Dynamic traffic flow control method for elevated drop-off ramp of large railway hub

Abstract

The application discloses a large railway hub overhead off-ramp dynamic traffic flow control method, realizes dynamic self-adaptive adjustment of vehicle stay time threshold according to real-time traffic state (queue length, lane occupancy rate and driving-in flow) of the ramp, automatically compresses the allowed stay time to accelerate vehicle turnover in peak congestion period, and appropriately relaxes in low peak idle period to improve humanized experience; a multi-dimensional dynamic correction factor system is constructed, time period characteristics (peak / flat / low), environmental factors (adverse weather such as rain, snow and fog, holidays / large events), historical overtime statistical rules and the like are brought into a dynamic control calculation model, and the adaptability and robustness of the control strategy are improved; a self-learning mechanism based on historical overtime rate feedback is introduced, so that the algorithm can automatically adjust the laxity according to actual operation data, stabilize the overtime rate in a preset target range, and realize continuous optimization iteration of the control strategy.

Description

Technical Field

[0001] This invention relates to the field of intelligent traffic management and control technology, specifically to a dynamic traffic flow control method for elevated passenger drop-off ramps in large railway hubs. Background Technology

[0002] As key nodes connecting urban and rural transportation, large railway hubs have elevated drop-off ramps that serve as drop-off points for various modes of transportation, including private vehicles, ride-hailing services, and taxis. With the continuous increase in passenger flow at these hubs, problems such as traffic congestion, illegal parking, and low traffic efficiency at these drop-off ramps have become increasingly prominent, becoming a "bottleneck" restricting the overall operational efficiency of the hubs.

[0003] Currently, the management and control measures for passenger drop-off ramps in major domestic railway hubs have mainly gone through three stages:

[0004] Phase 1: Static time limits + manual patrols. Static signs indicating "Limited Parking for X Minutes" are placed at ramp entrances or above lanes, and on-site security personnel or traffic police manually advise and disperse vehicles. This method relies on manual experience, is extensive in management, and is difficult to cope with dynamically changing traffic demands.

[0005] Phase Two: Video-based illegal parking detection + fixed-time penalties. High-definition cameras are deployed along the ramps to automatically capture images and issue penalties to vehicles exceeding a preset fixed time (e.g., 3 or 5 minutes). This method achieves automated management, but the "one-size-fits-all" fixed time limit cannot differentiate the reasonable stopping needs of different vehicle types (private cars, ride-hailing vehicles, taxis) nor can it be flexibly adjusted according to real-time congestion levels.

[0006] Phase Three: Radar-Visual Fusion Sensing + Static Hierarchical Management. Some advanced railway hubs have begun to adopt millimeter-wave radar and video fusion sensing technology to achieve vehicle trajectory tracking and dwell time statistics, and set different static time limits for taxis, ride-hailing vehicles, and private cars (e.g., 2 minutes for private cars and 4 minutes for ride-hailing vehicles). Although vehicle type-based management has been achieved, the dwell time thresholds are still pre-set static values ​​and cannot dynamically change with factors such as real-time ramp load, time of day, and weather.

[0007] Disadvantages or shortcomings of existing technologies:

[0008] (1) Static dwell time threshold: Whether it is 3 minutes, 5 minutes or fixed values ​​set according to vehicle type, it cannot respond to the dynamic changes in real-time traffic status (such as queue length and occupancy rate) of drop-off ramps, resulting in low vehicle turnover efficiency and congestion during peak hours, while excessively restricting reasonable vehicle dwell time during off-peak hours, resulting in poor user experience.

[0009] (2) Lack of multi-factor comprehensive adjustment capability: The existing scheme only considers vehicle type or simple time period division, and fails to incorporate multi-dimensional factors such as weather conditions (rain, snow, fog, etc.), special events (holidays, large-scale events), and historical overtime patterns into the dynamic allocation model of stay time, resulting in insufficient adaptability and robustness of the control strategy.

[0010] (3) Lack of self-learning and closed-loop optimization mechanism: Once the dwell time threshold is set, the traditional solution remains fixed for a long time and cannot be adaptively adjusted according to actual operation data (such as overtime rate and average dwell time). It lacks closed-loop feedback from "data collection → status assessment → strategy optimization". Summary of the Invention

[0011] This invention provides a dynamic traffic flow control method for elevated passenger drop-off ramps in large railway hubs to solve the above-mentioned technical problems in the prior art.

[0012] According to a first aspect, one embodiment provides a method for dynamic traffic flow control of elevated drop-off ramps in a large railway hub, the method comprising:

[0013] Real-time collection of vehicle characteristic information, real-time traffic status parameters, current traffic time period type, and environmental parameters for vehicles entering the drop-off ramp;

[0014] Based on the collected vehicle characteristic information, real-time traffic status parameters, the traffic time period type and environmental parameters of the current time, a dynamic set of correction factors is calculated, including load correction factor, time period correction factor, environmental correction factor and historical pattern correction factor.

[0015] Based on the current vehicle type, the corresponding basic dwell time is obtained from the pre-created vehicle type-basic dwell time mapping table. The basic dwell time is corrected and calculated using a set of dynamic correction factors, and a safety limit is applied to obtain the maximum dynamic allowable dwell time for the current vehicle.

[0016] Based on the current maximum dynamic allowable dwell time of vehicles and the actual dwell time monitored, traffic flow on ramps is guided and managed, and graded early warning processing is carried out in conjunction with a graded early warning mechanism.

[0017] Furthermore, real-time data is collected on vehicle characteristic information entering the drop-off / pick-up ramps, real-time traffic status parameters, the current traffic time period type, and environmental parameters, specifically including:

[0018] By deploying vehicle sensing units at the entrance of the drop-off ramp, vehicles entering the drop-off ramp are detected and identified in real time, and vehicle feature information including vehicle identity information, vehicle type and entry timestamp is extracted. The vehicle types include private cars, ride-hailing vehicles, taxis and special vehicles.

[0019] Furthermore, real-time data is collected on vehicle characteristic information entering the drop-off / pick-up ramps, real-time traffic status parameters, the current traffic time period type, and environmental parameters, specifically including:

[0020] By deploying integrated radar-visual devices or distributed sensor networks throughout the drop-off ramps, real-time traffic status parameters, including lane occupancy rate, average queue length, inbound traffic flow, and outbound traffic flow, are collected.

[0021] Furthermore, real-time data is collected on vehicle characteristic information entering the drop-off / pick-up ramps, real-time traffic status parameters, the current traffic time period type, and environmental parameters, specifically including:

[0022] The traffic time period is determined based on the current time, which includes peak hours, off-peak hours, and low-peak hours. Environmental parameters, including weather conditions and special events, are also obtained.

[0023] Furthermore, by deploying integrated radar-visual systems or distributed sensor networks throughout the drop-off ramps, real-time traffic status parameters, including lane occupancy rate, average queue length, inbound traffic flow, and outbound traffic flow, are collected, specifically including:

[0024] Lane occupancy rate The ratio of the length of the lane currently occupied by vehicles to the total length of the drop-off ramp;

[0025] Average queue length The average length of the vehicle queue extending upstream from the ramp exit;

[0026] Inbound traffic The number of vehicles that entered the ramp from inside the window in the recent period of time;

[0027] leaving the flow The number of vehicles that exited the ramp from inside the window in the most recent time.

[0028] Furthermore, based on the collected vehicle characteristic information, real-time traffic state parameters, the traffic time period type of the current moment, and environmental parameters, a dynamic set of correction factors is calculated, including load correction factor, time period correction factor, environmental correction factor, and historical pattern correction factor. Specifically, this set includes:

[0029] a. Load correction factor calculate:

[0030]

[0031] in:

[0032] The current lane occupancy rate, The congestion warning threshold is set to 0.7;

[0033] This represents the current average queue length. This is the total length of the drop-off ramps;

[0034] , For the weighting coefficients, satisfying ;

[0035] After calculation Limiting the amplitude: ;

[0036] b. Time Period Correction Factor calculate:

[0037] Values ​​are determined based on time period type:

[0038]

[0039] c. Environmental Correction Factors calculate:

[0040]

[0041] in:

[0042] : 0 for sunny days, and 0.1 to 0.3 for rainy, snowy, or foggy days;

[0043] The value is 0 for ordinary days and 0.1 to 0.4 for holidays or major event days.

[0044] After calculation Limiting the amplitude: ;

[0045] d. Historical pattern correction factor calculate:

[0046]

[0047] in:

[0048] The historical overtime rate is calculated as follows: the proportion of the number of vehicles whose actual dwell time exceeds the corresponding dynamic threshold within the same time period to the total number of vehicles entering the corresponding time period.

[0049] The target timeout rate is defined as 0.05 to 0.20.

[0050] To adjust the sensitivity coefficient, the value range is 1.0 to 3.0;

[0051] After calculation Limiting the amplitude: .

[0052] Furthermore, the base dwell time is corrected using a set of dynamic correction factors, and a safety limit is applied to obtain the maximum dynamic permissible dwell time for the current vehicle, specifically including:

[0053]

[0054] in, To correct the dynamic allowable dwell time, Based on the basic stay time, As the load correction factor, For time period correction factor, As an environmental correction factor, As a factor for correcting historical patterns;

[0055] Then apply a safety limit:

[0056]

[0057] in:

[0058] The minimum allowable dwell time ranges from 30 to 60 seconds.

[0059] This represents the absolute maximum permissible dwell time, ranging from 300 to 600 seconds.

[0060] This represents the maximum dynamic allowable dwell time that is ultimately obtained.

[0061] Furthermore, based on the current maximum dynamic permissible dwell time of vehicles and the monitored actual dwell time, a tiered early warning system is implemented, specifically including:

[0062] From the moment the vehicle entered The timing begins, and the location of the corresponding vehicle is continuously tracked through the integrated radar-visual system, recording its actual dwell time. , The current moment;

[0063] when achieve When the number of vehicles reaches 70% to 90%, a Level 1 warning is triggered: the license plate number of the corresponding vehicle is displayed on the roadside LED guidance screen, and a message is sent through the parking space geomagnetic broadcast system: "Please leave as soon as possible."

[0064] when When this occurs, a Level 2 warning is triggered: the guidance screen displays "Excessive stay will be dealt with according to law," automatically records evidence of the violation, and sends a reminder to the driver through the management platform system;

[0065] when When a Level 3 warning is triggered, the evidence of the violation will be packaged and pushed to the traffic management department's enforcement platform, whereby law enforcement officers will decide whether to issue a non-on-site penalty or notify on-site personnel to handle the situation.

[0066] This represents the maximum dynamic allowable dwell time for the current vehicle.

[0067] Furthermore, the method also includes:

[0068] This will include the entry time of each vehicle, vehicle type, actual stay time, and calculated data. Data on whether or not the timeout has occurred is stored in the historical database;

[0069] Recalculate the overtime rate for each time period and vehicle type every hour or day. And update the historical pattern correction factor. Historical statistics;

[0070] The system can automatically adjust the target timeout rate based on long-term data trends. Alternatively, the weighting coefficients of each correction factor can be used to continuously optimize the control strategy.

[0071] According to a second aspect, one embodiment provides a traffic flow control system for elevated drop-off ramps in a large railway hub, used to execute a dynamic traffic flow control method for elevated drop-off ramps in a large railway hub as described in any of the preceding claims, the system comprising:

[0072] Entrance detection gantry: Located 20 meters upstream of the starting point of the ramp, it is equipped with a license plate recognition camera and millimeter-wave radar to detect and identify vehicles entering the drop-off ramp in real time;

[0073] Mid-section radar-visual integrated unit: Radar-visual integrated units integrating millimeter-wave radar and high-definition video are deployed in the middle of the ramp and at the exit, for the purpose of tracking vehicle trajectory and counting dwell time throughout the entire process;

[0074] Roadside guidance screens: LED guidance screens are deployed at the entrance, middle section, and exit of the ramp to display license plate numbers and prompts;

[0075] Parking space geomagnetic sensors: A geomagnetic system is installed in the drop-off area parking spaces to identify vehicles and control dynamic broadcasts.

[0076] Broadcast speaker columns: Multiple directional broadcast speaker columns will be deployed along the ramps to cover the entire road section;

[0077] Edge computing unit: Deployed in the field cabinet, running algorithm models;

[0078] Central Management Platform: Deployed in the hub management center computer room, it is connected to the edge computing unit via fiber optic or 5G private network to realize data aggregation, visualization display, parameter configuration and law enforcement linkage.

[0079] This invention provides a dynamic traffic flow control method for elevated passenger drop-off ramps in large railway hubs, which has the following beneficial effects:

[0080] 1. Significantly improved traffic efficiency: through load correction factors The introduction of this feature automatically shortens the allowed dwell time during ramp congestion, accelerating vehicle turnaround. Tests using VISSIM simulation software show that during peak hours (occupancy > 0.7), this solution can reduce average vehicle dwell time by 25%–35%, increase ramp capacity (vehicles / hour) by 15%–25%, and reduce maximum queue length by 30%–40%.

[0081] 2. Enhanced humanization and adaptability of management strategies: through time-period adjustment factors. and environmental correction factors During off-peak hours or in inclement weather, the permitted stay time is appropriately extended, balancing management rigidity with user experience and avoiding complaints and disputes caused by "one-size-fits-all" control.

[0082] 3. Self-learning and closed-loop optimization capabilities: Historical pattern correction factor This allows the system to automatically adjust its leniency based on actual operational data, stabilizing the timeout rate around a preset target (such as 10%), achieving an evolutionary capability that "gets smarter with use" without the need for frequent manual parameter adjustments.

[0083] 4. Tiered early warning and flexible enforcement: A three-tiered early warning mechanism (flexible reminder → warning prompt → enforcement linkage) is set up to give drivers a reasonable buffer time, reduce the confrontational emotions caused by direct penalties, and improve the standardization and persuasiveness of enforcement evidence.

[0084] 5. Reduce manual management costs: The upgrade from "manual patrol + static capture" to "fully automatic dynamic control" is expected to reduce on-site security personnel by 30% to 50%, while also reducing the traffic complaint rate caused by mismanagement.

[0085] 6. Reduce the vehicle congestion density per unit length of drop-off ramps; shorten the travel time of drop-off vehicles in the entire ramp area. Attached Figure Description

[0086] Figure 1 A flowchart illustrating a dynamic traffic flow control method for elevated passenger drop-off ramps in a large railway hub, as provided in one embodiment of the present invention;

[0087] Figure 2This is a schematic diagram of the equipment system layout for a dynamic traffic flow control method for elevated passenger drop-off ramps in a large railway hub, provided as an embodiment of the present invention. Detailed Implementation

[0088] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the invention. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present invention are not shown or described in the specification. This is to avoid obscuring the core parts of the invention with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0089] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0090] The first embodiment of this invention provides a dynamic traffic flow control method for elevated passenger drop-off ramps in large railway hubs, which will be discussed below in conjunction with... Figure 1 Please provide a detailed explanation.

[0091] like Figure 1 As shown, in step S100, the vehicle characteristic information, real-time traffic status parameters, traffic time period type, and environmental parameters of vehicles entering the drop-off ramp are collected in real time.

[0092] The above steps specifically include:

[0093] S110 uses a vehicle sensing unit deployed at the entrance of the drop-off ramp to detect and identify vehicles entering the drop-off ramp in real time, and extracts vehicle feature information including vehicle identity information, vehicle type and entry timestamp. The vehicle type includes private cars, ride-hailing vehicles, taxis and special vehicles.

[0094] In this embodiment, a vehicle sensing unit (including but not limited to a combination of license plate recognition cameras and millimeter-wave radar) deployed at the entrance of the drop-off ramp is used to detect and identify vehicles entering the drop-off ramp in real time, and extract the following vehicle feature information:

[0095] (1) Vehicle identification information: at least the license plate number, preferably the license plate color as well;

[0096] (2) Vehicle type: Through the local vehicle database or the vehicle information database of the traffic management department, the vehicle type is divided into at least one of the following: private car, ride-hailing car, taxi, special vehicle (ambulance, police car, fire truck, etc.), of which ride-hailing car can be obtained in real time through the data interface with the ride-hailing platform;

[0097] Entry timestamp: Records the moment the vehicle enters the ramp detection area. The preferred accuracy is at the millisecond level.

[0098] S120 collects real-time traffic status parameters, including lane occupancy rate, average queue length, inbound traffic flow, and outbound traffic flow, through integrated radar-visual units or distributed sensor networks deployed throughout the drop-off ramps.

[0099] In this embodiment, the following traffic state parameters are collected in real time by using integrated radar-visual devices or distributed sensor networks deployed throughout the drop-off ramp area, with a sampling frequency of 0.1 to 10 seconds / time:

[0100] (1) Lane occupancy rate The ratio of the length of the lane currently occupied by vehicles to the total length of the drop-off ramp, with a value ranging from 0 to 1, preferably calculated using point cloud data from millimeter-wave radar;

[0101] (2) Average queue length : The average length of the queue of vehicles extending upstream from the exit ramp (end of the drop-off area), in meters, ranging from 0 to the total length of the ramp. ;

[0102] (3) Inbound traffic The number of vehicles entering the ramp within the most recent time window (preferably a sliding window of 60 to 300 seconds), in units of vehicles per minute;

[0103] leaving the flow The number of vehicles that exited the ramp within the window in the most recent time period, in vehicles per minute.

[0104] S130, determine the traffic time period type based on the current time, the traffic time period type includes peak time period, off-peak time period, and low-peak time period, and obtain environmental parameters including weather conditions and special events.

[0105] Time period characteristics: The time period type of the current time is determined according to the system clock. The time period division includes at least peak time, off-peak time and low-peak time. Peak time can be obtained by clustering historical traffic data (e.g., morning peak 7:30-9:30, evening peak 17:00-19:00) or dynamically determined by real-time traffic threshold.

[0106] Weather conditions: obtained through access to meteorological service API or on-site meteorological sensors, including but not limited to sunny, rainy, snowy, foggy, and windy days. When it is rainy, snowy, or foggy, it is marked as "severe weather".

[0107] Special event identifier: Obtained through external input or calendar rules, including but not limited to statutory holidays, the day of a large-scale event (such as a concert, sporting event, or exhibition), which are marked as "Special Event Day".

[0108] like Figure 1 As shown, in step S200, based on the collected vehicle characteristic information, real-time traffic state parameters, the traffic time period type of the current time, and environmental parameters, a dynamic correction factor set including load correction factor, time period correction factor, environmental correction factor, and historical pattern correction factor is calculated.

[0109] The above steps specifically include:

[0110] Calculate the following four types of dynamic correction factors, each with a value range of [0.5, 1.5]:

[0111] a. Load correction factor calculate:

[0112] The dwell time is dynamically adjusted based on the real-time congestion level of the ramp. The calculation formula is as follows:

[0113]

[0114] in:

[0115] The current lane occupancy rate, The congestion warning threshold is set to 0.7.

[0116] This represents the current average queue length (meters). The total length of the drop-off ramp (in meters) ranges from 50 to 300 meters.

[0117] For the weighting coefficients, satisfying , The value ranges from 0.4 to 0.8, with 0.6 being preferred. The value ranges from 0.2 to 0.6, with 0.4 being preferred;

[0118] After calculation Limiting the amplitude: That is, the heavier the load, the smaller the correction factor (the shorter the residence time), and the minimum is no less than 0.5 times the base value.

[0119] b. Time Period Correction Factor calculate:

[0120] Values ​​are determined based on time period type:

[0121]

[0122] Preferred scheme: 0.7 during peak hours, 1.0 during off-peak hours, and 1.2 during low-peak hours.

[0123] c. Environmental Correction Factors calculate:

[0124]

[0125] in:

[0126] : 0 for sunny days, and 0.1 to 0.3 for rainy, snowy, or foggy days, with 0.2 being preferred;

[0127] The value is 0 for ordinary days, and 0.1 to 0.4 for holidays or major event days, with 0.25 being preferred.

[0128] After calculation Limiting the amplitude: ;

[0129] d. Historical pattern correction factor calculate:

[0130] Feedback adjustments are made based on the historical timeout rate of the same vehicle type within the same time period (e.g., the same hour over the past 7 days):

[0131]

[0132] in:

[0133] The historical overtime rate is calculated as follows: the proportion of vehicles whose actual dwell time exceeds the corresponding dynamic threshold within the same time period to the total number of vehicles entering during that time period.

[0134] The target timeout rate is set between 0.05 and 0.20, with 0.10 (10%) being the preferred value.

[0135] To adjust the sensitivity coefficient, the value range is 1.0 to 3.0, with 2.0 being preferred;

[0136] After calculation Limiting the amplitude: .

[0137] This embodiment incorporates load factors (occupancy rate, queue length), time period factors, environmental factors (weather, events), and historical feedback factors into the dwell time calculation model, and adopts a multiplicative fusion architecture among the factors. Unlike the single static threshold or simple time period setting of existing technologies, this invention achieves "real-time, multi-dimensional, and adaptive" adjustment of the dwell time threshold.

[0138] By introducing The correction factor dynamically adjusts the current dwell time threshold based on the deviation between the historical timeout rate and the target timeout rate, forming a closed-loop control loop of "data acquisition → timeout rate statistics → threshold correction → re-acquisition". This is an intelligent self-optimization method not recorded in existing technologies.

[0139] In load correction factor Lane occupancy rate is adopted simultaneously in China and queue length Two parameters, and weighted by coefficients The fusion approach offers greater robustness and responsiveness compared to using either occupancy rate or queue length alone. In particular, when the queue length approaches the full length of the ramp, the system will proactively reduce dwell time to prevent backflow of congestion, even if the occupancy rate has not yet reached the threshold.

[0140] Furthermore, the calculation of each correction factor can also use the following alternative scheme:

[0141] 1) Alternative calculation models for load correction factors

[0142] Alternative Option A: Dynamic Adjustment Based on Queue Length Change Rate

[0143] In the original Based on this, the rate of change of queue length is introduced. (i.e., queue growth rate). When the queue grows faster, an additional contraction factor is applied. ,in This is the sensitivity coefficient. This scheme can respond to congestion trends earlier.

[0144] Alternative Option B: Prediction Model Based on Queuing Theory

[0145] use Queuing models or more sophisticated simulation models predict the average waiting time over the next 1 to 5 minutes, and use this predicted waiting time as the basis for load adjustments. This approach requires higher computing resources but provides forward-looking control capabilities.

[0146] 2) Expansion of historical pattern correction factors

[0147] Alternative solutions: Use Long Short-Term Memory (LSTM) networks or Transformer models.

[0148] Instead of a simple linear feedback formula, this method utilizes a neural network model to learn the complex nonlinear relationship between historical dwell time series and timeout rate, predicting the current optimal [condition / condition]. Value. This solution is suitable for large hubs with sufficient data accumulation (e.g., more than 6 months).

[0149] 3) Alternative solutions for environmental correction factors

[0150] Alternative solution: Dynamic calibration based on real-time access capabilities

[0151] Instead of directly using subjective weighting of weather events, it analyzes historical data to statistically determine the actual traffic capacity reduction factor under different weather conditions. For example, if big data analysis shows that "traffic capacity decreases by 15% in rainy weather," then the corresponding reduction factor will be... Set it to 1.15. This approach is more objective.

[0152] like Figure 1 As shown, in step S300, the corresponding basic dwell time is obtained from the pre-created vehicle type-basic dwell time mapping table according to the current vehicle type. The basic dwell time is corrected and calculated using a set of dynamic correction factors, and a safety limit is applied to obtain the maximum dynamic allowable dwell time of the current vehicle.

[0153] The above steps specifically include:

[0154] S310, based on the vehicle type identified in step one, retrieve the basic dwell time of the vehicle from the preset "Vehicle Type-Basic Dwell Time Mapping Table". The mapping relationship is as follows:

[0155]

[0156] Furthermore, the adaptive optimization of the base dwell time can also employ the following alternative schemes:

[0157] Extended approach: Dynamic base value adjustment based on reinforcement learning

[0158] Base stay time Treating each vehicle type as an optimizable variable, a reinforcement learning agent (such as Q-learning or PPO algorithm) is constructed. The agent uses a multi-objective reward function, aiming to minimize the timeout rate, minimize the average dwell time, and maximize the total traffic volume, to adjust the vehicle type in real time. Value. This scheme enables fully automated adaptive control.

[0159] S320, Comprehensive calculation of dynamic allowable dwell time

[0160] Multiplying the base dwell time by each correction factor yields the initial dynamic allowable dwell time:

[0161]

[0162] in, To correct the dynamic allowable dwell time, Based on the basic stay time, As the load correction factor, For time period correction factor, As an environmental correction factor, As a factor for correcting historical patterns;

[0163] Then apply a safety limit:

[0164]

[0165] in:

[0166] The minimum allowable dwell time is 30 to 60 seconds, preferably 45 seconds;

[0167] The absolute maximum allowable dwell time is set between 300 and 600 seconds, with 480 seconds (8 minutes) being the preferred value.

[0168] like Figure 1 As shown, in step S400, a graded early warning process is performed based on the current vehicle's maximum dynamic allowable dwell time and the monitored actual dwell time, combined with a graded early warning mechanism.

[0169] The above steps specifically include:

[0170] From the moment the vehicle entered The timing begins, and the location of the corresponding vehicle is continuously tracked through the integrated radar-visual system, recording its actual dwell time. , The current moment;

[0171] when achieve When the vehicle's license plate number is between 70% and 90% (preferably 80%), a Level 1 warning (flexible reminder) is triggered: the vehicle's license plate number is displayed on the roadside LED guidance screen, and a message is sent through the parking space's geomagnetic broadcast system: "Please leave as soon as possible."

[0172] when When the violation is triggered, a Level 2 warning (warning prompt) is displayed on the guidance screen: "Excessive stay will be dealt with according to law". The system automatically records evidence of the violation (including at least three timestamped photos or a video of more than 10 seconds) and sends a reminder to the driver through the management platform system (bound to the license plate number through the railway hub's official WeChat account).

[0173] when When this occurs, a Level 3 warning (law enforcement linkage) is triggered: the system packages and pushes the evidence of the violation to the traffic management department's law enforcement platform, where law enforcement personnel decide whether to issue a non-on-site penalty or notify on-site personnel to handle the situation.

[0174] This embodiment is configured as follows: (Minimum dwell time, 45 seconds) and (Absolute maximum time, 480 seconds) Double safety boundaries prevent unreasonable control due to extreme superposition of correction factors. Meanwhile, the warning trigger point is designed to be 80%. 100% 120% A three-tiered approach is adopted to form a progressive intervention strategy of "soft reminder → warning → enforcement".

[0175] Furthermore, alternative solutions can be adopted to achieve the replacement of early warning and law enforcement:

[0176] Alternative solution: Proactive alerts based on V2X vehicle-to-infrastructure (V2X) communication;

[0177] For vehicles with vehicle-road cooperative capabilities (equipped with OBU devices), information such as dwell time threshold and remaining time can be directly sent to the vehicle's dashboard or in-vehicle navigation via the roadside RSU, replacing traditional roadside guidance screens and broadcasts, and achieving more accurate and efficient information delivery.

[0178] In this embodiment, the method further includes:

[0179] The entry time, vehicle type, actual stay time, and calculated data for each vehicle are used to calculate... Data such as whether the timeout occurred is stored in the historical database;

[0180] Recalculate the overtime rate for each time period and vehicle type every hour or day. And update the historical pattern correction factor. Historical statistics;

[0181] Optionally, the system can automatically adjust the target timeout rate based on long-term data trends. Alternatively, the weighting coefficients of each correction factor can be used to continuously optimize the control strategy.

[0182] Taking the elevated drop-off ramp of a large high-speed railway station as an example, the total length of the ramp is... The system consists of six lanes for passenger drop-off in a single direction with two rows (the inner lane closest to the high-speed rail station is for taxis and ride-hailing vehicles, while the outer lane is for private vehicles). This invention also discloses a dynamic traffic flow control system for elevated passenger drop-off ramps in a large railway hub, with equipment deployment as follows: Figure 2 As shown, it specifically includes:

[0183] Entrance detection gantry: Located 20 meters upstream of the starting point of the ramp, it is equipped with a license plate recognition camera (resolution ≥ 4 million pixels, supports supplemental lighting) and millimeter-wave radar (detection distance ≥ 200 meters, ranging accuracy ± 0.1 meters).

[0184] Mid-section radar-visual integrated unit: One radar-visual integrated unit (integrating millimeter-wave radar and high-definition video) is deployed in the middle of the ramp (75 meters from the starting point) and at the exit (145 meters from the starting point) for full-process vehicle trajectory tracking and dwell time statistics;

[0185] Roadside guidance screens: Four LED guidance screens (P10 full color, size ≥1.5m×1.0m) are deployed at the entrance, middle section and exit of the ramp to display license plate numbers and prompts.

[0186] Parking space geomagnetic sensors: A geomagnetic system is installed in the drop-off area parking spaces to identify vehicles and control dynamic broadcasts.

[0187] Broadcast loudspeakers: Four directional broadcast loudspeakers will be deployed along the ramp to cover the entire road section;

[0188] Edge computing unit: Deployed in on-site racks (CPU ≥ 8 cores, clock speed ≥ 2.0GHz, memory ≥ 16GB, hard disk ≥ 512GB SSD, supports GPU acceleration), running algorithm models;

[0189] Central Management Platform: Deployed in the hub management center computer room, it is connected to the edge computing unit via fiber optic or 5G private network to realize data aggregation, visualization display, parameter configuration and law enforcement linkage.

[0190] Furthermore, alternatives to the perception layer can also include the following:

[0191] Alternative solution: Pure LiDAR solution;

[0192] Instead of using a radar-visual integrated machine, this solution deploys 3D LiDAR for vehicle detection and tracking, combined with edge computing to achieve high-precision positioning and behavior recognition. This solution performs better at night and in inclement weather, but is more expensive and suitable for high-value hub scenarios.

[0193] It should be noted that for a detailed description of the traffic flow control system for elevated passenger drop-off ramps in large railway hubs provided in this embodiment of the invention, please refer to the relevant description of the dynamic traffic flow control method for elevated passenger drop-off ramps in large railway hubs provided in this embodiment of the invention, which will not be repeated here.

[0194] The above examples illustrate the present invention only to aid in understanding it and are not intended to limit the scope of the invention. Those skilled in the art can make various simple deductions, modifications, or substitutions based on the ideas of this invention.

Claims

1. A method for dynamic traffic flow control of elevated drop-off ramps in large railway hubs, characterized in that, The method includes: Real-time collection of vehicle characteristic information, real-time traffic status parameters, current traffic time period type, and environmental parameters for vehicles entering the drop-off ramp; Based on the collected vehicle characteristic information, real-time traffic status parameters, the traffic time period type and environmental parameters of the current time, a dynamic set of correction factors is calculated, including load correction factor, time period correction factor, environmental correction factor and historical pattern correction factor. Based on the current vehicle type, the corresponding basic dwell time is obtained from the pre-created vehicle type-basic dwell time mapping table. The basic dwell time is corrected and calculated using a set of dynamic correction factors, and a safety limit is applied to obtain the maximum dynamic allowable dwell time for the current vehicle. Based on the current maximum dynamic allowable dwell time of vehicles and the actual dwell time monitored, a tiered early warning system is implemented to handle tiered warnings.

2. The dynamic traffic flow control method for elevated passenger drop-off ramps in a large railway hub as described in claim 1, characterized in that, Real-time collection of vehicle characteristic information, real-time traffic status parameters, current traffic time period type, and environmental parameters entering the drop-off ramp, specifically including: By deploying vehicle sensing units at the entrance of the drop-off ramp, vehicles entering the drop-off ramp are detected and identified in real time, and vehicle feature information including vehicle identity information, vehicle type and entry timestamp is extracted. The vehicle types include private cars, ride-hailing vehicles, taxis and special vehicles.

3. The dynamic traffic flow control method for elevated passenger drop-off ramps in a large railway hub as described in claim 1, characterized in that, Real-time collection of vehicle characteristic information, real-time traffic status parameters, current traffic time period type, and environmental parameters entering the drop-off ramp, specifically including: By deploying integrated radar-visual devices or distributed sensor networks throughout the drop-off ramps, real-time traffic status parameters, including lane occupancy rate, average queue length, inbound traffic flow, and outbound traffic flow, are collected.

4. The method for dynamic traffic flow control of elevated drop-off ramps in large railway hubs as described in claim 1, characterized in that, Real-time collection of vehicle characteristic information, real-time traffic status parameters, current traffic time period type, and environmental parameters entering the drop-off ramp, specifically including: The traffic time period is determined based on the current time, which includes peak hours, off-peak hours, and low-peak hours. Environmental parameters, including weather conditions and special events, are also obtained.

5. The dynamic traffic flow control method for elevated passenger drop-off ramps in a large railway hub as described in claim 3, characterized in that, By deploying integrated radar-visual systems throughout the drop-off ramp area, real-time traffic status parameters are collected, including lane occupancy rate, average queue length, inbound traffic flow, and outbound traffic flow. Specifically, these parameters include: Lane occupancy rate The ratio of the length of the lane currently occupied by vehicles to the total length of the drop-off ramp; Average queue length The average length of the vehicle queue extending upstream from the ramp exit; Inbound traffic The number of vehicles that entered the ramp within the window in the most recent 30-second period; leaving the flow The number of vehicles that exited the ramp within the window in the most recent time (30 seconds).

6. The dynamic traffic flow control method for elevated passenger drop-off ramps in a large railway hub as described in claim 1, characterized in that, Based on the collected vehicle characteristic information, real-time traffic state parameters, the traffic time period type at the current moment, and environmental parameters, a dynamic set of correction factors is calculated, including load correction factor, time period correction factor, environmental correction factor, and historical pattern correction factor. Specifically, this set includes: a. Load correction factor calculate: in: The current lane occupancy rate, The congestion warning threshold is set to 0.

7. This represents the current average queue length. This is the total length of the drop-off ramps; , Let be the weighting coefficient, satisfying ; After calculation Limiting the amplitude: ; b. Time Period Correction Factor calculate: Values ​​are determined based on time period type: c. Environmental Correction Factors calculate: in: : 0 for sunny days, and 0.1 to 0.3 for rainy, snowy, or foggy days; The value is 0 for ordinary days and 0.1 to 0.4 for holidays or major event days. After calculation Limiting the amplitude: ; d. Historical pattern correction factor calculate: Feedback adjustments are made based on the historical timeout rate of the same vehicle type within the same time period (e.g., the same hour over the past 7 days). in: The historical overtime rate is calculated as follows: the proportion of the number of vehicles whose actual dwell time exceeds the corresponding dynamic threshold within the same time period to the total number of vehicles entering the corresponding time period. The target timeout rate ranges from 0.05 to 0.

20. To adjust the sensitivity coefficient, the value range is 1.0 to 3.0; After calculation Limiting the amplitude: .

7. The dynamic traffic flow control method for elevated passenger drop-off ramps in a large railway hub as described in claim 1, characterized in that, The base dwell time is corrected and calculated using a set of dynamic correction factors, and a safety limit is applied to obtain the maximum dynamic permissible dwell time for the current vehicle. Specifically, this includes: in, To correct the dynamic allowable dwell time, Based on the basic stay time, As the load correction factor, For time period correction factor, As an environmental correction factor, As a factor for correcting historical patterns; Then apply a safety limit: in: The minimum allowable dwell time ranges from 30 to 60 seconds. This represents the absolute maximum permissible dwell time, ranging from 300 to 600 seconds. This represents the maximum dynamic allowable dwell time that is ultimately obtained.

8. The dynamic traffic flow control method for elevated passenger drop-off ramps in a large railway hub as described in claim 1, characterized in that, Based on the current maximum dynamic permissible dwell time of vehicles and the monitored actual dwell time, a tiered early warning management system is implemented, which specifically includes: From the moment the vehicle entered The timing begins, and the location of the corresponding vehicle is continuously tracked through the integrated radar-visual system, recording its actual dwell time. , For the current moment; when achieve When the number of vehicles reaches 70% to 90%, a Level 1 warning is triggered: the license plate number of the corresponding vehicle is displayed on the roadside LED guidance screen, and a message is sent through the parking space geomagnetic broadcast system: "Please leave as soon as possible." when When this occurs, a Level 2 warning is triggered: the guidance screen displays "Excessive stay will be dealt with according to law," automatically records evidence of the violation, and sends a reminder to the driver through the management platform system; when When a Level 3 warning is triggered, the evidence of the violation will be packaged and pushed to the traffic management department's enforcement platform, whereby law enforcement officers will decide whether to issue a non-on-site penalty or notify on-site personnel to handle the situation. This represents the maximum dynamic allowable dwell time for the current vehicle.

9. A dynamic traffic flow control method for elevated passenger drop-off ramps in a large railway hub as described in claim 6, characterized in that, The method further includes: This will include the entry time of each vehicle, vehicle type, actual stay time, and calculated data. Data on whether or not the timeout has occurred is stored in the historical database; Recalculate the overtime rate for each time period and vehicle type every hour or day. And update the historical pattern correction factor. Historical statistics; The system can automatically adjust the target timeout rate based on long-term data trends. Alternatively, the weighting coefficients of each correction factor can be used to continuously optimize the control strategy.

10. A dynamic traffic flow control system for elevated drop-off ramps in a large railway hub, used to execute the dynamic traffic flow control method for elevated drop-off ramps in a large railway hub as described in any one of claims 1-9, characterized in that, The system includes: Entrance detection gantry: Located 20 meters upstream of the starting point of the ramp, it is equipped with a license plate recognition camera and millimeter-wave radar to detect and identify vehicles entering the drop-off ramp in real time; Mid-section radar-visual integrated unit: Radar-visual integrated units integrating millimeter-wave radar and high-definition video are deployed in the middle of the ramp and at the exit, for the purpose of tracking vehicle trajectory and counting dwell time throughout the entire process; Roadside guidance screens: LED guidance screens are deployed at the entrance, middle section, and exit of the ramp to display license plate numbers and prompts; Parking space geomagnetic sensors: A geomagnetic system is installed in the drop-off area parking spaces to identify vehicles and control dynamic broadcasts. Broadcast speaker columns: Multiple directional broadcast speaker columns will be deployed along the ramps to cover the entire road section; Edge computing unit: Deployed in on-site cabinets to run algorithm models; Central Management Platform: Deployed in the hub management center computer room, it is connected to the edge computing unit via fiber optic or 5G private network to realize data aggregation, visualization display, parameter configuration and law enforcement linkage.

Images