CN115206115A - Road network congestion area control method based on multi-source data edge calculation in intelligent networking environment - Google Patents

Abstract

The invention discloses a road network congestion area control method based on multi-source data edge calculation under an intelligent networking environment, which comprises the following steps: 1. establishing a road network traffic flow database, 2, uploading multi-source data, 3, preprocessing the data by an edge computing center, 4, judging whether a congestion area exists by the data center, 5, formulating congestion area management and control measures, and 6, implementing management and control. The method can effectively utilize the advantages of multi-source traffic data and the network connection automatic driving vehicle, and accurately identify the congestion area which changes in real time in the urban road network, thereby controlling the traffic flow which is close to the boundary of the congestion area and enters the direction of the congestion area, and more efficiently solving the problem of managing and controlling the congestion area of the urban road network.

Description

Management and control of road network congestion areas based on multi-source data edge computing in intelligent network environment method

technical field

The invention belongs to the field of intelligent transportation, and in particular relates to a road network congestion area management and control method based on multi-source data edge computing in an intelligent network connection environment.

Background technique

In the process of urbanization and motorization, the phenomenon of urban traffic congestion is becoming more and more frequent, especially in the road network in many urban centers. Real-time detection of the traffic flow of the road network in the city center area, and control of the traffic flow from the periphery of the city center area into the city center area according to the detection results will help ease the congestion in the city center area.

In recent years, with the rapid development of 5G communication technology and connected autonomous driving technology, Connected and Automated Vehicle (CAV) has gradually broken away from theory and become a reality. It can be foreseen that in the near future, the traffic flow on the road will enter the stage of CAV and human-driven vehicle (Human-driven Vehicle, referred to as HV) mixed. In the environment of mixed traffic flow (traffic flow in which connected autonomous vehicles and human-driven vehicles are mixed), using the advantages of connected autonomous vehicles helps to obtain more accurate vehicle driving data and determine the real-time traffic status of the road network. Accelerates the rate at which congestion dissipates in high-traffic-density areas in urban centers and improves traffic control. At the same time, the application of traffic data acquisition sensors and vehicle-road coordination equipment has been popularized, providing a variety of judgment basis for the identification of road network traffic status. However, the transmission of massive traffic data with various types and complex structures also puts forward extremely high requirements on the system communication bandwidth and processing capacity, which brings huge processing pressure to the data center. As a distributed computing architecture, edge computing decomposes large-scale services that were originally processed by the data center, and collects and preprocesses data at the end close to the data source. There is no need to upload a large amount of original data, and the processing delay is smaller and the security is The reliability is higher, and it is suitable for the scenario of judging the road network status based on a variety of complex multi-source data.

In solving regional urban congestion management and control methods, the traditional method divides the entire urban road network into several fixed sub-road networks in advance, that is, sub-districts, and judges whether the sub-districts are congested according to the changes in the traffic density in the sub-road network, and then changes the boundary location. The signal timing is limited to limit the traffic flow into the congested sub-area. However, the road congestion area actually changes in real time. The congestion area judgment based on fixed sub-areas has poor flexibility and cannot be applied to the dynamic change of the congestion area, which affects the effect of the control on the entrance boundary of the congestion area. In addition, the existing boundary control method reduces the traffic volume entering the high traffic density area by shortening the phase duration of the intersection signal from the peripheral area of the congested area into the direction of the congested area. Due to the fixed boundary, a large number of vehicles line up in the upstream section of the congestion area or even overflow the intersection, affecting traffic safety and system stability. Finally, the existing control strategies cannot control the non-signal controlled intersections at the boundary of the congested area, which affects the control effect.

SUMMARY OF THE INVENTION

In order to solve the above-mentioned shortcomings of the prior art, the present invention proposes a road network congestion area management and control method based on multi-source data edge computing in an intelligent networked environment, in order to effectively utilize multi-source traffic data and networked automatic driving Using the advantages of the vehicle itself, it can accurately identify the congested areas in the urban road network that change in real time, so that it can control the traffic flow in the direction of the congested area near the border of the congested area, so as to solve the management and control problems of the congested area of the urban road network more efficiently.

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

The invention is a road network congestion area control method based on multi-source data edge computing in an intelligent networked environment, which is applied to m edge computing centers, n networked automatic driving vehicles, m road section units and 1 data center. In the formed urban road network in which the connected automatic driving vehicle and the artificial driving vehicle coexist, the feature is that the method for managing and controlling the congestion area of the road network includes the following steps:

Step 1: Consider the road in each flow direction between every two adjacent intersections as a road segment unit, and obtain the history of the l-th road segment unit by considering the lane number n _l and length information L _l of any l-th road segment unit Traffic flow data and corresponding weather data, thereby establishing a traffic flow database including each road segment unit; l=1,2,3,...,m;

Set the time interval between two adjacent control cycles as T; the current number of control cycles is k, and initialize k=1;

Step 2: Under the k-th control cycle, the i-th network-connected autonomous vehicle sends its own driving speed v _i (k) and position information to the edge computing center on the roadside of the road unit where it is located, and the The traffic flow data collected by the sensor is also transmitted to the edge computing center on the roadside of the road section unit; i=1, 2, 3, ..., n; n represents the total number of connected autonomous vehicles;

Step 3: Under the kth control cycle, the lth edge computing center processes the traffic flow data to obtain the total number of vehicles N _l (k) in the lth road segment unit;

According to the location information of each networked autonomous vehicle, the lth edge computing center deletes the information of the networked autonomous vehicles in the non-lth road segment unit, and counts the total number of networked automatic driving vehicles on the lth road segment unit.

Calculate the average traffic speed of the lth road segment unit based on the driving speed of each connected autonomous vehicle on the same road segment

从而计算第l个路段单元的流量Q_l(k)＝n_lK_l(k)V_l(k)；Obtain the traffic density of the l-th road segment unit by the total number of vehicles N _l (k) in the l-th road segment unit

Thereby, the flow Q _l (k)=n _l K _l (k) V _l (k) of the l-th road segment unit is calculated;

The lth edge computing center is networked to obtain the current weather conditions of the urban road network, and inquires the free flow speed V _l ^f of the lth road segment unit under the current weather conditions in the traffic flow database to determine the traffic state of the lth road segment unit :

When V _l (k)>V _l ^f ×δ%, it means that the l-th road segment unit is in a smooth state;

When V _l (k)≤V _l ^f ×δ%, it means that the l-th road segment unit is in a congested state;

车辆总数N_l(k)、流量Q_l(k)以及交通状态的判断结果所构成的交通状态信息上传至数据中心；其中，δ表示所设定的阈值；The lth edge computing center calculates the total number of connected autonomous vehicles in the lth road segment unit.

The traffic state information composed of the total number of vehicles N _l (k), the flow rate Q _l (k) and the judgment result of the traffic state is uploaded to the data center; wherein, δ represents the set threshold;

Step 4: The data center uses the density-based spatial clustering algorithm to cluster the road section units in the congested state according to the traffic state information of each road section unit in the urban road network, and obtains several clustering results, each clustering The result is a cluster composed of several congested road section units; the smallest polygon of the area occupied by each clustering result in the urban road network is recorded as the area;

When the total length S _A of a certain clustering result in the occupied area A is greater than the threshold S _cr , the urban road network is considered to be in a congested state, and the corresponding area is identified as a congested area; otherwise, it indicates that the corresponding area is congested. The area is a clear area, and there is no need to implement boundary control or traffic guidance measures on the road network. After the k-th control cycle ends, assign k+1 to k, and then return to step 2; where, S _cr =S ₀ ×Δ%, S ₀ is the total length of the road unit in the urban road network; Δ represents the set threshold;

Step 5: Denote the road section unit in the congested area as set Θ, denote the road section unit in the direction of exiting the congested area near the border of the congested area as set Ο, and denote the road section unit in the direction of entering the congested area near the border of the congested area as set Ω;

及拥堵区域内车辆总数

则拥堵区域内网联自动驾驶车辆渗透率为p(k)＝N^CAV(k)/N(k)；Calculate the total number of connected autonomous vehicles in a congested area

and the total number of vehicles in the congested area

Then the penetration rate of connected autonomous vehicles in the congested area is p(k)=N ^CAV (k)/N(k);

Based on the traffic data point set in the traffic flow database consisting of the total number of vehicles in the congested area and the travel completion rate of the congested area under the total number of vehicles in the congested area, generate the road network of the congested area when the penetration rate of connected autonomous vehicles is p(k). The macroscopic basic map, so as to fit the macroscopic basic map function G _A (x) of the congested area and generate a 95% prediction band, so as to obtain the total number of vehicles in the congested area corresponding to the maximum value G _max (k) of G _A (x) The critical value of N _cr (k); where the travel completion rate of a certain area represents the number of vehicles leaving the area per unit time; x represents the total number of vehicles in the congested area; the value of G _A (x) represents the vehicles in the congested area The trip completion rate when the total number is x; the 95% prediction band is the region to which 95% of the data points belong at the time of fitting;

Step 6: Judging the validity of boundary control according to the total number of vehicles in the current congestion area and the travel completion rate, if the boundary control is valid, go to Step 7; if the boundary control is invalid, go to Step 9;

l∈Ω；对于拥堵区边界处有信号灯控制的交叉口，将进入拥堵区方向的相位时长由t₀调整为u(k)t₀，以降低驶入拥堵区的交通量，其中，

表示未进行边界控制时第l个路段单元的限速值，l∈Ω；Step 7: Taking the travel completion rate of the congested area reaching the maximum value G _max (k) as the goal, implement model predictive control on the boundary of the congested area, and calculate the control rate u(k) of the congested area boundary in the kth control period; Adjust the speed limit value of the lth road segment unit in the direction of entering the congestion area near the boundary of the congestion area as

l∈Ω; for the intersection with signal light control at the boundary of the congestion area, the phase duration in the direction of entering the congestion area is adjusted from t ₀ to u(k)t ₀ to reduce the amount of traffic entering the congestion area, where,

Represents the speed limit value of the l-th road segment unit without boundary control, l∈Ω;

拥堵区域以及管控区域，第l个边缘计算中心将调整后的限速值

传输给第l个路段单元内的网联自动驾驶车辆，使得网联自动驾驶车辆减速行驶，l∈Ω；其中，管控区域表示拥堵区域及临近拥堵区域边界的驶入拥堵区方向的路段单元所占面积的最小多边形；Step 8: The data center transmits the adjusted speed limit value of the lth road segment unit to the lth edge computing center

Congested area and control area, the speed limit value adjusted by the lth edge computing center

It is transmitted to the networked autonomous vehicle in the l-th road segment unit, so that the networked autonomous driving vehicle decelerates, l∈Ω; where, the control area represents the congestion area and the road segment unit near the boundary of the congestion area that enters the congestion area. The smallest polygon that occupies an area;

Step 9: At the upstream intersection of the congested area, continuously inform the downstream area of congestion through the roadside variable information sign within the time interval T, and provide the speed limit information and suggested driving route of the surrounding road units to achieve the purpose of traffic guidance ; After the k-th control cycle ends, assign k+1 to k, and return to step 2.

The feature of the method for managing and controlling the road network congested area according to the present invention is also that the judgment of the effectiveness of the boundary control in the step 6 includes:

Step 6.1: According to the total number of vehicles N(k) in the congested area and the macro basic graph fitting function G _A (x), obtain the theoretical value of the trip completion rate G _A (N(k)) in the congested area;

与理论值G_A(N(k))之间的残差ΔG_A；其中，

Step 6.2: Calculate the actual value of the trip completion rate for the congested area

Residual _ΔGA from the theoretical value GA (N( _k )); where,

构成的交通数据点是否处于95％预测带内，若处于，则边界控制有效；否则，表示边界控制失效。Step 6.3: Determine the total number of vehicles N(k) in the current congestion area and the actual value of the trip completion rate according to the residual ΔG _A

Whether the constituted traffic data points are within the 95% prediction band, if so, the boundary control is valid; otherwise, the boundary control is invalid.

The congested area boundary model predictive control in step 7 includes:

Step 7.1: According to the total number of vehicles N(k) in the congested area and the critical value _Ncr (k) of the total number of vehicles in the congested area in the k-th control cycle, the state variable of the system is obtained as X(k)=N(k)- N _cr (k), let the control variable of the system be U(k)=u(k), and the state equation of the system is X(k+1)=[-β(k)G′ _A (N(k))+ 1]X(k)+[q _in (k)Tq _out (k)T]U(k), the output of the state equation is the total number of vehicles N(k+1) and the number of vehicles in the congested area of the k+1th control cycle. The difference between the total critical value N _cr (k+1); among them, β(k) indicates that in the k-th control cycle, the travel destination of the vehicles in the congested area is the vehicle in the congested area, which accounts for 5% of the total travel vehicles. ratio; q _in (k) represents the traffic flow rate into the congested area in the k-th control cycle; q _out (k) represents the traffic flow rate out of the congested area in the k-th control cycle; G′ _A (N(k)) represents The value of the derivative function G′ _A (x) of the macroscopic basic graph fitting function G _A (x) when x=N(k);

Step 7.2: Obtain the state variables of the system in the next n _p control cycles according to the state equation of the system, and calculate the control rate in the next n _p control cycles, n _c <n _p , use the matrix X=[X(k), X(k+1),…X(k+n _c ),…X(k+n _p )] represents the state variables of the system in the current control cycle and the next n _p control cycles, the current control cycle and the future n _p The control variable matrix of the system in the control period is U=[U(k), U(k+1),...U(k+n _c ),...U(k+n _p )], when the predicted k'th When the control period k+n _c <k′≤k+n _p , let U(k′)=U(k+n _c ); where X(k+n _c ) represents the k+n _c th control period. State variable; U(k+n _c ) represents the control variable of the k+n _c -th control cycle;

Step 7.3: The output of the system is the state variable of the congested area in the current control cycle and the next n _p control cycles;

When the output of the system approaches 0, the travel completion rate is maximized, and the objective function Z solved in the control period is: Z=minX ^T ·H·X; where H is the weight matrix; X ^T represents the rotation of X set;

Step 7.4: Obtain the control variables in each control cycle by solving the objective function Z, and select the first element value U(k) of the control variable matrix U as the boundary control rate of the kth control cycle.

Compared with the prior art, the present invention has the following beneficial effects:

(1) The present invention uses edge computing technology to preprocess multi-source data, and then uploads the judgment result to the data center after judging the traffic state of the road section, which avoids the heavy processing tasks of the data center, improves the efficiency of data processing, and The preprocessed data is less, the time required for data upload is shorter, the bandwidth requirements are lower, and the security and reliability are higher; in addition, the information uploaded to the data center only involves traffic information at the road level, and the specific individual vehicle Data does not need to be uploaded, which can effectively protect personal privacy.

(2) The present invention realizes the identification of non-fixed boundary congestion areas by clustering the road section units in the congestion state, has higher flexibility, and is more suitable for the real-time change characteristics of the actual road congestion areas, which can effectively improve the access to the congestion areas. The effect of boundary control.

(3) The present invention forms a mobile bottleneck by decelerating the network-connected automatic driving vehicle, achieves the speed limit effect, restricts the traffic flow entering the congested area, and solves the problem that the existing boundary control method cannot control the borderless control intersection in the congested area. The boundary control method based on speed limit can effectively solve the problem of vehicle overflow caused by the existing boundary control method when there are too many queuing vehicles, and the safety benefits are significant; The variable speed limit measures implemented through roadside variable speed limit information signs face the problem of low compliance rate of manually driven vehicles, which can effectively improve the implementation effect of variable speed limit.

Description of drawings

Fig. 1 is the system flow chart of the present invention;

2 is a schematic diagram of a local congestion of a city road network provided by the present invention;

3 is a schematic diagram of a macroscopic basic diagram and a 95% prediction band provided by the present invention;

FIG. 4 is a schematic diagram of the congestion zone management and control measures of the present invention;

FIG. 5 is a schematic diagram of the present invention implementing speed limit through a networked autonomous vehicle.

Detailed ways

In this embodiment, a method for managing and controlling road network congestion areas based on multi-source data edge computing in an intelligent networked environment is based on the advantages of multi-source traffic data, combined with the data processing capability of an edge computing center, to realize the control of road section units. Traffic status judgment. At the same time, according to the spatial distribution of the congested road sections and the connection between the units of the congested road sections, the data center can quickly and accurately identify the congested areas of the road network and formulate a congestion evacuation strategy. Finally, using the characteristics of the highly obedience of the connected self-driving vehicle to the command, the networked self-driving vehicle decelerates to form a mobile bottleneck, reduces the speed of the traffic flow on the road section, and achieves the effect of restricting the traffic flow into the congested area, and at the same time solves the existing congestion area boundary control method It is impossible to adjust the problem of traffic control at intersections without signals, and can effectively prevent the overflow of queues in road sections. Specifically, as shown in Figure 1, the method includes the following steps:

Step 1: Consider the road in each flow direction between every two adjacent intersections as a road segment unit, and obtain the history of the l-th road segment unit by considering the lane number n _l and length information L _l of any l-th road segment unit Traffic flow data and corresponding weather data, thereby establishing a traffic flow database including each road segment unit; l=1,2,3,...,m;

Set the time interval between two adjacent control cycles as T; the current number of control cycles is k, and initialize k=1;

Step 2: Under the k-th control cycle, the i-th network-connected autonomous vehicle sends its own driving speed v _i (k) and position information to the edge computing center on the roadside of the road unit where it is located, and the The traffic flow data collected by the sensor is also transmitted to the edge computing center on the roadside of the road section unit; i=1, 2, 3, ..., n; n represents the total number of connected autonomous vehicles;

Step 3: Under the kth control cycle, the lth edge computing center processes the traffic flow data to obtain the total number of vehicles N _l (k) in the lth road segment unit;

According to the location information of each networked autonomous vehicle, the lth edge computing center deletes the information of the networked autonomous vehicles in the non-lth road segment unit, and counts the total number of networked automatic driving vehicles on the lth road segment unit.

Calculate the average traffic speed of the lth road segment unit based on the driving speed of each connected autonomous vehicle on the same road segment

从而计算第l个路段单元的流量Q_l(k)＝n_lK_l(k)V_l(k)；Obtain the traffic density of the l-th road segment unit by the total number of vehicles N _l (k) in the l-th road segment unit

Thereby, the flow Q _l (k)=n _l K _l (k) V _l (k) of the l-th road segment unit is calculated;

The lth edge computing center is connected to the network to obtain the current weather conditions of the urban road network, and inquires the free flow speed V _l ^f of the lth road segment unit under the current weather conditions in the traffic flow database, and judges the traffic state of the lth road segment unit:

When V _l (k)>V _l ^f ×50%, it means that the l-th road segment unit is in a smooth state;

When V _l (k)≤V _l ^f ×50%, it means that the l-th road section unit is in a congested state;

车辆总数N_l(k)、流量Q_l(k)以及交通状态的判断结果所构成的交通状态信息上传至数据中心；The lth edge computing center calculates the total number of connected autonomous vehicles in the lth road segment unit.

The traffic state information composed of the total number of vehicles N _l (k), the flow rate Q _l (k) and the judgment result of the traffic state is uploaded to the data center;

Step 4: According to the traffic state information of each road section unit in the urban road network, the data center uses the density-based spatial clustering algorithm to cluster the road section units in the congested state, and obtains several clustering results, each clustering result is A cluster composed of several congested road section units; the smallest polygon of the area occupied by each clustering result in the urban road network is recorded as an area;

When the total length S _A of a certain clustering result in the occupied area A is greater than the threshold S _cr , the urban road network is considered to be in a congested state, and the corresponding area is identified as a congested area, as shown in Figure 2 Otherwise, it means that the corresponding area is a clear area, and no boundary control or traffic guidance measures are required to implement the road network. After the k-th control period ends, after assigning k+1 to k, return to step 2; where, S _cr =S ₀ × 8%, S ₀ is the total length of the road unit in the urban road network;

Step 5: Denote the road section unit in the congested area as set Θ, denote the road section unit in the direction of exiting the congested area near the border of the congested area as set Ο, and denote the road section unit in the direction of entering the congested area near the border of the congested area as set Ω;

及拥堵区域内车辆总数

则拥堵区域内网联自动驾驶车辆渗透率为p(k)＝N^CAV(k)/N(k)；Calculate the total number of connected autonomous vehicles in a congested area

and the total number of vehicles in the congested area

Then the penetration rate of connected autonomous vehicles in the congested area is p(k)=N ^CAV (k)/N(k);

Based on the traffic data point set in the traffic flow database consisting of the total number of vehicles in the congested area and the travel completion rate of the congested area under the total number of vehicles in the congested area, generate the road network of the congested area when the penetration rate of connected autonomous vehicles is p(k). The macroscopic basic map, so as to fit the macroscopic basic map function G _A (x) of the congested area and generate a 95% prediction band, so as to obtain the total number of vehicles in the congested area corresponding to the maximum value G _max (k) of G _A (x) The critical value of N _cr (k) is shown in Figure 3; among them, the travel completion rate of a certain area represents the number of vehicles driving out of the area per unit time; x represents the total number of vehicles in the congested area; G _A (x) The value represents the trip completion rate when the total number of vehicles in the congested area is x, G _A (x) = ax ³ +bx ² +cx, a, b, c are the fitting parameters of the macroscopic fundamental graph function G _A (x); 95 % The prediction band is the area where 95% of the data points belong to when fitting;

Step 6: Judging the validity of boundary control according to the total number of vehicles in the current congestion area and the travel completion rate. If the boundary control is valid, go to Step 7; if the boundary control is invalid, go to Step 9;

Step 6.1: According to the total number of vehicles N(k) in the congested area and the macro basic graph fitting function G _A (x), obtain the theoretical value of the trip completion rate G _A (N(k)) in the congested area;

与理论值G_A(N(k))之间的残差ΔG_A；其中，

Step 6.2: Calculate the actual value of the trip completion rate for the congested area

Residual _ΔGA from the theoretical value GA (N( _k )); where,

构成的交通数据点是否处于95％预测带内，若处于，则边界控制有效；否则，表示边界控制失效。Step 6.3: Determine the total number of vehicles N(k) in the current congestion area and the actual value of the trip completion rate according to the residual ΔG _A

Whether the constituted traffic data points are within the 95% prediction band, if so, the boundary control is valid; otherwise, the boundary control is invalid.

l∈Ω；对于拥堵区边界处有信号灯控制的交叉口，将进入拥堵区方向的相位时长由t₀调整为u(k)t₀，以降低驶入拥堵区的交通量，其中，

表示未进行边界控制时第l个路段单元的限速值，l∈Ω；拥堵区边界模型预测控制方法步骤包括：Step 7: Taking the travel completion rate of the congested area to reach the maximum value G _max (k) as the goal, implement model predictive control on the boundary of the congested area, and calculate the control rate u(k) of the congested area boundary in the kth control period. Adjust the speed limit value of the lth road segment unit in the direction of entering the congestion area near the boundary of the congestion area as

l∈Ω; for the intersection with signal light control at the boundary of the congestion area, the phase duration in the direction of entering the congestion area is adjusted from t ₀ to u(k)t ₀ to reduce the amount of traffic entering the congestion area, where,

Represents the speed limit value of the l-th road segment unit without boundary control, l∈Ω; the steps of the congestion zone boundary model predictive control method include:

Step 7.1: According to the total number of vehicles N(k) in the congested area and the critical value _Ncr (k) of the total number of vehicles in the congested area in the k-th control cycle, the state variable of the system is obtained as X(k)=N(k)- N _cr (k), let the control variable of the system be U(k)=u(k), and the state equation of the system is X(k+1)=[-β(k)G′ _A (N(k))+ 1]X(k)+[q _in (k)Tq _out (k)T]U(k), the output of the state equation is the total number of vehicles N(k+1) and the number of vehicles in the congested area of the k+1th control cycle. The difference between the total critical value N _cr (k+1); among them, β(k) indicates that in the k-th control cycle, the travel destination of the vehicles in the congested area is the vehicle in the congested area, which accounts for 5% of the total travel vehicles. ratio; q _in (k) represents the traffic flow rate into the congested area in the k-th control cycle; q _out (k) represents the traffic flow rate out of the congested area in the k-th control cycle; G′ _A (N(k)) represents The value of the derivative function G′ _A (x) of the macroscopic basic graph fitting function G _A (x) when x=N(k);

Step 7.2: Obtain the state variables of the system in the next n _p control cycles according to the state equation of the system, and calculate the control rate in the next n _p control cycles, n _c <n _p , use the matrix X=[X(k), X(k+1),…X(k+n _c ),…X(k+n _p )] represents the state variables of the system in the current control cycle and the next n _p control cycles, the current control cycle and the future n _p The control variable matrix of the system in the control period is U=[U(k), U(k+1),...U(k+n _c ),...U(k+n _p )], when the predicted k'th When the control period k+n _c <k′≤k+n _p , let U(k′)=U(k+n _c ); where X(k+n _c ) represents the k+n _c th control period. State variable; U(k+n _c ) represents the control variable of the k+n _c -th control cycle;

Step 7.3: The output of the system is the state variable of the congested area in the current control cycle and the next n _p control cycles;

When the output of the system approaches 0, the travel completion rate is maximized, and the objective function Z solved in the control period is: Z=minX ^T ·H·X; where H is the weight matrix; X ^T represents the rotation of X set;

Step 7.4: Obtain the control variables in each control cycle by solving the objective function Z, and select the first element value U(k) of the control variable matrix U as the boundary control rate of the kth control cycle.

拥堵区域以及管控区域，第l个边缘计算中心将调整后的限速值

传输给第l个路段单元内的网联自动驾驶车辆，使得网联自动驾驶车辆减速行驶，l∈Ω，临近拥堵区域边界的驶入拥堵区方向的路段单元中，网联自动驾驶车辆减速后在路段中形成移动瓶颈，使得人工驾驶车辆难以以高于限速值的速度行驶，迫使人工驾驶车辆减速，最终形成一个个以网联自动驾驶车辆为头车的车辆队列，如图4、图5所示，调整后的限速得以快速有效的实施落实，路段单元内的交通流流速降低，单位时间内驶入拥堵区域的车辆减少；其中，管控区域表示拥堵区域及临近拥堵区域边界的驶入拥堵区方向的路段单元所占面积的最小多边形，如图2所示；Step 8: The data center transmits the adjusted speed limit value of the lth road segment unit to the lth edge computing center

Congested area and control area, the speed limit value adjusted by the lth edge computing center

It is transmitted to the connected self-driving vehicle in the l-th road segment unit, so that the connected self-driving vehicle decelerates, l∈Ω, in the road segment unit near the border of the congested area entering the direction of the congested area, after the networked self-driving vehicle decelerates A mobile bottleneck is formed in the road section, making it difficult for manually-driven vehicles to drive at a speed higher than the speed limit, forcing the manually-driven vehicles to slow down, and finally forming a vehicle queue with connected autonomous vehicles as the lead vehicles, as shown in Figure 4 and Figure 4. As shown in Figure 5, the adjusted speed limit can be implemented quickly and effectively, the traffic flow velocity in the road section unit is reduced, and the number of vehicles entering the congested area per unit time is reduced; among them, the control area refers to the congested area and the traffic near the border of the congested area. The smallest polygon of the area occupied by the road section unit in the direction of entering the congestion area, as shown in Figure 2;

Step 9: At the upstream intersection of the congested area, continuously inform the downstream area of congestion through the roadside variable information sign within the time interval T, and provide the speed limit information of the surrounding road section units and the recommended driving route to achieve the purpose of traffic guidance; After the k-th control cycle ends, assign k+1 to k, and then return to step 2.

Claims (3)

1. A road network congestion area control method based on multi-source data edge computing in an intelligent networked environment, which is applied to m edge computing centers, n networked autonomous vehicles, m road section units and 1 data center. In the urban road network formed by the coexistence of networked autonomous vehicles and manually driven vehicles, it is characterized in that the method for managing and controlling the congestion area of the road network includes the following steps: Step 1: Consider the road in each flow direction between every two adjacent intersections as a road segment unit, and obtain the history of the l-th road segment unit by considering the lane number n _l and length information L _l of any l-th road segment unit Traffic flow data and corresponding weather data, thereby establishing a traffic flow database including each road segment unit; l=1,2,3,...,m; Set the time interval between two adjacent control cycles as T; the current number of control cycles is k, and initialize k=1; Step 2: Under the k-th control cycle, the i-th network-connected autonomous vehicle sends its own driving speed v _i (k) and position information to the edge computing center on the roadside of the road unit where it is located, and the The traffic flow data collected by the sensor is also transmitted to the edge computing center on the roadside of the road section unit; i=1, 2, 3, ..., n; n represents the total number of connected autonomous vehicles; Step 3: Under the kth control cycle, the lth edge computing center processes the traffic flow data to obtain the total number of vehicles N _l (k) in the lth road segment unit; According to the location information of each networked autonomous vehicle, the lth edge computing center deletes the information of the networked automatic driving vehicles in the non-lth road segment unit, and counts the total number of networked automatic driving vehicles on the lth road segment unit N _l ^CAV (k); Calculate the average traffic speed of the lth road segment unit based on the driving speed of each connected autonomous vehicle on the same road segment

Obtain the traffic density of the l-th road segment unit by the total number of vehicles N _l (k) in the l-th road segment unit

Thereby, the flow Q _l (k)=n _l K _l (k) V _l (k) of the l-th road segment unit is calculated; The lth edge computing center is networked to obtain the current weather conditions of the urban road network, and inquires the free flow speed V _l ^f of the lth road segment unit under the current weather conditions in the traffic flow database to determine the traffic state of the lth road segment unit : When V _l (k)>V _l ^f ×δ%, it means that the l-th road segment unit is in a smooth state; When V _l (k)≤V _l ^f ×δ%, it means that the l-th road segment unit is in a congested state; The lth edge computing center uses the judgment results of the total number of connected autonomous vehicles N _l ^CAV (k), the total number of vehicles N _l (k), the flow Q _l (k) and the traffic state in the lth road segment unit. The constituted traffic status information is uploaded to the data center; wherein, δ represents the set threshold; Step 4: The data center uses the density-based spatial clustering algorithm to cluster the road section units in the congested state according to the traffic state information of each road section unit in the urban road network, and obtains several clustering results, each clustering The result is a cluster composed of several congested road section units; the smallest polygon of the area occupied by each clustering result in the urban road network is recorded as the area; When the total length S _A of a certain clustering result in the occupied area A is greater than the threshold S _cr , the urban road network is considered to be in a congested state, and the corresponding area is identified as a congested area; otherwise, it indicates that the corresponding area is congested. The area is a clear area, and there is no need to implement boundary control or traffic guidance measures on the road network. After the k-th control cycle ends, assign k+1 to k, and then return to step 2; where, S _cr =S ₀ ×Δ%, S ₀ is the total length of the road unit in the urban road network; Δ represents the set threshold; Step 5: Denote the road section unit in the congested area as set Θ, denote the road section unit in the direction of exiting the congested area near the border of the congested area as set Ο, and denote the road section unit in the direction of entering the congested area near the border of the congested area as set Ω; Calculate the total number of connected autonomous vehicles in a congested area

and the total number of vehicles in the congested area

Then the penetration rate of connected autonomous vehicles in the congested area is p(k)=N ^CAV (k)/N(k); Based on the traffic data point set in the traffic flow database consisting of the total number of vehicles in the congested area and the travel completion rate of the congested area under the total number of vehicles in the congested area, generate the road network of the congested area when the penetration rate of connected autonomous vehicles is p(k). The macroscopic basic map, so as to fit the macroscopic basic map function G _A (x) of the congested area and generate a 95% prediction band, so as to obtain the total number of vehicles in the congested area corresponding to the maximum value G _max (k) of G _A (x) The critical value of N _cr (k); where the travel completion rate of a certain area represents the number of vehicles leaving the area per unit time; x represents the total number of vehicles in the congested area; the value of G _A (x) represents the vehicles in the congested area The trip completion rate when the total number is x; the 95% prediction band is the region to which 95% of the data points belong at the time of fitting; Step 6: Judging the validity of boundary control according to the total number of vehicles in the current congestion area and the travel completion rate, if the boundary control is valid, go to Step 7; if the boundary control is invalid, go to Step 9; Step 7: Taking the travel completion rate of the congested area reaching the maximum value G _max (k) as the goal, implement model predictive control on the boundary of the congested area, and calculate the control rate u(k) of the congested area boundary in the kth control period; Adjust the speed limit value of the lth road segment unit in the direction of entering the congestion area near the boundary of the congestion area as

l∈Ω; for the intersection with signal light control at the boundary of the congestion area, the phase duration in the direction of entering the congestion area is adjusted from t ₀ to u(k)t ₀ to reduce the amount of traffic entering the congestion area, where,

Represents the speed limit value of the l-th road segment unit without boundary control, l∈Ω; Step 8: The data center transmits the adjusted speed limit value of the lth road segment unit to the lth edge computing center

Congested area and control area, the speed limit value adjusted by the lth edge computing center

It is transmitted to the networked autonomous vehicle in the l-th road segment unit, so that the networked autonomous driving vehicle decelerates, l∈Ω; where, the control area represents the congestion area and the road segment unit near the boundary of the congestion area that enters the congestion area. The smallest polygon that occupies an area; Step 9: At the upstream intersection of the congested area, continuously inform the downstream area of congestion through the roadside variable information sign within the time interval T, and provide the speed limit information and suggested driving route of the surrounding road units to achieve the purpose of traffic guidance ; After the k-th control cycle ends, assign k+1 to k, and return to step 2. 2. The method for managing and controlling a road network congested area according to claim 1, wherein the boundary control validity judgment in the step 6 comprises: Step 6.1: According to the total number of vehicles N(k) in the congested area and the macro basic graph fitting function G _A (x), obtain the theoretical value of the trip completion rate G _A (N(k)) in the congested area; Step 6.2: Calculate the actual value of the trip completion rate for the congested area

Residual _ΔGA from the theoretical value GA (N( _k )); where,

Step 6.3: Determine the total number of vehicles N(k) in the current congestion area and the actual value of the trip completion rate according to the residual ΔG _A

Whether the constituted traffic data points are within the 95% prediction band, if so, the boundary control is valid; otherwise, the boundary control is invalid. 3. The road network congested area management and control method according to claim 1, wherein the congested area boundary model predictive control in step 7 comprises: Step 7.1: According to the total number of vehicles N(k) in the congested area and the critical value _Ncr (k) of the total number of vehicles in the congested area in the k-th control cycle, the state variable of the system is obtained as X(k)=N(k)- N _cr (k), let the control variable of the system be U(k)=u(k), and the state equation of the system is X(k+1)=[-β(k)G′ _A (N(k))+ 1]X(k)+[q _in (k)Tq _out (k)T]U(k), the output of the state equation is the total number of vehicles N(k+1) and the number of vehicles in the congested area of the k+1th control cycle. The difference between the total critical value N _cr (k+1); among them, β(k) indicates that in the k-th control cycle, the travel destination of the vehicles in the congested area is the vehicle in the congested area, which accounts for 5% of the total travel vehicles. ratio; q _in (k) represents the traffic flow rate into the congested area in the k-th control cycle; q _out (k) represents the traffic flow rate out of the congested area in the k-th control cycle; G′ _A (N(k)) represents The value of the derivative function G′ _A (x) of the macroscopic basic graph fitting function G _A (x) when x=N(k); Step 7.2: Obtain the state variables of the system in the next n _p control cycles according to the state equation of the system, and calculate the control rate in the next n _p control cycles, n _c <n _p , use the matrix X=[X(k), X(k+1),…X(k+n _c ),…X(k+n _p )] represents the state variables of the system in the current control cycle and the next n _p control cycles, the current control cycle and the future n _p The control variable matrix of the system in the control period is U=[U(k), U(k+1),...U(k+n _c ),...U(k+n _p )], when the predicted k'th When the control period k+n _c <k′≤k+n _p , let U(k′)=U(k+n _c ); where X(k+n _c ) represents the k+n _c th control period. State variable; U(k+n _c ) represents the control variable of the k+n _c -th control cycle; Step 7.3: The output of the system is the state variable of the congested area in the current control cycle and the next n _p control cycles; When the output of the system approaches 0, the travel completion rate is maximized, and the objective function Z to be solved in the control period is: Z=min X ^T · H · X; where H is the weight matrix; X ^T represents the Transpose; Step 7.4: Obtain the control variables in each control cycle by solving the objective function Z, and select the first element value U(k) of the control variable matrix U as the boundary control rate of the kth control cycle.

Images