Design of Real-Time Dynamic Reversible Lane in Intelligent Cooperative Vehicle Infrastructure System

.e rapidly growing traffic demand and the slowly increasing traffic supply have produced an mounting contradiction, which is mainly manifested in cities as road congestion and unbalanced bidirectional traffic flow. Most of the reversible lanes are implemented on fixed sections and fixed times and are mainly guided by ground markings, road signs, railings, and traffic police officer. It requires a lot of human and material costs. And, the control effect is lagging and inaccurate. Aiming at these problems, a real-time dynamic reversible lane scheme in the Intelligent Cooperative Vehicle Infrastructure System (CVIS) was proposed. Traffic information was collected in real time through the CVIS, and a reversible lane scheme was established based on the realtime service level V/C and BRP functions. A lane change control model was applied to determine the number of lanes and the timing of lane changes. .en, the reversible lanes were managed in real time through intelligent road stud lights and light curtain walls. Buffer sections and no-entry sections were set to ensure reversible lanes operating safely and efficiently. VISSIM simulation was used for case analysis, and the results showed that compared with the traditional time-controlled reversible lane scheme, the real-time dynamic reversible lane scheme could reduce the average vehicle delay by 27.4% and decrease the vehicle VOC, CO and NOX emissions by 13.5%.


Introduction
e national highway mileage reached 4,486,500 km, an increase of 73,100 km and 1.53% over the previous year. e rapidly increasing transportation demand and the slowly increasing transportation supply have produced increasing contradictions between supply and demand [1,2]. Some urban roads have serious congestion and unbalanced bidirectional traffic flow. In 2017, the number of connected vehicles in China reached 17.8 million. With the integration and innovation of big data, cloud computing, and other technologies, the penetration rate of connected vehicles will continue to accelerate in the future [3][4][5][6].
Intelligent road stud light is currently mainly used for underground parking lot vehicle indication, detection, and lane keeping reminder and has not been used as a signal indicating device on the road. In the proposed reversible lane scheme, intelligent solar road stud lights were used, which were controlled in real time under connected vehicle environment, to achieve the same indication function as road traffic lights.
A huge amount of study has been carried out on reversible lanes. Brian Wolshon and Laurence Lambert studied the application status, control and management measures, and evaluation methods of reversible lanes, and found that although there was no unified planning and well-established standards, the introducing of reversible lanes still achieved the expected goals and were accepted by the public [7]; Matthew and Peter et al. established a linear programming model to calculate the optimal configuration of variable lanes and used the information collected by traffic sensors to determine the direction of reversible lanes on the road. e direction was dynamically changed, and the results showed that the road capacity can be effectively improved by introducing the dynamic reversible lane scheme [8]; Golub et al. put forward suggestions from the perspective of economic benefits and traffic benefits from the design of reversible lanes, marking, and management [9]; Yu and Tian studied the setting conditions of the reversible lanes and established a two-level planning model based on the road network. e upper layer of the model integrates and optimizes the lane allocation, and the lower layer is a network equilibrium model for predicting the driver's response to lane allocation [10]; Cui and Liu investigated the possibility and necessity of introducing reversible lane schemes based on field survey and put forward solutions to manufacturingrelated hidden traffic hazards of the reversible lane schemes [11]; Sun established an optimization model for the allocation of reversible lanes, dividing 24 hours of a day into several continuous time sections, which were used to allocate the number of tidal lanes in different sections according to the traffic flow characteristics of different time sections [12][13][14]; Dai et al. analyzed the current status of reversible lane schemes in foreign countries, proposed three steps for setting reversible lane schemes, and verified the effect of introducing reversible lane schemes through simulation [15,16].
Reversible lanes have been studied by scholars all over the world, and they have been widely used in urban roads, but most of them are fixed time and fixed lanes, which require manual operation depending on the traffic situation. At present, no intelligent lane-combined technology with road studs has been achieved for the dynamic control of reversible lanes [17][18][19][20]. Where road information in CVIS was collected, the control model was used to allocate reversible lanes and intelligent road studs were used to guide traffic [21,22]. Compared with the existing reversible lane control methods, the biggest advantage is that reversible lane allocation can be dynamically changed according to the traffic volume of the road, no manual operation is needed, and energy conservation and environmental protection are achieved, thereby achieving optimal use of road resources. For the study of travel time on the road, the BPR model is most frequently used, and the road resistance function is as follows:

Design Principles and Schemes of Real-Time
where t is the travel time, t 0 is the travel time for free flow, V is the traffic volume, C is the traffic capacity, and α and β are parameters equal to 0.15 and 4, respectively. Based on the minimum impedance, the path resistance function model is as follows: where V 1 and m refer to the traffic volume and number of lanes in one direction, V 2 and n refer to the traffic volume and number of lanes in the other direction, C represents the traffic capacity, I is the number of reversible lanes, K o is the delay coefficient ranging from 1 to 1.2, L is the length of reversible lane, V s is the design speed of the road, r is a parameter, if the nonmotorized lane and motor lane are separated, r equals to, otherwise r equals to 0.8, and η equals to 1 if the width of lanes is 3.5 meters.

Determination of the Two-Way Traffic Conditions after the Lane Change Based on the Adjusted Service Level.
then the congestion service level can be calculated as follows: e adjusted service level of the uncongested segment is not lower than the service level of the congested segment, then we have the critical state:

Establishment of Dynamic Control Model of Lane
Change. e lane change dynamic control model obtains road traffic information in CVIS and evaluates the road traffic flow conditions based on the service level of the road. en, the corresponding control module is used to solve the optimal switching scheme and determine whether the scheme meets the preset conditions. If it is satisfied, the optimal solution is executed; if it is not satisfied, the existing solution is maintained. e flow chart of dynamic control model is shown in Figure 1.

Road Signaling.
e intelligent road stud is used as a road signal indicating device to dynamically change the lane allocation and guide the vehicle throughout the road. e intelligent road stud light uses lithium battery and solar battery to power. It can work continuously for 36 days under the supply of lithium battery power. It uses 6 2000 mcd LED lights to meet the signal indication requirements under different conditions. rough LoRa's new generation of IoT communication technology, point-to-point and point-tocenter ultra-long-distance communication is realized at the same time.
e road stud light adopts cast aluminum housing, with IP68 waterproof feature, and the working temperature range is −40°to 85°; it can withstand 20 tons of static pressure, which can meet the requirements of different scenarios and different roads.
e Intelligent road stud has three light colors, red, yellow, and green, and can be displayed in flashing. Red flashing means reminding to leave the lane and prohibiting traffic; green flashing means the lane is about to open and allowing traffic; the yellow light indicates the buffer section, which means to drive away as soon as possible and warn drivers that there is a no-entry section ahead. ere are six patterns for road stud lights, as shown in Figure 2. e road management agency can change the color of the lights through the remote communication function to achieve the indication of different road traffic conditions. e remote instructions of the road stud lights are encrypted by RSA to achieve the security of traffic infrastructure control.
In combination with the existing laser projection technology, a virtual wall curtain idea formed by light projection was proposed. A light curtain wall was provided with a small projection device in a stud light to project light onto a reversible lane to form a virtual wall curtain; the brightness and color of the wall curtain are consistent with the stud lights. e light curtain wall can cooperate with road stud lights to provide instructions and reminders to ensure that vehicles can safely change lanes and drive. At present, the light curtain wall technology does not meet the requirements for practical application to roads. According to technical research and experimental analysis, light curtain wall technology can be applied to roads as a traffic signal indicator in the future.

Conflict Zones.
A conflict zone of the reversible lane is divided into a starting conflict zone and an ending conflict zone. e objective of setting conflict zones is to make the vehicles in both directions change lanes smoothly and avoid conflicts. e starting conflict zone consists of a no-entry section.
e ending conflict zone consists of two buffer sections and a no-entry section. e vehicle can complete the lane change or stay for a while waiting for a lane change in the buffer section. During the operation of the variable lane, the traffic is allowed to enter the no-entry section, forming a spatial separation zone for traffic flow to avoid oncoming conflicts.
When setting a conflict zone on a city road, the location should be as far away as possible from intersections, bus stops, and other places where traffic flow will be disturbed. If the reversible lane is at the beginning or end of the intersection, it is not necessary to set a conflict zone. e intersection is an effective area to avoid oncoming conflict. erefore, when the distance between the start or end of the reversible lane of the road section and the center of the intersection is within 300 m and the intersection satisfies the conditions of introducing the reversible lane, the reversible lane should be directly extended to the intersection. When the intersection does not meet the requirements for setting a reversible lane, the reversible lane should be shortened to ensure that the distance from the conflict zone to the intersection is more than 300 m. Because the reversible lane extending to the intersection does not need to set the conflict zone, we will only discuss the corresponding setting parameters for setting the conflict zone far away from the intersections. e starting conflict zone of the reversible lane is relatively simple to set.
ere is no direct oncoming conflict for two-way traffics in starting conflict zone, so we only need to avoid conflicts happen in the lane change process. e distance of the starting conflict zone depends on the vehicle's safe following distance and road stud laying interval which is generally 5-10 m.
In order to alleviate traffic congestion of eastbound traffic, vehicles in the eastbound traffic can borrow a part of the middle lane of the westbound traffic. Figure 3 shows the starting conflict area of the reversible lane. Vehicles in the eastbound traffic can choose to merge left to enter the reversible lane in the starting conflict area, and vehicles in the westbound traffic can merge left to get back to its original lane after the starting conflict zone. In Figure 4, the ending conflict area is the end area of the lane occupancy for vehicles in the eastbound traffic and the starting area of lane occupancy by the eastbound traffic for vehicles in the westbound traffic. erefore, vehicles in the eastbound traffic must drive back to the original lane, and vehicles in the westbound traffic must drive into other lanes in the eastbound direction to avoid oncoming conflicts.

Design of No-Entry Section.
e no-entry section is the isolated area of the vehicle during the operation of the reversible lane. It is set at the beginning and end of the reversible lane where the two-way traffic conflicts. In order to ensure the safety of reversible lane operation, the longer the distance of the no-entry section is, the better. However, in the perspective of utilization of road resources, the shorter no-entry section is preferred. erefore, it is necessary to make the distance of the no-entry section as small as possible while ensuring the safety operation of the traffic. e length of the starting no-entry zone is the length of the starting conflict zone, and it is the safety following distance between vehicles. Generally, it is 5-10 m. e parameters of the ending on-entry zone are discussed below.
If a driver in one direction enters the end of the no-entry section, then they must be able to make a complete stop before the end of the no-entry section; if two drivers from opposite directions have entered the no-entry section, they must have sufficient space to stop cars to avoid a collision. erefore, the length of the no-entry section must meet the sight distance of the vehicle and ensure the safe braking distance. e shortest length is determined by the reaction distance (l r1 , l r2 ), breaking distance (l b1 , l b2 ), and safety following distance (l 0 ) of two opposite traveling drivers as shown in Figure 5.
Assume the vehicle braking performance and response time of drivers are the same, then where t is the reaction time ranging from 1 to 2.5 seconds, V 0 is the vehicle speed before breaking, g is the gravity, and φ is the coefficient of adhesion of wheels.

Design of Buffer
where θ is the cautious coefficient of drivers ranging from 0 to 1, the larger the number, the more cautious the driver is, α 1 and α 2 are parameters depends on traffic condition and number of interval lanes between the current lane and the target lane, ρ is the traffic density, ρ jam is the traffic congestion density, and n is the number of reversible lanes. Vehicle braking deceleration is related to the driver's judgment of the braking distance and is also related to the driver's driving habits, road conditions, and vehicle speed on the road. From the perspective of the braking capacity of a car, the maximum deceleration of the car during emergency braking is generally 7.5～8m/s 2 . For a normal braking, the average deceleration of the car should be 3～4m/s 2 . However, in real situation, except for emergency situations, the braking deceleration should normally be 1.5～2.5m/s 2 .
us, a value of 2 m/s 2 of deceleration is used in this paper. If the driver takes breaks at the beginning of the buffer section, the braking distance of the vehicle is calculated based on the vehicle's speed, braking performance, and road friction coefficient. e driver's reaction time from making a braking decision to force the brake is called the driver's reaction time t 0 . e time interval from when the brake is forced to when the brake is in effect is called the driver's operating time, which is represented by t 1 . Assume total the reaction time t � t 0 + t 1 , and define t 2 is the time interval from when brake is in effect to when the car completely stops.
e total braking distance of the vehicle could be calculated as follows:

East West
No-entry section Figure 3: Configuration of starting conflict zone.

Buffer section Buffer section East West
No-entry section Figure 4: Configuration of ending conflict zone.

Journal of Advanced Transportation
Let t � 2.5 s, and t 2 is negligible so the second term drops out. e reaction distance S 1 and breaking distance S 2 can be calculated as follows: e length of a buffer section is

Data Analysis.
A survey was conducted to collect traffic flow data from 18 : 00 to 19 : 00 of an 8-lane highway in Huai'an city, as shown in Table 2. e microscopic simulation software VISSIM and its secondary development are used to simulate and evaluate the vehicle running effect of the road in timing control mode and dynamic control mode. e timing control mode is controlled by the traffic police officer according to the traffic conditions of the road. It was turned on at 18 : 40 in the simulation. e dynamic control mode is controlled using the lane change control model of the proposed scheme.
rough the analysis of simulation data, with the continuous increase of road traffic, the average delay of vehicles in the dynamic control mode is relatively slow (Figure 7), the overall average delay of vehicles is reduced by 27.4%, and the vehicle VOC, CO, and NOX emissions are reduced by 13.5%. erefore, the dynamic control mode has better effects in improving traffic efficiency and environmental protection. e volume reduction of 13.5% effectively improves the imbalance of traffic flow and alleviates road congestion, improves road traffic efficiency, reduces energy consumption, and provides a new solution for the promotion of new energy-saving and environmentally friendly reversible lane designs. With the continuous development and application of urban traffic guidance system and vehicle-road collaboration technology, combining it with this solution can more effectively optimize road resource allocation and improve vehicle traffic efficiency. e proposed dynamic signal control mode with intelligent road stud lights can guide urban roads in different road conditions. e combination with bus priority can effectively improve the utilization of existing bus lanes and other facilities. Adhering to the concept of energy saving and environmental protection, optimizing road resources, and innovating signal control modes, a new solution for the optimization of smart city transportation systems was proposed.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
Lina Mao conceived and designed the paper; Wenquan Li conducted the model and simulation; Lina Mao and Pengsen Hu wrote the paper; Guiliang Zhou designed the real-time dynamic reversible laneschemes; Huiting Zhang analyzed the simulation results, Jin Dai collected traffic data.  Journal of Advanced Transportation 7