjudgment.

[0061] The emergency response module includes: The warning unit triggers a yellow warning when a non-emergency vehicle stays for an extended period, and a red warning when an emergency vehicle approaches. The yellow warning is indicated by a warning message displayed on the information screen, while the red warning is indicated by the activation of an audible and visual alarm device and the display of an emergency rescue channel on the information screen. The vehicle control unit implements whitelist management for operating vehicles. Overtime records are automatically uploaded to the traffic management platform. For regular vehicles, if they do not leave after a three-minute timer, the camera will capture the license plate and preserve the chain of evidence. The rescue coordination unit includes a passive rescue mode and an active rescue mode. The passive rescue mode is automatically triggered when an accident is detected. The passive rescue mode includes activating a yellow light strip to guide the rescue route. The active rescue mode includes sending evacuation instructions to nearby vehicles, simultaneously reporting to the traffic police command center and recording routine handling logs, and coordinating rescue resources to arrive at the scene.

[0062] In this technical solution, the system automatically triggers yellow (non-emergency vehicle overstaying) or red (emergency vehicle approaching / accident occurring) warnings based on the risk type, enabling differentiated handling of risk levels. Yellow warnings display warning signs on the information screen, guiding drivers to proactively correct their behavior; red warnings activate audible and visual alarms and emergency lane guidance on the information screen, creating a dual visual and auditory warning to avoid secondary accident risks. A whitelist mechanism is implemented for commercial vehicles such as buses and hazardous materials transport vehicles, with overstaying records automatically synchronized to the traffic management platform, facilitating industry supervision and credit assessment, and improving the efficiency of key vehicle management. For non-commercial vehicles, the system sets a 3-minute overstay timer. After the time limit is exceeded, it automatically captures the license plate and generates a complete evidence chain including timestamps, location coordinates, and video clips of the violation, supporting post-event tracing and penalties, reducing manual evidence collection costs. The remaining stay time is dynamically displayed on the information screen ("2 minutes remaining"), combined with voice prompts ("Please leave as soon as possible"), guiding drivers to proactively avoid violations and reducing conflict risks. When an accident occurs, the system automatically activates the yellow light strip inside the tunnel, forming a continuous light strip along the optimal rescue route to guide rescue vehicles to the scene quickly. Simultaneously, it shuts off traffic lights in the conflict area to prevent other vehicles from entering. The system sends evacuation instructions ("Accident ahead, please change lanes immediately") to vehicles near the accident site, simultaneously reporting to the traffic police command center and recording the response log, coordinating external resources such as fire and medical personnel. Furthermore, it pushes a 3D map of the tunnel and real-time traffic conditions to rescue vehicles via vehicle-to-infrastructure (V2I) terminals, shortening response time.

[0063] Understandably, the early warning unit can distinguish between routine violations (excessive stay) and emergency events (accidents, rescue needs), avoiding excessive interference with normal traffic flow by a single early warning mode, while ensuring that key risks are not overlooked.

[0064] The above embodiments, which describe the specific features of the present invention, are only used to further illustrate the present invention and should not be construed as limiting the scope of protection of the present invention. Any non-essential improvements and adjustments made to the present invention by those skilled in the art based on the above description of the invention shall fall within the scope of protection of the present invention.

Claims

1. A method for safely guiding emergency stopping lanes in tunnels, characterized in that, Includes the following steps: S1, the emergency stopping lane area is equipped with radar, rangefinder, telephoto camera, snapshot camera and smart workstation, and the vehicle is equipped with vehicle-road cooperative terminal and on-board OBU device. The smart workstation supports a unified vehicle-road cooperative communication protocol so that the multi-dimensional data acquired in real time by radar, rangefinder, telephoto camera, snapshot camera, vehicle-road cooperative terminal and on-board OBU device are all sent to the smart workstation. S2 uses machine learning algorithms to dynamically predict the average speed, traffic flow, and vehicle density of emergency stopping lanes and adjacent lanes based on real-time multi-dimensional data and historical traffic data. S3 identifies rear-end collision risks, side collision risks, and lane change conflict risks based on preset driving habit models, real-time multi-dimensional data, and historical traffic data. S4. Based on the dynamic prediction results and risk identification results, generate the intelligence board control strategy and dynamically adjust the intelligence board indication status. S5 builds a strategy library, dynamically updates strategies based on real-time data, and sets emergency trigger conditions.

2. The method for safe guidance of an emergency stopping lane in a tunnel according to claim 1, characterized in that, The multi-dimensional data includes vehicle dynamic parameters, vehicle spacing parameters, environmental parameters, and vehicle-specific parameters. The vehicle dynamic parameters include the vehicle's position, speed, acceleration, and direction within the parking area; The vehicle spacing parameter refers to the distance to vehicles in the lane adjacent to the parking lane, including both near and far distances. The environmental parameters include visibility inside the tunnel and road surface slipperiness; The vehicle's own status parameters are the vehicle's own status information within the parking area, including tire condition and engine condition.

3. A method for safe guidance of an emergency stopping lane in a tunnel according to claim 2, characterized in that, In step S2, the machine learning algorithm includes time series analysis, regression analysis, and spatial interpolation algorithms. The dynamic prediction of average vehicle speed, traffic flow, vehicle density, and potential conflict points in the emergency stopping lane and adjacent lanes using the machine learning algorithm includes: Based on historical vehicle speed data, current vehicle speed data, and adjacent lane vehicle speed data, the average vehicle speed over a certain period of time in the future is predicted using a time series analysis algorithm. Based on historical traffic data, current vehicle queue length, and entrance ramp traffic flow, a regression analysis algorithm is used to predict traffic flow changes over a certain period of time in the future. Based on vehicle location data and tunnel geometric parameters, a spatial interpolation algorithm is used to calculate the current and predicted vehicle density distribution over a certain future period.

4. A method for safe guidance of an emergency stopping lane in a tunnel according to claim 3, characterized in that, In step S3, the establishment of the preset driving habit model includes: Based on drivers' historical travel data, frequency and magnitude of accelerator / brake pedal operation, lane change frequency, and following distance, drivers are divided into aggressive and conservative types. The aggressive type includes drivers with high accelerator pedal operation frequency, large brake pedal operation magnitude, high lane change frequency, and short average following distance. The conservative type includes drivers with low accelerator pedal operation frequency, small brake pedal operation magnitude, low lane change frequency, and long average following distance.

5. A method for safe guidance of an emergency stopping lane in a tunnel according to claim 1, 2, 3, or 4, characterized in that, The identification of rear-end collision risk, side collision risk, and lane change conflict risk includes: Identify rear-end collision risk, side collision risk, and lane change conflict risk; calculate the rear-end collision risk index; when the risk index exceeds the first preset threshold, it is determined that there is a rear-end collision risk. The side collision risk index is calculated based on the vehicle's speed, direction, distance from vehicles in adjacent lanes, and the speed and direction of vehicles in adjacent lanes. When the risk index exceeds the second preset threshold, it is determined that there is a risk of side collision. Based on the vehicle's lane change intention, lane change speed, distance to vehicles in front and behind in the target lane, and the speed and acceleration of vehicles in front and behind in the target lane, a lane change conflict risk index is calculated. When the risk index exceeds the third preset threshold, it is determined that there is a risk of lane change conflict.

6. A method for safe guidance of an emergency stopping lane in a tunnel according to claim 4, characterized in that, The information board control strategy includes: When an increase in traffic flow is predicted at the emergency stopping lane exit, the permitted passage time is extended; when a decrease in traffic flow is predicted, the permitted passage time is shortened. In areas with a high risk of rear-end collisions, extend the waiting time; in areas with a high risk of lane change conflicts, install directional traffic lights. For aggressive drivers, shorten the permitted passage time; for conservative drivers, increase the permitted passage time.

7. A method for safe guidance of an emergency stopping lane in a tunnel according to claim 6, characterized in that, The S5 also includes: Store the information board control strategy, its corresponding triggering conditions, and execution instructions in the strategy library; Set emergency trigger conditions to open emergency escape routes, guide rescue vehicles to the scene, and issue emergency evacuation instructions when an emergency vehicle is detected to require rescue.

8. A tunnel emergency stopping lane safety guidance system, applied to the tunnel emergency stopping lane safety guidance method according to any one of claims 1-7, characterized in that, include: The multi-source perception module deploys telephoto cameras, snapshot cameras, radar, and rangefinders in the emergency parking lane area, as well as vehicle-road cooperative terminals and on-board OBU devices on vehicles, forming a data acquisition network. It collects vehicle position coordinates, velocity vectors, longitudinal acceleration, and heading angle data in real time within the parking lane area, obtains license plate recognition information through snapshot cameras, and collects tire pressure and engine operating conditions through on-board OBU devices. The edge computing module, configured with a smart workstation, is used for preprocessing the collected data, predicting traffic flow, assessing risks, and detecting conflicts. The control execution module dynamically adjusts the information board control strategy and displays parking lane information based on the output of the edge computing module. The emergency response module is used to respond to emergencies and coordinate rescue resources.

9. A tunnel emergency stopping lane safety guidance system according to claim 8, characterized in that, The control execution module includes: The information board control terminal dynamically adjusts the permitted passage phase time based on traffic flow forecast results, and increases the waiting frequency in high-risk areas; Yellow flashing guidance signs are placed at a certain distance in front of the parking lane entrance. When an emergency vehicle is detected, the flashing warning mode is activated. The parking area features an information screen that displays the number of remaining parking spaces, estimated waiting time, and safety reminders in real time. For vehicles that exceed the waiting time limit, the screen displays the license plate number and a warning sign.

10. A tunnel emergency stopping lane safety guidance system according to claim 9, characterized in that, The emergency response module includes: The warning unit triggers a yellow warning when a non-emergency vehicle stays for an extended period, and a red warning when an emergency vehicle approaches. The yellow warning is indicated by a warning message displayed on the information screen, while the red warning is indicated by the activation of an audible and visual alarm device and the display of an emergency rescue route on the information screen. The vehicle control unit implements whitelist management for operating vehicles. Overtime records are automatically uploaded to the traffic management platform. For regular vehicles, if they do not leave after a three-minute timer, the camera will capture the license plate and preserve the chain of evidence. The rescue coordination unit includes a passive rescue mode and an active rescue mode. The passive rescue mode is automatically triggered when an accident is detected. The passive rescue mode includes activating a yellow light strip to guide the rescue route. The active rescue mode includes sending evacuation instructions to nearby vehicles, simultaneously reporting to the traffic police command center and recording routine handling logs, and coordinating rescue resources to arrive at the scene.

Images