Vehicle-road collaborative management and control system for typical scene of urban expressway

Abstract

The invention discloses a vehicle-road collaborative management and control system for typical scenes of urban expressways, and belongs to the field of traffic. The system comprises a road scene recognition system, a multi-source traffic information collection system, a multi-source traffic data processing system, a vehicle-road cooperative control system and a management and control method selection system. According to the research, the driving efficiency of intelligent network connection vehicles and the operation effect of mixed traffic flow in three typical urban expressway scenes including an urban expressway basic road section, an expressway exit and an expressway entrance are considered; and designing an optimal vehicle collaborative management and control strategy of the intelligent network connection vehicle under three types of urban expressway typical scenes, different CAV permeability and different traffic demands, and maximizing the traffic flow operation efficiency. According to the method, for three typical traffic scenes (such as basic road sections, entrance and exit road sections and the like of the urban expressway) in the urban expressway, the limitation that design is only oriented to a single scene in the prior art is broken through, collaborative optimization of traffic flow can be achieved in various complex urban traffic environments, and the overall utilization rate of road resources is effectively improved.

Description

Technical Field

[0001] This invention belongs to the field of transportation. Background Technology

[0002] With rapid urban development and the continuous growth of motor vehicle ownership, urban expressways, as vital channels connecting the main urban area with surrounding regions, are facing unprecedented traffic pressure. Problems such as traffic bottlenecks, congestion at key points, frequent interference at entrances and exits, and declining traffic efficiency are becoming increasingly prominent, severely restricting the expressways' intended rapid traffic flow function. Simultaneously, due to the lack of a unified and coordinated management mechanism and intelligent control measures, the operational efficiency of expressways is difficult to optimize, and traffic safety hazards and environmental burdens are also increasing. These problems not only affect citizens' travel experience but also pose challenges to the overall stability and sustainable development of the urban transportation system.

[0003] Currently, research in the field of intelligent transportation mainly focuses on the design of control strategies for specific scenarios, typically optimizing under specific traffic conditions, such as the entrances and exits of urban expressways. While these studies have achieved some results in specific scenarios, they lack comprehensive consideration of various typical urban expressway scenarios. Therefore, existing control strategies often appear limited and ineffective when facing complex urban traffic environments. Especially with the low penetration rate of autonomous vehicles, the control effect of vehicle-road cooperative systems is often insufficient, failing to effectively alleviate traffic pressure and reduce congestion.

[0004] Furthermore, with the rapid development of connected vehicles (CAVs), more and more autonomous and assisted driving vehicles are entering the roads, changing the structure of traffic flow. Traditional traffic management methods often lack adaptability to the mixed flow of autonomous and human-driven vehicles, causing management strategies to fail to achieve the desired results in the new traffic environment. Therefore, how to design an efficient management strategy that takes into account both human-driven and autonomous vehicles has become an urgent problem to be solved in the field of intelligent transportation.

[0005] The invention patent with application number "CN202411348830.3" for autonomous vehicle platoon control method provides an autonomous vehicle platoon control method and system based on operating condition recognition. The method involves the lead vehicle perceiving the surrounding environment, judging the current driving conditions of the platoon, and adjusting the speed of the following vehicles according to the expected distance between adjacent vehicles.

[0006] The mixed traffic flow optimization control method combining variable speed limit and lane changing, with application number "CN202210258362.5", provides a mixed traffic flow optimization control method based on the combination of variable speed limit and lane changing. It uses vehicle networking technology to customize variable speed limit values ​​for each vehicle at different locations and directly sends the speed limit to the corresponding individual vehicle.

[0007] The method for controlling and evaluating urban expressway weaving areas with reduced game time, as proposed in application number "CN202410313815.9", provides a method for formulating control strategies for expressway weaving areas based on the geometric design parameters of the expressway weaving areas and the significance test results of traffic flow. Through the established vehicle coding rules, dynamic strategy management of vehicle game under different conditions can be achieved, thereby reducing vehicle game time and improving road capacity.

[0008] Current patent literature largely focuses on improving the operational efficiency of urban expressways through autonomous driving platooning control, speed-limited lane-changing coordination optimization, and dynamic management of weaving zones, achieving positive results. However, overall, existing research mostly focuses on single typical scenarios of urban expressways, such as platooning, lane changing, or weaving zone behavior optimization. The designed control strategies are often targeted at specific traffic functions or operating conditions, lacking a systematic and comprehensive control strategy for multiple scenarios, multiple elements, and multiple modes of traffic conditions on urban expressways. Therefore, there is an urgent need for a more flexible, dynamic, and efficient vehicle-road cooperative control strategy to adapt to the changing environment under different CAV penetration rates and traffic scenarios. By designing corresponding control strategies based on the characteristics of different traffic scenarios, traffic flow optimization, road utilization improvement, and congestion reduction can be achieved, thereby promoting the development of urban traffic management towards intelligence and refinement.

[0009] To address the following problems and shortcomings in the existing urban transportation system:

[0010] (1) Existing control strategies are generally designed for single traffic scenarios and lack unified coordination and optimization for multiple typical traffic scenarios, making it difficult to meet the diverse traffic operation needs of complex urban expressway systems.

[0011] (2) The existing solution is not effective when the CAV penetration rate in the vehicle-road cooperative system is low. It lacks the ability to dynamically adapt to traffic flow characteristics under different CAV penetration rates, resulting in poor adaptability and practicality under different application conditions.

[0012] (3) Existing technologies have limited effectiveness in addressing urban traffic congestion and improving traffic flow efficiency. They cannot effectively alleviate problems such as congestion, delays, and stagnation in traffic operations, thus affecting road capacity and operational quality.

[0013] The technical problem to be solved by this invention is to provide a vehicle-road cooperative management strategy that is applicable to a variety of typical traffic scenarios on urban expressways, has good adaptability to CAV penetration, and can improve traffic flow efficiency and alleviate traffic congestion, so as to realize intelligent, efficient and flexible management of urban expressway traffic systems. Summary of the Invention

[0014] To address the above issues, this invention discloses a vehicle-road cooperative management and control system for typical urban expressway scenarios, comprising a road scene recognition system, a multi-source traffic information collection system, a multi-source traffic data processing system, a vehicle-road cooperative control system, and a management and control method selection system.

[0015] The specific technical solution is as follows:

[0016] This invention discloses a vehicle-road cooperative management and control method for typical urban expressway scenarios, including a road scene recognition system, a multi-source traffic information collection system, a multi-source traffic data processing system, a vehicle-road cooperative control system, and a management and control method selection system. This research considers the driving efficiency of intelligent connected vehicles and the operational effects of mixed traffic flow in three typical urban expressway scenarios: basic expressway sections, expressway exits, and expressway entrances. It designs optimal vehicle cooperative management and control strategies for intelligent connected vehicles under different CAV penetration rates and traffic demands in these three typical urban expressway scenarios to maximize traffic flow efficiency.

[0017] Road scene recognition system

[0018] The core function of this system is to collect and analyze road information in real time through the vehicle's own perception module and vehicle-to-everything (V2X) module to determine the current road scenario in which the vehicle is located. The system utilizes onboard sensors (such as LiDAR, cameras, and ultrasonic sensors) and V2X communication technology to acquire information on surrounding traffic conditions, road surface conditions, and traffic signs, thereby accurately identifying the type of road scenario the vehicle is in (such as a basic section of an urban expressway, an entrance / exit section, etc.). The system can monitor changes in the road environment in real time, providing precise scenario data support for subsequent management and control strategies.

[0019] Multi-source traffic information collection system

[0020] This system collects various traffic information, including road traffic volume, CAV penetration rate, lane position, vehicle speed, and distance to key locations, through the vehicle's perception module and vehicle-to-everything (V2X) module. Specifically, it includes:

[0021] Road traffic volume: Real-time traffic volume of each lane is obtained through the sensing module;

[0022] CAV penetration rate: The ratio of autonomous vehicles (CAVs) to manually driven vehicles in the surrounding area;

[0023] Vehicle position and lane: Determine the lane in which the vehicle is located and its position relative to other vehicles;

[0024] Distance to key locations: including expressway exits, expressway entrances, etc.;

[0025] Vehicle speed: Collects speed information of the vehicle itself and surrounding vehicles;

[0026] Headway: Collects time distance information between the vehicle and surrounding vehicles. This multi-source information will help determine traffic flow conditions and provide basic data for optimizing control strategies.

[0027] Multi-source traffic data processing system

[0028] Based on the collection of multi-source traffic information, this system performs fusion processing and feature calculation on raw data from video monitoring, geomagnetic coils, radar, on-board units (OBU), roadside units (RSU), and other sensors to form reliable traffic operation status indicators, providing support for vehicle-road cooperative management and control decisions.

[0029] First, the system performs unified preprocessing on multi-source data. All data is synchronized according to a unified timestamp, and vehicle positioning data is projected onto the actual road centerline coordinate system using a map matching algorithm to achieve spatiotemporal alignment. For data with noise or missing values, the system uses linear interpolation and median filtering for correction, and identifies outlier samples using an outlier detection method (based on Z-score and IQR threshold), thereby ensuring the stability and reliability of the input data.

[0030] In the data calculation phase, the system performs real-time calculations on core traffic parameters. Instantaneous vehicle speeds are obtained directly from location tracking or radar speed measurement, while interval average speeds are calculated using a weighted average or harmonic average.

[0031]

[0032] Where: v i (t): The instantaneous velocity of the i-th vehicle;

[0033] N: The total number of vehicles detected within the current time t;

[0034] Average speed of the road segment;

[0035] The average speed of the road section.

[0036] Traffic flow q is calculated from the number of vehicles passing through the detection section per unit time:

[0037]

[0038] Where: q(t): the vehicle flow rate passing through the detection section within the time interval Δt (unit: veh / h);

[0039] Δt: Observation time interval.

[0040] Traffic density k is derived from the ratio of the number of vehicles on a road segment to the length of the road segment:

[0041]

[0042] Wherein: the number of vehicles within the detection interval length L;

[0043] L: Length of the road segment being inspected (unit: km);

[0044] k(t): Vehicle density per unit length (veh / km).

[0045] And satisfy the basic traffic relationship formula The vehicle's headway is obtained by detecting the time difference between the vehicle's departure and arrival at the designated point.

[0046] h t =t i -t i-1

[0047] Wherein: the timestamp of the i-th vehicle passing through the detection section;

[0048] t i-1 : The timestamp of the (i-1)th vehicle passing through this section;

[0049] h t Time interval between the front ends of two adjacent vehicles (unit: seconds).

[0050] Average headway h t Reflecting the compactness of traffic flow, it is an important auxiliary parameter for density estimation. Occupancy is calculated as the ratio of the occupancy time of a detection unit to the observation period.

[0051]

[0052] Wherein: T occ The total time that the detection unit is occupied by vehicles during the observation period;

[0053] T: Observation period length;

[0054] Occupancy: The percentage of inspection points occupied (reflecting the density of traffic flow).

[0055] To reduce the impact of sensor noise, the system employs a joint algorithm combining exponentially weighted moving average (EMA) and Kalman filtering. The EMA smoothing formula is:

[0056]

[0057] Wherein: the currently observed velocity value;

[0058] Speed ​​estimate after smoothing;

[0059] α: Smoothing coefficient, which controls the weighting of new and old observations (generally taken as 0.1 to 0.3);

[0060] Δt: Sampling time step.

[0061] When multiple sources of velocity or location information exist, a weighted fusion model is used:

[0062]

[0063] in: The variance of the measurement error of the i-th data source is used to represent its uncertainty.

[0064] x i : The measurement value of the i-th data source (such as video detection, radar, OBU);

[0065] w i The weight of the data source is defined as the reciprocal of the measurement variance.

[0066] During the real-time calculation phase, the system employs a multi-scale sliding window mechanism (1s, 10s, 60s, 5min) for rolling statistics, and utilizes an incremental update algorithm to quickly calculate the mean, variance, and distribution characteristics. To achieve abnormal traffic or congestion detection, the system constructs a congestion index model based on speed and density.

[0067]

[0068] Where: v free Average velocity under free-flow conditions (e.g., 80 km / h);

[0069] k crit The critical density at which traffic flow reaches its maximum capacity;

[0070] α: Adjustment coefficient, used to balance the combined effects of speed decay and density increase on the congestion index (generally taken as 0.5 to 1.5).

[0071] When CI(t) is close to 1, it indicates severe congestion; when it is close to 0, it indicates smooth traffic.

[0072] Various parameters (such as free flow velocity v) free Critical density k crit The smoothing coefficient (α, etc.) can be determined by calibrating historical traffic data or by offline training based on traffic simulation platforms (such as SUMO and VISSIM) to ensure the applicability of the model in different road sections and scenarios.

[0073] Furthermore, the system can predict short-term traffic conditions using machine learning models (such as XGBoost or lightweight LSTM). Input features include traffic flow, density, average speed, headway, and occupancy rate. The output is the probability of speed and congestion level within the next 5 to 30 minutes, providing a forward-looking decision-making basis for the vehicle-road cooperative control system.

[0074] Ultimately, the system outputs structured traffic status data, including average speed, flow rate, density, occupancy rate, average headway, congestion index and its confidence level, and updates it in real time through edge computing nodes to ensure data transmission latency of less than 200ms, providing high-precision and low-latency basic data support for subsequent vehicle-road cooperative control modules.

[0075] Vehicle-Road Cooperative Control System

[0076] This system is responsible for implementing traffic control through vehicle-road cooperation based on different expressway scenarios and traffic conditions. For different road scenarios, the system will select and optimize control strategies using three pre-established typical scenario models. These models are based on existing CAV (Consumer Access Control) methods and, considering scenario characteristics (such as basic sections and entrance / exit sections of urban expressways) and traffic flow characteristics, establish control strategy models adapted to different CAV penetration rates and traffic demands. Based on real-time traffic data, the control system can dynamically adjust control strategies to ensure optimized traffic flow and improved road utilization efficiency.

[0077] (1) Basic road segment scenario control strategy

[0078] Existing vehicle-road cooperative systems typically employ fixed lane management strategies when handling mixed traffic flows, resulting in low utilization rates of bus lanes during off-peak hours, while regular lanes are prone to congestion. Furthermore, the lack of dynamic and intelligent lane-changing strategies prevents the effective utilization of road resources while ensuring bus priority. To improve road resource utilization efficiency and facilitate the coordinated passage of traditional vehicles, highway buses, and intelligent connected vehicles in basic road segment scenarios, this invention proposes a dynamic bus lane borrowing strategy based on CAV (Continuous Aerial Vehicle) behavioral decision-making. This strategy, by judging lane-changing motives and safety conditions, allows CAVs to dynamically borrow bus lanes without interfering with bus priority, thereby improving overall traffic efficiency.

[0079] 1) Strategy Steps

[0080] The strategy mainly consists of the following three stages:

[0081] ① Determining the Motive for Changing Lanes

[0082] Whether a CAV (Carrier Available Vehicle) has the motivation to switch lanes to the bus lane is the primary entry point for strategy execution. At this point, the system will comprehensively assess whether there are traffic restrictions in the current lane and whether adjacent bus lanes offer better traffic conditions.

[0083] When the distance between this vehicle and the vehicle in front is insufficient to support the current acceleration increase or approach the maximum desired speed, it indicates that the driving environment is limited:

[0084]

[0085] In the formula: d i (t)——Distance between the front of the vehicle and the vehicle in front;

[0086] —The current speed of the vehicle (m / s);

[0087] Δt—Decision time interval (s), usually taken as 0.1 to 1 second;

[0088] —Acceleration of the vehicle in front (m / s²) 2 );

[0089] v max —Maximum permissible speed of the vehicle (m / s);

[0090] l — the length of the CAV vehicle.

[0091] If there is a larger headway in front of the bus lane on the right, it indicates that there is room to improve traffic efficiency.

[0092]

[0093] In the formula: d i-rl (t)——The distance between the vehicle and the vehicle in front of the adjacent right-hand bus lane.

[0094] Based on the MOBIL model framework, the overall benefit ε of lane changing is calculated to see if it exceeds a set threshold E. ε is determined by the change in its own acceleration and its impact on surrounding vehicles.

[0095]

[0096] a SV —The current acceleration of the target vehicle;

[0097] —The acceleration of the target vehicle after changing lanes;

[0098] a FVi —The current acceleration of the vehicle behind the target vehicle in the original lane;

[0099] —The acceleration of vehicles following the target vehicle in the original lane after the target vehicle changes lanes;

[0100] a PFVi —The current acceleration of the vehicle behind in the target lane;

[0101] —The acceleration of the vehicle following the target vehicle in the target lane after the target vehicle changes lanes;

[0102] E(k) – Lane-changing threshold. The lane-changing threshold E is determined by calculating the ε distribution offline based on historical trajectories and simulation samples, and using ROC / grid search to determine the baseline value E0. It is dynamically adjusted using an adaptive function E(k) in combination with traffic density or congestion index. At the same time, hard safety constraints are applied to the maximum deceleration of the following vehicle and the minimum headway before the decision is made. Lane changing is only allowed when ε>E(k) and the safety constraints are met.

[0103] ε—Lane-changing efficiency;

[0104] p – Politeness coefficient, set to 0.8.

[0105] Adaptive formula:

[0106]

[0107] Where: β is the calibration coefficient (β∈[0.2,1.0] is recommended);

[0108] k: Traffic density.

[0109] If the benefit ε exceeds the set threshold E, then lane changing is considered valuable, and the motivation for lane changing is valid. Only when all three conditions are met will the system proceed to the next stage of safety assessment.

[0110] ② Determining the safety conditions for lane changing

[0111] If a CAV has the motivation to change lanes, it is necessary to further determine whether the lane-changing behavior is safe and ensure that it will not affect vehicles behind in the bus lane.

[0112] First, determine whether this will force vehicles behind in the target lane to brake suddenly:

[0113] d i-rf (t)-l>v rf (t)·Δt

[0114] In the formula: d i-rf (t)——Distance between the target vehicle and the vehicle behind it in the right-hand bus lane;

[0115] v rf (t)——The current speed (m / s) of the following vehicle in the target lane.

[0116] This indicates that after changing lanes, vehicles in the target lane have enough time to react and do not need to decelerate suddenly.

[0117] Next, determine whether the safety distance requirements are met:

[0118] d i-rf (t)>d safe -l

[0119] In the formula d safe —Minimum safe headway.

[0120] This is the basic safety standard for changing lanes laterally, to avoid vehicles getting too close.

[0121] If either of the above two conditions is not met, the lane-changing attempt will be rejected, and the vehicle will continue to travel in the original lane.

[0122] ③ Priority constraints for public transport vehicles

[0123] Even if the motivation to change lanes is strong and the lane change itself is safe, the system still needs to further determine whether it interferes with the passage of buses in the bus lane. If the vehicle behind the target lane is a bus, the system introduces a "bus clearance distance" constraint to ensure that the CAV (Carrier Air Vehicle) cannot forcibly insert itself too close in front of the bus.

[0124] d i-rf (t)-l>L C

[0125] Among them, the minimum clearing distance L for buses C for:

[0126]

[0127] In the formula: —Bus speed;

[0128] b max —The maximum deceleration of the bus is set to 2.5 m / s² by default. 2 ;

[0129] t h —Following the lead time;

[0130] s0 — Minimum static safety distance.

[0131] CAVs are only allowed to change lanes in front of buses if the clearance distance condition is met, to ensure that buses are not interfered with.

[0132] Once the above-mentioned motivation assessment, safety verification, and bus priority check are all passed, the system can instruct the CAV to perform a lane-changing operation, briefly using the bus lane to improve its own and the overall system's traffic efficiency. If any step determines that the conditions are not met, the current lane-changing request is cancelled, and the CAV continues to operate in the current lane.

[0133] 2) Implementation Mechanism

[0134] When a CAV is traveling on a regular road, if it meets both the motivation and safety conditions for lane changing and does not violate the minimum clearance distance requirement for buses, the system will control it to change lanes to the bus lane; otherwise, it will maintain its current lane to ensure a balance between traffic safety and efficiency.

[0135] (1) This invention proposes a variety of targeted control strategies for three typical traffic scenarios in urban expressways (such as the basic road section and the entrance and exit road sections of urban expressways). It breaks through the limitation of existing technologies that are only designed for a single scenario. It can achieve coordinated optimization of traffic flow in a variety of complex urban traffic environments and effectively improve the overall utilization rate of road resources.

[0136] (2) This invention fully considers the different performance of vehicle-road cooperative systems under different CAV penetration rates, constructs multi-level adjustable control strategies for each typical scenario, and can dynamically select the optimal solution according to the actual vehicle network penetration rate, effectively solving the problem that the control effect of existing technologies is not obvious under low penetration rates, thereby enhancing the practicality and adaptability of the control system.

[0137] (3) By making comprehensive use of a flexible combination of multiple scenarios and strategies, this invention can significantly reduce traffic flow conflicts, stagnation and inefficient operation, improve traffic flow organization, enhance overall traffic capacity, effectively alleviate traffic congestion during peak hours and in special scenarios, and improve the quality and service level of urban expressway traffic operation.

[0138] (4) The control strategy proposed in this invention has good feasibility and scalability, is applicable to different urban expressway environments and traffic operation conditions, has high engineering feasibility and promotion and application value, and is of great significance to the construction of intelligent transportation systems. Attached Figure Description

[0139] Figure 1 Schematic diagram of dedicated lane layout.

[0140] Figure 2 Schematic diagram of variable speed limits at expressway exits. Detailed Implementation

[0141] To adapt to different stages of CAV penetration, this invention also designs three typical implementation strategies:

[0142] Option 1: No dedicated CAV lane is set up. CAVs utilize the bus lane and can share the road with conventional vehicles. Figure 1 Scenario 1;

[0143] Option 2: Establish a dedicated CAV lane, where CAVs can use both the dedicated CAV lane and the bus lane simultaneously, but cannot share the road with conventional vehicles. Figure 1 Scenario 2;

[0144] Option 3: Set up a dedicated CAV lane, while allowing mixed traffic. Figure 1 Scenario 3.

[0145] When the CAV penetration rate is below 0.3, Scheme 1 is adopted as the control scheme; when the penetration rate is in the range of [0.3-0.5], Scheme 2 is adopted as the implementation scheme; when the penetration rate is in the range of [0.5,1], Scheme 3 is adopted as the operation scheme.

[0146] (2) Expressway Exit Control Strategy

[0147] Existing expressway exit control methods primarily rely on fixed speed limit signs or simple traffic flow guidance, which cannot be dynamically adjusted according to real-time traffic conditions. When upstream traffic volume surges suddenly, these static strategies fail to effectively restrict vehicles entering the bottleneck area, easily leading to severe downstream congestion and even triggering a chain reaction that affects mainline traffic. Furthermore, existing systems typically do not consider the characteristics of intelligent connected vehicles and cannot fully utilize their rapid response capabilities to optimize control strategies.

[0148] To mitigate the reduced capacity and congestion risks caused by traffic flow convergence in expressway exit areas (i.e., divergence zones), a guidance strategy based on Variable Speed ​​Limit (VSL) is proposed. The traffic complexity in divergence zones is generally higher than in basic road sections, and their capacity is significantly reduced by changes in physical structure and flow direction. If upstream traffic volume consistently exceeds the maximum service capacity of the divergence zone, it can easily lead to queuing, congestion, and even spillover effects to further upstream areas. Therefore, a speed limit control zone is established upstream of the divergence zone to reduce the inflow per unit time by limiting vehicle speed, effectively preventing local oversaturation. To this end, this strategy is set as follows: Figure 2 The three functional zones shown are: Variable Speed ​​Limit Control Zone (Speed ​​Limit Segment 1): Located upstream of the bottleneck zone, it dynamically controls the speed and traffic density of vehicles entering the bottleneck, limiting the number of vehicles entering the bottleneck zone per unit time; Acceleration Transition Zone (Speed ​​Limit Segment 2): Located between the speed limit zone and the bottleneck zone, it is used to gradually restore vehicles to free-flow speed after passing through the VSL zone, ensuring that vehicles enter the bottleneck section in the optimal state; Bottleneck Zone: i.e., the expressway exit or divergence point, its capacity is the core constraint for the entire system's regulation.

[0149] The system collects real-time traffic density and speed information from each cell on the expressway, and simultaneously obtains the current vehicle-to-everything (V2X) autonomous vehicle (CAV) penetration rate γ. CAVs have fast response speeds and short following distances, therefore, the length of the speed-limited zone can be shortened under high penetration conditions. The distance between the speed-limited section and the exit is set as follows under different CAV penetration rates: when γ∈[0.1,0.9], to ensure the speed limit effect is fully realized, the distance from the end of the speed-limited section to the exit is set to 900 meters; when γ>0.9, considering the fast response speed and short following distance of CAVs, the length of the speed-limited section is set to 700 meters to improve traffic efficiency. By initializing traffic data and key parameters, the spatial basis for speed limit control is established.

[0150] To describe the real-time evolution of traffic conditions, this paper employs a Cellular Transport Model (CTM) for modeling and computation. As the computational core of the system, the CTM model enables dynamic simulation updates of traffic density, speed, and flow, serving as the foundation for judging congestion trends, setting speed limits, and adjusting speed-limited zones. It ensures the entire guidance and control process possesses computability, responsiveness, and a feedback mechanism. The road is divided into equal-length cells, with n vehicles per cell. i,j The principle of vehicle conservation is satisfied as it changes over time:

[0151] n i,j =n i-1,j +y i-1,j-1 -y i,j-1

[0152]

[0153] Ni=ρj·Δx

[0154] In the formula: v f —Free-flow vehicle speed;

[0155] n i,j — The number of vehicles in cell i at time j;

[0156] y i,j —The traffic flow transmitted from cell i to cell i+1 at time j;

[0157] N i,j —The capacity of cell i at time j;

[0158] Q i,j —The passage capacity of cell i at time j;

[0159] w—Propagation speed of the congestion reverse wave;

[0160] ρ j —The theoretical congestion (saturation) density (veh / km) of lane / road segment at time j;

[0161] Δx — Cell length (km).

[0162] To achieve dynamic guidance, it is necessary to first determine whether there is a risk of congestion. To detect the congestion risk in bottleneck areas, the system determines in real time whether the target area shows an oversaturation trend (i.e., the traffic density exceeds the critical density ρ). c2 When the system detects that traffic flow begins to decrease and the density continues to rise beyond the critical value ρ, c2 When a vehicle is identified as "potentially congested," the speed limit control module is triggered to intervene in the upstream vehicle speed.

[0163] Speed ​​limits are dynamically calculated based on traffic flow parameters, with the goal of reducing traffic speeds to maintain the maximum bottleneck capacity Q. max The state. Combining the traffic three-phase flow model, the vehicle speed v... i,j Expressed as density ρ i,j Piecewise functions:

[0164]

[0165] In the formula: ρ c1 —Critical density for free flow;

[0166] ρ c2 —Critical congestion density;

[0167] ρ c —Theoretical critical density, where traffic flow reaches maximum flow rate Q max The density;

[0168] ρ j —The upper limit of congestion density can be estimated by the concentration point at the right end of the actual road segment flow-density map (observing the congestion limit density);

[0169] u w —Blocking wave speed.

[0170] When congestion is imminent, the system determines the density ρ of the upstream control area. i Based on this, the controllable maximum flow rate Q is calculated. sl and speed limit value v sl :

[0171]

[0172] This rate limit is used to constrain the transmission capacity of the upstream cell:

[0173]

[0174] At the same time, adjustments were made to the downstream receiving capacity:

[0175]

[0176] In the formula: α i,j --The sending capability of cell i;

[0177] β i+1,j —The receiving capacity of cell i+1;

[0178] n i — The number of vehicles in cell i;

[0179] v i,j-1 —The average velocity of the i-th cell at time j-1.

[0180] This creates traffic flow constraints under speed limits, ultimately outputting the actual traffic volume:

[0181] q i,j =min{α i,j ,β i+1,j}

[0182] Once the speed limit policy is activated, vehicles in the guidance zone will proceed at the set speed limit v. sl Run, and gradually restore to free-flow velocity v in the downstream buffer. f The acceleration zone ensures vehicles can smoothly pass through bottleneck sections, preventing secondary congestion and cutting in, and improving mainline traffic efficiency. This process achieves a smooth transition of traffic conditions by continuously updating cell density and speed. Vehicle acceleration does not cause conflicts, effectively improving traffic safety and comfort.

[0183] The entire VSL control system operates in a closed-loop structure. As traffic conditions and CAV penetration change, the system updates density and flow parameters in real time, dynamically adjusting the length of the speed-limited zone and the speed limit value v. sl This enables continuous management and guidance of bottleneck areas. By linking with the CAV sensing system, the system can achieve second-level response and dynamic strategy iteration, further improving control accuracy and system robustness.

[0184] (3) Expressway entrance control strategy

[0185] To address traffic disruptions caused by the merging of ramps and main road traffic at expressway entrances, a virtual fleet collaborative control strategy based on a vehicle-to-everything (V2X) environment is proposed. This strategy targets vehicles on both main and auxiliary roads within the merging area, effectively mitigating merging conflicts and improving road efficiency and operational stability through coordinated vehicle scheduling and longitudinal control.

[0186] In existing expressway systems, vehicles on ramps competing for limited space with those on the main road when entering the main road can easily experience deceleration, swerves, and disturbance waves, causing fluctuations in upstream traffic flow or even congestion, severely impacting operational efficiency and safety. Especially during peak hours or when ramp traffic is heavy, merging areas become bottlenecks in the traffic system.

[0187] This invention, supported by a vehicle-to-everything (V2X) environment, utilizes real-time information interaction between vehicles to construct a virtual fleet model without physical connections. This model enables unified scheduling and acceleration control of vehicles in merging areas, thereby guiding vehicles through merging points in an orderly manner. Simultaneously, a collaborative lane-changing mechanism is designed based on the current operating status of vehicles and the level of system congestion to further enhance the ramp's capacity.

[0188] The implementation mechanism of this control strategy is as follows:

[0189] 1) Vehicle information perception and arrival time prediction

[0190] Using a vehicle-to-everything (V2X) sensing system, the speed v of vehicles on the main road and ramps, the distance d between their current location and the merging point are obtained in real time, and their estimated arrival time at the merging point is calculated.

[0191]

[0192] In the formula: —The estimated time, in seconds, for the i-th vehicle on the main road to arrive at the merging point;

[0193] —The estimated arrival time of the i-th vehicle on the auxiliary road at the merging point, s;

[0194] — The distance between vehicles on the main road and the merging point, in meters;

[0195] —Distance between vehicles on the auxiliary road and the merging point, in meters;

[0196] —The speed of vehicles on the main road at the current moment, in m / s;

[0197] —The speed of vehicles on the main road at the current moment, in m / s.

[0198] 2) Virtual fleet construction and sorting control

[0199] The system sorts all vehicles approaching the merging point according to their estimated arrival time, generating a virtual convoy sequence arranged chronologically. Vehicles do not require physical formation, but respond to the actions of the vehicle ahead in a queue-like manner within the control logic, achieving virtual coordination.

[0200] 3) Longitudinal acceleration control model

[0201] In the virtual convoy, each vehicle calculates its own acceleration based on the status of the vehicle in front (acceleration, speed, and distance).

[0202] a n (t)=K a *a n-1 (t)+K d [Δx n (t)-E d (t)]+K v [v n-1 (t)-vn ( t)]#

[0203] In the formula: a n (t)——The acceleration of vehicle n at time t, m / s² 2 ;

[0204] a n-1 (t)——The acceleration of the leading vehicle n-1 of vehicle n at time t;

[0205] Δx n (t)——The distance between vehicle n and its leading vehicle n-1 at time t;

[0206] v n (t)——The speed of vehicle n at time t, m / s;

[0207] K a ,K d ,K v —The control coefficients of the model are calculated using measured vehicle trajectories in the merging zone and particle swarm optimization.

[0208] The method uses speed error as the minimum objective for calibration;

[0209] E d (t)——The expected following distance between the target vehicle n and the leading vehicle n-1.

[0210] 4) Calculation of expected spacing (refer to IDM model)

[0211] Where the expected spacing E d (t) Dynamic adjustments are made based on the IDM (Intelligent Driver Model):

[0212]

[0213] Where: T—the expected headway in the IDM model, taken as 1.5s;

[0214] d — the following distance in congested traffic, in meters;

[0215] a——Maximum vehicle acceleration, m / s²2 ;

[0216] b — Vehicle comfort deceleration, m / s 2 .

[0217] Among them, a, b, and d can all be calibrated using particle swarm optimization algorithm based on measured data. Simultaneously, the system also sets a basic distance threshold based on CAV permeability for adjusting controller initialization parameters. When the permeability is less than 0.1, the distance threshold is set to 350m; when the permeability is in the range [0.1, 0.3], the distance threshold is 150m; and when the permeability is greater than 0.3, the distance threshold is set to 450m.

[0218] 5) Freewheel acceleration control

[0219] When the target vehicle is the lead vehicle of the convoy or there are no vehicles within the target vehicle's influence range, the vehicle is in a free-driving state and is considered a free-moving vehicle not participating in the virtual convoy dispatch. Its acceleration is calculated based on the difference between the current speed and the road speed limit.

[0220]

[0221] In the formula: —Acceleration outside the convoy;

[0222] —This represents the speed limit for the road, and the acceleration is constrained by the maximum / minimum limit and the rate of change.

[0223] The following constraints are set to ensure smooth and comfortable vehicle control:

[0224]

[0225] In the formula: a differ —Maximum rate of change of acceleration.

[0226] 6) Coordinated lane-changing control mechanism

[0227] When lane changing is permitted in the outer lane of the main road, the system checks for on-ramp vehicles nearby. If present, it initiates the lane-change decision process. Lane changes must meet the following conditions:

[0228] (1) The system determines the spatial and speed relationships between the vehicles in front and behind the target lane and the vehicle itself, which must simultaneously satisfy:

[0229]

[0230] In the formula:

[0231] v n v rear These are the speeds of the vehicle itself and the vehicles following it in the target lane, respectively.

[0232] T represents the desired headway (1.5s);

[0233] d0 is the minimum spacing (2m);

[0234] b represents the comfortable deceleration (2.0 m / s²). 2 ).

[0235] If the front and rear distances are both greater than the safety threshold, then the lateral safety conditions are considered met.

[0236] (2) Acceleration gain threshold condition;

[0237]

[0238] The parameters are the same as above.

[0239] (3) Conflict detection and priority determination

[0240] When multiple vehicles simultaneously meet the lane-changing conditions, the system calculates the remaining time t for each vehicle to reach the merging zone. reach And lane change benefit ε:

[0241]

[0242] in:

[0243] t reach The remaining time for the vehicle to arrive at the merging area;

[0244] d merge : Coordinates of the starting point of the merging zone;

[0245] x n The current longitudinal position of the vehicle;

[0246] v n Current vehicle speed.

[0247] Prioritize satisfying t reach Smaller vehicles (i.e., those closer to the ramp entrance);

[0248] If t reach If they are the same, compare ε, and the one with higher profit will switch lanes first;

[0249] The remaining vehicles will temporarily postpone changing lanes and will be reassessed in the next moment.

[0250] Through the above mechanisms, the system can identify in advance main road vehicles that may conflict with ramp vehicles before entering the merging zone, and dynamically adjust their lane-changing behavior, enabling vehicles to complete cooperative avoidance in a safe and efficient manner.

[0251] 7) Control cycle and dynamic update

[0252] The control system executes the above process at a fixed cycle (e.g., 0.1 seconds), dynamically updating the status information, queue structure, and control commands of each vehicle, thereby achieving second-level cascaded control, real-time queue adjustment, and safe merging guidance.

[0253] The virtual platoon control strategy of this invention achieves coordinated management of vehicles on main and auxiliary roads in merging areas without relying on physical platooning. This strategy can effectively reduce merging point conflicts, lower the incidence of cutting in, and improve traffic efficiency. Combined with a lane-changing coordination mechanism, it can further enhance road capacity and ramp acceptance capacity, improve the overall safety and stability of expressway entrance operations, and has good engineering application value and promotion prospects.

Claims

1. A vehicle-road cooperative management and control system for typical urban expressway scenarios, characterized in that... include: 1) Multi-source traffic information collection system This system collects various traffic information, including road traffic volume, CAV penetration rate, lane position, vehicle speed, and distance to key locations, through the vehicle's perception module and vehicle-to-everything (V2X) module; specifically, it includes: Road traffic volume: Real-time traffic volume of each lane is obtained through the sensing module; CAV penetration rate: The ratio of autonomous vehicles (CAVs) to manually driven vehicles in the surrounding area; Vehicle position and lane: Determine the lane in which the vehicle is located and its position relative to other vehicles; Distance to key locations: including expressway exits and expressway entrances; Vehicle speed: Collects speed information of the vehicle itself and surrounding vehicles; Vehicle headway: Collects time distance information between this vehicle and surrounding vehicles; 2) Multi-source traffic data processing system Based on the collection of multi-source traffic information, this system performs fusion processing and feature calculation on raw data from video monitoring, geomagnetic coils, radar, on-board units (OBU), roadside units (RSU) and other sensors to form reliable traffic operation status indicators, providing support for vehicle-road cooperative management and control decisions. First, the system performs unified preprocessing on multi-source data; all data is synchronized according to a unified timestamp, and the vehicle positioning data is projected onto the actual road centerline coordinate system through a map matching algorithm to achieve spatiotemporal alignment; for data with noise or missing values, the system uses linear interpolation and median filtering for correction, and identifies abnormal samples through outlier detection methods; In the data calculation phase, the system performs real-time calculations on core traffic parameters; vehicle instantaneous speeds are obtained directly from positioning or radar speed measurement, and interval average speeds are calculated using weighted averages or harmonic averages. Where: v i (t): The instantaneous velocity of the i-th vehicle; N: The total number of vehicles detected within the current time t; Average speed of the road segment; Harmonized average speed of road segment; Traffic flow q is calculated from the number of vehicles passing through the detection section per unit time: Where: q(t): the vehicle flow rate passing through the detection section within the time interval Δt (unit: veh / h); Δt: Observation time interval; Traffic density k is derived from the ratio of the number of vehicles on a road segment to the length of the road segment: Wherein: the number of vehicles within the detection interval length L; L: Length of the road segment being inspected (unit: km); k(t): Vehicle density per unit length (veh / km); And satisfy the basic traffic relationship formula The vehicle's headway is obtained by detecting the time difference between the vehicle's departure and arrival at the designated point. h t =t i -t i-1 Wherein: the timestamp of the i-th vehicle passing through the detection section; t i-1 : The timestamp of the (i-1)th vehicle passing through this section; h t Time interval between the front ends of two adjacent vehicles (unit: seconds); Average headway h t Reflecting the compactness of traffic flow, it is an important auxiliary parameter for density estimation; occupancy is calculated as the ratio of the occupancy time of a detection unit to the observation period. Wherein: T occ The total time that the detection unit is occupied by vehicles during the observation period; T: Observation period length; Occupancy: The percentage of inspection points occupied reflects the density of traffic flow; The system introduces a joint algorithm of exponentially weighted moving average (EMA) and Kalman filtering; the EMA smoothing formula is: Wherein: the currently observed velocity value; Speed ​​estimate after smoothing; α: Smoothing coefficient, which controls the weighting of new and old observations (generally taken as 0.1 to 0.3); Δt: Sampling time step; When multiple sources of velocity or location information exist, a weighted fusion model is used: in: The variance of the measurement error of the i-th data source is used to represent its uncertainty. x i : The measurement value of the i-th data source; w i The weight of this data source is defined as the reciprocal of the measurement variance; During the real-time calculation phase, the system employs a multi-scale sliding window mechanism for rolling statistics and utilizes an incremental update algorithm to quickly calculate the average, variance, and distribution characteristics; a congestion index model based on speed and density is constructed. Where: v free Average velocity under free-flow conditions; k crit The critical density at which traffic flow reaches its maximum capacity; α: Adjustment coefficient, ranging from 0.5 to 1.5; When CI(t) is close to 1, it indicates severe congestion; when it is close to 0, it indicates smooth traffic. In addition, the system can predict short-term traffic conditions through machine learning models; the input features include traffic flow, density, average speed, headway, and occupancy rate, and the output is the speed and congestion level probability in the next 5 to 30 minutes, providing a forward-looking decision-making basis for the vehicle-road cooperative control system. Ultimately, this system outputs structured traffic status data, including average speed, flow rate, density, occupancy rate, average headway, congestion index and its confidence level, and updates it in real time through edge computing nodes to ensure that the data transmission latency is less than 200ms. 3) Vehicle-Road Cooperative Control System (1) Basic road segment scenario control strategy 1) Strategy Steps The strategy comprises the following three phases: ① Determining the Motive for Changing Lanes Whether the CAV has the motivation to switch lanes to the bus lane is the primary entry point for strategy execution; at this time, the system will comprehensively assess whether there are traffic restrictions in the current lane and whether the adjacent bus lane provides better traffic conditions. When the distance between this vehicle and the vehicle in front is insufficient to support the current acceleration increase or approach the maximum desired speed, it indicates that the driving environment is limited: In the formula: d i (t)——Distance between the front of the vehicle and the vehicle in front; —The current speed of the vehicle (m / s); Δt—Decision time interval (s), usually taken as 0.1 to 1 second; —Acceleration of the vehicle in front (m / s²) 2 ); v max —Maximum permissible speed of the vehicle (m / s); l — the length of the CAV vehicle; If there is a larger headway in front of the bus lane on the right, it indicates that there is room to improve traffic efficiency. In the formula: d i-rl (t)——The distance between the vehicle and the vehicle in front of the adjacent right-hand bus lane; Calculate whether the overall benefit ε of lane changing is greater than the set threshold E. ε is determined by the change in its own acceleration and the impact on surrounding vehicles. a SV —The current acceleration of the target vehicle; —The acceleration of the target vehicle after changing lanes; a FVi —The current acceleration of the vehicle behind the target vehicle in the original lane; —The acceleration of vehicles following the target vehicle in the original lane after the target vehicle changes lanes; a PFVi —The current acceleration of the vehicle behind in the target lane; —The acceleration of the vehicle following the target vehicle in the target lane after the target vehicle changes lanes; E(k) – Lane-changing threshold. The lane-changing threshold E is determined by calculating the ε distribution offline based on historical trajectories and simulation samples, and using ROC / grid search to determine the baseline value E0. It is dynamically adjusted using an adaptive function E(k) in combination with traffic density or congestion index. At the same time, hard safety constraints are applied to the maximum deceleration of the following vehicle and the minimum headway before the decision is made. Lane changing is only allowed when ε>E(k) and the safety constraints are met. ε—Lane-changing efficiency; p – Politeness coefficient, set to 0.8; Adaptive formula: Where: β is the calibration coefficient, β∈[0.2,1.0]; k: Traffic density; If the benefit ε exceeds the set threshold E(k), then lane changing is considered valuable and the lane changing motivation is valid; only when all three conditions above are met will the system enter the next stage of safety judgment. ② Determining the safety conditions for lane changing If a CAV has the motivation to change lanes, it is necessary to further determine whether the lane-changing behavior is safe and ensure that it will not affect vehicles behind in the bus lane. First, determine whether this will force vehicles behind in the target lane to brake suddenly: d i-rf (t)-l>v rf (t)·Δt In the formula: d i-rf (t)——Distance between the target vehicle and the vehicle behind it in the right-hand bus lane; v rf (t)——The current speed (m / s) of the vehicle behind in the target lane; This indicates that after changing lanes, vehicles in the target lane have enough time to react and do not need to decelerate suddenly. Next, determine whether the safety distance requirements are met: d i-rf (t)>d safe -l In the formula d safe —Minimum safe headway; This is the basic safety standard for changing lanes laterally, to avoid vehicles getting too close; If either of the above two conditions is not met, the lane-changing attempt will be rejected and the vehicle will continue to travel in the original lane. ③ Priority constraints for public transport vehicles Even if the motivation to change lanes is strong and the lane change itself is safe, the system still needs to further determine whether it interferes with the passage of buses in the bus lane; if the vehicle behind the target lane is a bus, the system introduces a "bus clearance distance" constraint to ensure that the CAV cannot forcibly insert itself too close to the front of the bus. d i-rf (t)-l>L C Among them, the minimum clearing distance L for buses C for: In the formula: —Bus speed; b max —The maximum deceleration of the bus is set to 2.5 m / s² by default. 2 ; t h —Following the lead time; s0 — Minimum static safety distance; Only when the clearance distance condition is met is a CAV allowed to change lanes in front of a bus to ensure that the bus is not interfered with. Once the above motivation assessment, safety verification, and bus priority check are all passed, the system can instruct the CAV to perform a lane-changing operation and briefly use the bus lane; if any step determines that the conditions are not met, the current lane-changing request will be canceled, and the CAV will continue to operate in the current lane. 2) Implementation Mechanism When a CAV is traveling on a regular road, if it meets both the motivation and safety conditions for lane changing and does not violate the minimum clearance distance requirement for buses, the system will control it to change lanes to the bus lane; otherwise, it will maintain its current lane operation.

2. The vehicle-road cooperative management and control system for typical urban expressway scenarios according to claim 1, characterized in that... include: To adapt to different stages of CAV penetration testing, three typical implementation strategies were designed: Option 1: No dedicated lane for CAVs will be set up. CAVs will use the bus lane and can share the road with conventional vehicles. Option 2: Set up a dedicated lane for CAVs, which can use both the dedicated CAV lane and the bus lane simultaneously, and cannot be mixed with traditional vehicles; Option 3: Set up a dedicated lane for CAVs, while allowing mixed traffic; When the CAV penetration rate is below 0.3, Scheme 1 is adopted as the control scheme; when the penetration rate is in the range of [0.3-0.5], Scheme 2 is adopted as the implementation scheme; when the penetration rate is in the range of [0.5,1], Scheme 3 is adopted as the operation scheme.

3. The vehicle-road cooperative management and control system for typical urban expressway scenarios according to claim 1, characterized in that... include: (2) Expressway Exit Control Strategy The system is divided into three functional zones: Variable Speed ​​Limit Control Zone (VSL) or Speed ​​Limit Segment 1: Located upstream of the bottleneck zone, it dynamically controls the speed and traffic density of vehicles entering the bottleneck, limiting the number of vehicles entering the bottleneck zone per unit time; Acceleration Transition Zone (VSL) or Speed ​​Limit Segment 2: Located between the speed limit zone and the bottleneck zone, it allows vehicles to gradually return to free-flow speed after passing through the VSL zone, ensuring vehicles enter the bottleneck zone in optimal condition; Bottleneck Zone: The expressway exit or divergence point. The system collects traffic density and speed information of each cell on the expressway in real time, and at the same time obtains the penetration rate γ of connected vehicle autonomous vehicles (CAV) in the current environment; Under different CAV penetration rates, the distance between the speed-limiting section and the exit is set as follows: when γ∈[0.1,0.9], in order to ensure that the speed-limiting effect is fully utilized, the distance from the end of the speed-limiting section to the exit is set to 900 meters; when γ>0.9, the length of the speed-limiting section is set to 700 meters. To describe the evolution of traffic conditions in real time, a cellular transport model (CTM) was used for modeling and calculation. In the CTM model, roads are divided into equal-length cells, and each cell contains n vehicles. i,j The principle of vehicle conservation is satisfied as it changes over time: n i,j =n i-1,j +and i-1,j-1 -and i,j-1 Ni=ρj·Δx In the formula: v f —Free-flow vehicle speed; n i,j — The number of vehicles in cell i at time j; y i,j —The traffic flow transmitted from cell i to cell i+1 at time j; N i,j —The capacity of cell i at time j; Q i,j —The passage capacity of cell i at time j; w—Propagation speed of the congestion reverse wave; ρ j —Theoretical congestion density (veh / km) of lane / segment at time j; Δx — Cell length (km); To achieve dynamic guidance, it is necessary to first determine whether there is a risk of congestion; to detect the congestion risk in bottleneck areas, the system determines in real time whether there is an oversaturation trend in the target area, i.e., the traffic density exceeds the critical density ρ. c2 When the system detects that traffic flow begins to decrease and the density continues to rise beyond the critical value ρ, c2 When a vehicle is identified as "potentially congested," the speed limit control module is triggered to intervene in the upstream vehicle speed. Speed ​​limits are dynamically calculated based on traffic flow parameters, with the goal of reducing traffic speeds to maintain the maximum bottleneck capacity Q. max The state; combined with the traffic three-phase flow model, the vehicle speed v i,j Expressed as density ρ i,j Piecewise functions: In the formula: ρ c1 —Critical density for free flow; ρ c2 —Critical congestion density; ρ c —Theoretical critical density, where traffic flow reaches maximum flow rate Q max The density; ρ j — Maximum congestion density; u w —Blocking wave speed; When congestion is imminent, the system determines the density ρ of the upstream control area. i Based on this, the controllable maximum flow rate Q is calculated. sl and speed limit value v sl : This rate limit is used to constrain the transmission capacity of the upstream cell: At the same time, adjustments were made to the downstream receiving capacity: In the formula: α i,j --The sending capability of cell i; β i+1,j —The receiving capacity of cell i+1; n i — The number of vehicles in cell i; v i,j-1 —The average velocity of the i-th cell at time j-1; This creates traffic flow constraints under speed limits, ultimately outputting the actual traffic volume: q i,j =min{a i,j ,b i+1,j } Once the speed limit policy is activated, vehicles in the guidance zone will proceed at the set speed limit v. sl Run, and gradually restore to free-flow velocity v in the downstream buffer. f .

4. The vehicle-road cooperative management and control system for typical urban expressway scenarios according to claim 1, characterized in that, The specific control strategy for expressway entrances is as follows: The implementation mechanism of this control strategy is as follows: 1) Vehicle information perception and arrival time prediction Using a vehicle-to-everything (V2X) sensing system, the speed v of vehicles on the main road and ramps, the distance d between their current location and the merging point are obtained in real time, and their estimated arrival time at the merging point is calculated. In the formula: —The estimated time, in seconds, for the i-th vehicle on the main road to arrive at the merging point; —The estimated arrival time of the i-th vehicle on the auxiliary road at the merging point, s; — The distance between vehicles on the main road and the merging point, in meters; —Distance between vehicles on the auxiliary road and the merging point, in meters; —The speed of vehicles on the main road at the current moment, in m / s; —The speed of vehicles on the main road at the current moment, in m / s; 2) Virtual fleet construction and sorting control All vehicles approaching the merging point are sorted according to their expected arrival time, generating a virtual convoy sequence arranged chronologically. Vehicles do not need to be physically convoyed, but they respond to the behavior of the vehicle in front in a queue manner in the control logic, realizing virtual coordination. 3) Longitudinal acceleration control model In the virtual convoy, each vehicle calculates its own acceleration based on the status of the vehicle in front (acceleration, speed, and distance). a n (t)=K a *a n-1 (t)+K d [Δx n (t)-E d (t)]+K v [v n-1 (t)-v n (t)]# In the formula: a n (t)——The acceleration of vehicle n at time t, m / s² 2 ; a n-1 (t)——The acceleration of the leading vehicle n-1 of vehicle n at time t; Δx n (t)——The distance between vehicle n and its leading vehicle n-1 at time t; v n (t)——The speed of vehicle n at time t, m / s; K a ,K d ,K v —The control coefficients of the model are calibrated using the measured vehicle trajectories in the merging zone and the particle swarm optimization algorithm with the speed error as the minimum objective. E d (t)——The expected following distance between the target vehicle n and the leading vehicle n-1; 4) Calculation of expected spacing (refer to IDM model) Where the expected spacing E d (t) Dynamic adjustments are made based on the IDM (Intelligent Driver Model): Where: T—the expected headway in the IDM model, taken as 1.5s; d — the following distance in congested traffic, in meters; a——Maximum vehicle acceleration, m / s² 2 ; b — Vehicle comfort deceleration, m / s 2 ; in, a, b, and d can all be calibrated using particle swarm optimization algorithm based on measured data; simultaneously, a basic distance threshold is set according to the CAV permeability for adjusting the controller initialization parameters. When the permeability is less than 0.1, the distance threshold is set to 350m; when the permeability is in the range of [0.1, 0.3], the distance threshold is 150m; when the permeability is greater than 0.3, the distance threshold is set to 450m. 5) Freewheel acceleration control When the target vehicle is the lead vehicle of the convoy or there are no vehicles within the target vehicle's influence range, the vehicle is in a free-driving state and is considered a free-moving vehicle not participating in the virtual convoy dispatch. Its acceleration is calculated based on the difference between the current speed and the road speed limit. In the formula: —Acceleration outside the convoy; —This is the speed limit for the road; acceleration is constrained by the maximum / minimum limits and the rate of change. The following constraints are set to ensure smooth and comfortable vehicle control: In the formula: a differ —Maximum rate of change of acceleration; 6) Coordinated lane-changing control mechanism If lane changing is permitted in the outer lane of the main road, determine if there are any vehicles approaching the ramp nearby. If so, proceed with the lane change determination process. Lane changes must meet the following conditions: (1) To determine the spatial and speed relationships between the vehicles in front and behind the target lane and the vehicle itself, the following conditions must be met simultaneously: In the formula: v n v rear These are the speeds of the vehicle itself and the vehicles following it in the target lane, respectively. T represents the desired headway; d0 is the minimum spacing; b represents comfortable deceleration; If the front and rear distances are both greater than the safety threshold, then the lateral safety conditions are considered to be met. (2) Acceleration gain threshold condition; (3) Conflict detection and priority determination When multiple vehicles are eligible to change lanes simultaneously, calculate the remaining time t for each vehicle to reach the merging zone. reach And lane change benefit ε: in: t reach The remaining time for the vehicle to arrive at the merging area; d merge : Coordinates of the starting point of the merging zone; x n The current longitudinal position of the vehicle; v n Current vehicle speed; ●Prioritize satisfying t reach Smaller vehicles, meaning those closer to the ramp entrance; ●If t reach If they are the same, compare ε, and the one with higher profit will switch lanes first; ●The remaining vehicles will postpone changing lanes and will be reassessed in the next moment; 7) Control cycle and dynamic update The above process is executed at fixed intervals to dynamically update the status information, queue structure and control commands of each vehicle, thereby achieving second-level cascaded control, real-time queue adjustment and safe merging guidance.

Images