阅读说明：本技术 一种考虑扰动的城市交通区域迭代学习边界控制方法 (Urban traffic area iterative learning boundary control method considering disturbance ) 是由 闫飞 王昆 田建艳 于 2021-06-03 设计创作，主要内容包括：本发明公开了一种考虑扰动的城市交通区域迭代学习边界控制方法,基于城市交通区域的宏观基本图特性以及宏观交通流的周期性特征,在车辆平衡方程中引入扰动项,将迭代学习边界控制计算的结果与非对称信号配时策略结合得出边界交叉口各相位信号时长,通过对边界交叉口信号进行多次迭代学习控制,将控制子区的累计车辆数量稳定到期望值,从而提高路网的整体运行效率。本发明考虑了实际城市交通系统运行中的扰动因素对控制作用性能的影响,构建的车辆平衡方程更接近路网实际情况；且开闭环PD型迭代学习控制可以克服开环控制的不足,能同时利用系统前一次和当前的输出误差信息,进一步抑制系统的不确定性,提高系统的鲁棒性。(The invention discloses an urban traffic area iterative learning boundary control method considering disturbance, which introduces a disturbance item into a vehicle balance equation based on the macroscopic basic graph characteristics of an urban traffic area and the periodic characteristics of macroscopic traffic flow, combines the results of iterative learning boundary control calculation with an asymmetric signal timing strategy to obtain the signal duration of each phase of a boundary intersection, and stabilizes the accumulated vehicle number of a control subarea to a desired value by carrying out iterative learning control on signals of the boundary intersection for multiple times, thereby improving the overall operation efficiency of a road network. The influence of disturbance factors on the control action performance in the operation of an actual urban traffic system is considered, and the constructed vehicle balance equation is closer to the actual situation of a road network; and the open-close loop PD type iterative learning control can overcome the defect of open-loop control, and can utilize the previous and current output error information of the system at the same time, further inhibit the uncertainty of the system and improve the robustness of the system.)

1. An iterative learning boundary control method for an urban traffic region considering disturbance is characterized by comprising the following steps:

1) selecting a road network to be researched, and dividing the road network into a plurality of homogeneous areas;

2) drawing a macroscopic basic diagram of the road network subareas divided in the step 1) according to the vehicle data collected by the detector in the road network;

3) constructing a vehicle balance equation after considering the disturbance;

4) judging the running state of the road network, and if the running state is the congestion state, implementing boundary control; otherwise, continuously monitoring the running state of the road network;

5) and implementing boundary control on the road network under the action of disturbance of different degrees, introducing an iterative control scheme by utilizing the repetitive characteristics of urban traffic flow, and setting the green light time length of each phase of the boundary intersection according to the green light time length increment obtained by iterative learning control and an asymmetric signal timing strategy in a combined manner, so that the error value between the actual vehicle number and the expected vehicle number in the road network is gradually reduced until an ideal state is reached.

2. The iterative learning boundary control method for urban traffic areas with disturbance considered according to claim 1, wherein in the step of dividing homogeneous areas, the iterative learning boundary control method comprises the steps of:

1.1) selecting a road network G to be researched, obtaining a road topological structure of the road network, designing and installing detectors at the positions of an inlet road and an outlet road of each signalized intersection, and obtaining actual traffic flow data of each intersection in the road network, wherein the steps comprise: the number N of lanes between adjacent intersections, the traffic flow Q, the number h of traffic flow branches and the average travel time T;

1.2) calculating the traffic association degree between adjacent intersections based on the obtained road topological structure and the actual traffic flow data, wherein the calculation formula of the traffic association degree between the adjacent intersection i and the intersection j is as follows:

wherein, w_ijIs the traffic association degree between the adjacent intersection i and the adjacent intersection j; i is_fThe road section flow imbalance coefficient is obtained; h is the number of branches for the traffic flow from the upstream intersection to enter;the sum of the traffic flow reaching the downstream intersection; q_maxThe maximum traffic flow in the branch for the traffic flow from the upstream intersection is Q_sMaximum value of (1); t is the average travel time of the vehicle between two adjacent intersections; n is the number of lanes from the upstream intersection to the downstream intersection;

1.3) calculating a density Laplace matrix L of a road network G according to the relevance information between adjacent signalized intersections, constructing a clustering feature matrix X, carrying out rapid global K-means clustering on the clustering feature matrix X, and dividing signalized intersections belonging to the same cluster into the same control subarea.

3. The iterative learning boundary control method for urban traffic areas with disturbance considered according to claim 1, wherein the step of drawing a macroscopic basic graph of the divided road network sub-areas comprises:

2.1) collecting vehicle data of a region to be researched from a low peak state to a high peak state, acquiring the number of vehicles of main roads in a road network meeting macroscopic basic diagram drawing conditions, and counting the number N of vehicles in the region at the time t;

2.2) processing the acquired vehicle data, eliminating data points with obvious errors and acquiring effective macroscopic basic pattern cost values;

2.3), drawing a road network MFD scatter diagram by taking the total number N of vehicles driving away from the road network in the sampling time interval delta t in the road network as an abscissa and taking the total number N of the vehicles driving away from the road network in the sampling time interval delta t in the road network as an ordinate, and fitting the MFD curve of the road network according to the scatter data to obtain the MFD curve of an area to be researched and the critical cumulative number N of the vehicles in the area_cr。

4. The iterative learning boundary control method for urban traffic areas with consideration of disturbance according to claim 1, wherein the step of constructing a vehicle balance equation with consideration of disturbance comprises:

3.1) setting an urban road network, dividing the urban road network into a plurality of well-defined macroscopic basic graph sub-areas according to the traffic flow density or other correlation parameters, wherein inflow and outflow of traffic flows among the sub-areas are relatively stable, the sub-area 1 is an area in an oversaturated state or extremely prone to traffic congestion, and the sub-area 2 is a peripheral area thereof, so that the vehicle dynamic balance equation is expressed as follows:

wherein x is_ij(t) the cumulative number of vehicles in the sub-area i at time t and with the sub-area j as the destination; q. q.s₁₁(t) is the time tThe internal traffic requirements of zone 1; q. q.s_ij(t) traffic flow rate from sub-zone i to sub-zone j at time t; u. of_ij(t) is the boundary control rate from sub-area i to sub-area j at time t; x is the number of₁(t) is the accumulated number of vehicles in the sub-area 1 at the time t; g₁(x₁(t)) is the cumulative number of vehicles in sub-area 1 to reach x₁(t) a travel vehicle completion amount;

3.2), setting a control area boundary and a boundary control gate, and considering three disturbance forms which are easy to influence the accumulated vehicle number in the operation of the actual urban traffic system: (a) vehicles enter and exit the parking lot, (b) pedestrians cross the road randomly, and (c) sudden traffic accidents;

3.3) in the calculation process by adopting the boundary control method, in the current control period T, the phase green light time of each signal period is kept unchanged until the next control period, and then the green light time is adjusted according to the green light time obtained by calculation of the previous control period, so that the discretization of the vehicle balance equation after disturbance is added is obtained:

where T is the sampling period, q_ξij(k) Disturbance affecting the running state of the vehicle or the accumulated number of vehicles in the road network; eta (k) is the error produced by the output of the data acquisition equipment in the subarea.

5. The iterative learning boundary control method for urban traffic areas with disturbance considered according to claim 2, wherein the step of judging the road network operation state comprises the steps of:

4.1) according to the MFD characteristics of the road network, the accumulated vehicle number in the road network has a critical value, and when the accumulated vehicle number reaches a preset threshold range near the critical value, the finished quantity of the travel vehicles in the road network is maximum; when the actual accumulated vehicle number is smaller than the minimum value of the preset threshold range, the road network is in a smooth state, and the finished quantity of the traveling vehicles is increased along with the increase of the accumulated vehicle number; once the actual accumulated vehicle number is larger than the maximum value of the preset threshold range, the completion amount of the traveling vehicles in the road network is reduced, so that the traffic in the area is in a congestion state;

4.2), if the cumulative number of vehicles in the road network is greater than the critical cumulative number of vehicles N_crIf the road network is in a congestion state, regarding signalized intersections at the outermost periphery of the area as boundary nodes, and performing boundary control on the area; if the accumulated number of vehicles in the road network is less than the critical accumulated number of vehicles N_crAnd continuing to monitor the running state of the road network.

6. The iterative learning boundary control method for the urban traffic area considering the disturbance according to claim 1, wherein the boundary control is implemented on the road network under the action of different degrees of disturbance, an iterative control scheme is introduced by utilizing the repetitive characteristics of urban traffic flow, and the green light time length of each phase of the boundary intersection is set according to the green light time length increment obtained by the iterative learning control and the asymmetric signal timing strategy in a combined manner, so that the error value between the actual number of vehicles and the expected number of vehicles in the road network is gradually reduced until the ideal state is reached, the iterative learning boundary control method comprises the following steps:

5.1), in order to meet the influence of pedestrian crossing time and signal period factors, the green light duration of each phase of the traffic signal must be set within a reasonable range, namely, the control input after iterative learning control can meet the actual traffic condition:

u_n(k)∈[u_min(k),u_max(k)]

5.2) in order to reduce the influence of disturbance on the performance of the control system and enhance the robustness of the system, the iterative control scheme adopts open-closed loop PD type iteration, and the iterative learning control law is as follows:

u_n+1(k)＝sat[u_n(k)]+β₁(e_n(k+1)-e_n(k))+β₂e_n+1(k)

wherein u is_n(k) To control input, sat [ u ]_n(t)]A saturation function that is a control input; beta is a₁For iterative learning gain, beta₂Controlling the gain for feedback;

5.3) when the road network does not enter the congestion state, setting the expected curve as the accumulated vehicle number in fixed time distribution; when the road network accumulated vehicle number exceeds the critical accumulated vehicle number, setting the expected curves as the critical accumulated vehicle number;

5.4) the first iteration adopts the original fixed timing scheme, the error is made between the second iteration and the accumulated vehicle number collected last time, the control input of the current moment is calculated, and the green light time length of each phase of the boundary intersection, namely g, is set according to the green light time length increment obtained by iterative learning control and the asymmetric signal timing strategy_k+1(t)＝g_k(t)+Δu；

5.5) in the current control period, the phase green signal ratio of each signal period is kept unchanged, and when the next control period is reached, the green signal ratio is adjusted according to the green signal ratio calculated in the next control period.

Technical Field

The invention relates to the technical field of traffic control, in particular to an urban traffic area iterative learning boundary control method considering disturbance.

Background

Due to the expansion of the traffic jam range and the complexity of traffic flow, the traditional single-point and trunk line coordination control method can only partially improve the traffic state of a single intersection, and is difficult to solve the problem of large-scale traffic jam. The Macro Fundamental Diagram (MFD) characteristic of the road network can reflect the relationship between the traffic in the road network and the operating efficiency of the road network, can monitor the operating state of the road network from a macro level, and provides a good tool for solving traffic congestion.

The prior MFD-based boundary control has abundant research results, but most of the MFD-based boundary control methods are still model-based feedback control algorithms, and the control effect of the feedback control algorithms is easily influenced by model parameter selection; from a macroscopic view, the traffic flow of each day has similar space-time distribution characteristics aiming at a specific city area, so that the urban traffic flow can be considered to have the repetitive characteristics. Iterative Learning Control (ILC) may be used to solve the tracking control problem of a repeatedly-operated system within a limited interval, and it aims at complete tracking, does not depend on an accurate model of the system, has strong adaptability, is easy to implement, obtains a control input capable of generating a desired output trajectory only through information of previous operations to improve control quality, and is suitable for use in traffic control with repeatability.

The actual engineering system is often influenced by various interference and uncertainty factors, and the influence of interference of different degrees on a controlled system can weaken the performance of control action and even lead to unpredictable results, so that the influence of disturbance on the control performance is very important to research. The existing regional boundary control method is mostly only suitable for road networks in an ideal state, and the influence of disturbance factors such as traffic accidents, pedestrian crossing streets, parking lots and the like in the operation of an actual urban traffic system is not considered, and the disturbance factors not only influence the control performance of a traffic signal controller, but also possibly further aggravate the generated traffic jam.

Disclosure of Invention

In order to overcome the defects of the prior art, the invention provides an iterative learning boundary control method for an urban traffic area with disturbance considered.

In order to achieve the above object, the present invention provides a disturbance-considered boundary control method for iterative learning of an urban traffic region, including:

1) selecting a road network to be researched, and dividing the road network into a plurality of homogeneous areas;

2) drawing a macroscopic basic diagram of the road network subareas divided in the step 1) according to the vehicle data collected by the detector in the road network;

3) constructing a vehicle balance equation after considering the disturbance;

4) judging the running state of the road network, and if the running state is the congestion state, implementing boundary control; otherwise, continuously monitoring the running state of the road network;

5) and implementing boundary control on the road network under the action of disturbance of different degrees, introducing an iterative control scheme by utilizing the repetitive characteristics of urban traffic flow, and setting the green light time length of each phase of the boundary intersection according to the green light time length increment obtained by iterative learning control and an asymmetric signal timing strategy in a combined manner, so that the error value between the actual vehicle number and the expected vehicle number in the road network is gradually reduced until an ideal state is reached.

Wherein, in the step of dividing the homogeneous region, the method comprises the steps of:

1.1) selecting a road network G to be researched, obtaining a road topological structure of the road network, designing and installing detectors at the positions of an inlet road and an outlet road of each signalized intersection, and obtaining actual traffic flow data of each intersection in the road network, wherein the steps comprise: the number N of lanes between adjacent intersections, the traffic flow Q, the number h of traffic flow branches and the average travel time T;

1.2) calculating the traffic association degree between adjacent intersections based on the obtained road topological structure and the actual traffic flow data, wherein the calculation formula of the traffic association degree between the adjacent intersection i and the intersection j is as follows:

wherein, w_ijIs the traffic association degree between the adjacent intersection i and the adjacent intersection j; i is_fThe road section flow imbalance coefficient is obtained; h is the number of branches for the traffic flow from the upstream intersection to enter;the sum of the traffic flow reaching the downstream intersection; q_maxThe maximum traffic flow in the branch for the traffic flow from the upstream intersection is Q_sMaximum value of (1); t is the average travel time of the vehicle between two adjacent intersections; n is the number of lanes from the upstream intersection to the downstream intersection;

1.3) calculating a density Laplace matrix L of a road network G according to the relevance information between adjacent signalized intersections, constructing a clustering feature matrix X, carrying out rapid global K-means clustering on the clustering feature matrix X, and dividing signalized intersections belonging to the same cluster into the same control subarea.

The step of drawing the macroscopic basic graph of the divided road network subareas comprises the following steps:

2.1) collecting vehicle data of a region to be researched from a low peak state to a high peak state, acquiring the number of vehicles of main roads in a road network meeting macroscopic basic diagram drawing conditions, and counting the number N of vehicles in the region at the time t;

2.2) processing the acquired vehicle data, eliminating data points with obvious errors and acquiring effective macroscopic basic pattern cost values;

2.3), drawing a road network MFD scatter diagram by taking the total number N of vehicles driving away from the road network in the sampling time interval delta t in the road network as an abscissa and taking the total number N of the vehicles driving away from the road network in the sampling time interval delta t in the road network as an ordinate, and fitting the MFD curve of the road network according to the scatter data to obtain the MFD curve of an area to be researched and the critical cumulative number N of the vehicles in the area_cr。

Wherein the step of constructing a vehicle balance equation after considering the disturbance comprises:

3.1) setting an urban road network, dividing the urban road network into a plurality of well-defined macroscopic basic graph sub-areas according to the traffic flow density or other correlation parameters, wherein inflow and outflow of traffic flows among the sub-areas are relatively stable, the sub-area 1 is an area in an oversaturated state or extremely prone to traffic congestion, and the sub-area 2 is a peripheral area thereof, so that the vehicle dynamic balance equation is expressed as follows:

wherein x is_ij(t) the cumulative number of vehicles in the sub-area i at time t and with the sub-area j as the destination; q. q.s₁₁(t) is the internal traffic demand of sub-area 1 at time t; q. q.s_ij(t) traffic flow rate from sub-zone i to sub-zone j at time t; u. of_ij(t) is the boundary control rate from sub-area i to sub-area j at time t; x is the number of₁(t) is the accumulated number of vehicles in the sub-area 1 at the time t; g₁(x₁(t)) is the cumulative number of vehicles in sub-area 1 to reach x₁(t) a travel vehicle completion amount;

3.2), setting a control area boundary and a boundary control gate, and considering three disturbance forms which are easy to influence the accumulated vehicle number in the operation of the actual urban traffic system: (a) vehicles enter and exit the parking lot, (b) pedestrians cross the road randomly, and (c) sudden traffic accidents;

3.3) in the calculation process by adopting the boundary control method, in the current control period T, the phase green light time of each signal period is kept unchanged until the next control period, and then the green light time is adjusted according to the green light time obtained by calculation of the previous control period, so that the discretization of the vehicle balance equation after disturbance is added is obtained:

where T is the sampling period, q_ξij(k) Disturbance affecting the running state of the vehicle or the accumulated number of vehicles in the road network; eta (k) is the error produced by the output of the data acquisition equipment in the subarea.

The step of judging the road network running state comprises the following steps:

4.1) according to the MFD characteristics of the road network, the accumulated vehicle number in the road network has a critical value, and when the accumulated vehicle number reaches a preset threshold range near the critical value, the finished quantity of the travel vehicles in the road network is maximum; when the actual accumulated vehicle number is smaller than the minimum value of the preset threshold range, the road network is in a smooth state, and the finished quantity of the traveling vehicles is increased along with the increase of the accumulated vehicle number; once the actual accumulated vehicle number is larger than the maximum value of the preset threshold range, the completion amount of the traveling vehicles in the road network is reduced, so that the traffic in the area is in a congestion state;

4.2), if the cumulative number of vehicles in the road network is greater than the critical cumulative number of vehicles N_crIf the road network is in a congestion state, regarding signalized intersections at the outermost periphery of the area as boundary nodes, and performing boundary control on the area; if the accumulated number of vehicles in the road network is less than the critical accumulated number of vehicles N_crAnd continuing to monitor the running state of the road network.

The method comprises the following steps of implementing boundary control on the road network under the action of disturbance of different degrees, introducing an iterative control scheme by utilizing the repetitive characteristic of urban traffic flow, setting the green light time length of each phase of a boundary intersection according to the green light time length increment obtained by iterative learning control and an asymmetric signal timing strategy in a combined manner, and gradually reducing the error value between the actual vehicle number and the expected vehicle number in the road network until an ideal state is reached, wherein the method comprises the following steps:

5.1), in order to meet the influence of pedestrian crossing time and signal period factors, the green light duration of each phase of the traffic signal must be set within a reasonable range, namely, the control input after iterative learning control can meet the actual traffic condition:

u_n(k)∈[u_min(k),u_max(k)]

5.2) in order to reduce the influence of disturbance on the performance of the control system and enhance the robustness of the system, the iterative control scheme adopts open-closed loop PD type iteration, and the iterative learning control law is as follows:

u_n+1(k)＝sat[u_n(k)]+β₁(e_n(k+1)-e_n(k))+β₂e_n+1(k)

wherein u is_n(k) To control input, sat [ u ]_n(t)]A saturation function that is a control input; beta is a₁For iterative learning gain, beta₂Controlling the gain for feedback;

5.3) when the road network does not enter the congestion state, setting the expected curve as the accumulated vehicle number in fixed time distribution; when the road network accumulated vehicle number exceeds the critical accumulated vehicle number, setting the expected curves as the critical accumulated vehicle number;

5.4) the first iteration adopts the original fixed timing scheme, the error is made between the second iteration and the accumulated vehicle number collected last time, the control input of the current moment is calculated, and the green light time length of each phase of the boundary intersection, namely g, is set according to the green light time length increment obtained by iterative learning control and the asymmetric signal timing strategy_k+1(t)＝g_k(t)+Δu；

5.5) in the current control period, the phase green signal ratio of each signal period is kept unchanged, and when the next control period is reached, the green signal ratio is adjusted according to the green signal ratio calculated in the next control period.

The method is characterized in that a disturbance item is introduced into a vehicle balance equation based on the macroscopic basic diagram characteristics of the urban traffic area and the periodic characteristics of the macroscopic traffic flow, the time length of each phase signal of the boundary intersection is obtained by combining the result of iterative learning boundary control calculation and an asymmetric signal timing strategy, and the accumulated vehicle number of the control subarea is stabilized to a desired value by carrying out iterative learning control on the signals of the boundary intersection for multiple times, so that the overall operation efficiency of the road network is improved. The influence of disturbance factors on the control action performance in the operation of an actual urban traffic system is considered, and the constructed vehicle balance equation is closer to the actual situation of a road network; and the open-close loop PD type iterative learning control can overcome the defect of open-loop control, and can utilize the previous and current output error information of the system at the same time, further inhibit the uncertainty of the system and improve the robustness of the system.

Drawings

The invention will be further described with reference to the accompanying drawings and examples, in which:

FIG. 1 is a schematic flow chart of an iterative learning boundary control method for an urban traffic area with consideration of disturbance according to the present invention;

FIG. 2 is a schematic view of a traffic sub-region macroscopic traffic flow model of an urban traffic region iterative learning boundary control method considering disturbance according to the present invention;

fig. 3 is a schematic diagram of three common disturbance forms in an urban traffic system of the urban traffic area iterative learning boundary control method considering disturbance according to the present invention.

Fig. 4 is a schematic block diagram of an open-close loop PD type iterative learning control principle of the disturbance-considered urban traffic area iterative learning boundary control method provided by the invention.

Fig. 5 is a schematic phase diagram of a boundary intersection in an urban traffic area iterative learning boundary control method considering disturbance according to the present invention.

Fig. 6 is a schematic diagram of a two-phase asymmetric signal timing modification scheme at a boundary intersection in an urban traffic area iterative learning boundary control method considering disturbance.

Detailed Description

For a more clear understanding of the technical features, objects and effects of the present invention, embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

Referring to fig. 1, the invention provides an iterative learning boundary control method for an urban traffic area considering disturbance, which comprises the following steps:

1) selecting a road network to be researched, and dividing the road network into a plurality of homogeneous areas:

1.1) selecting a road network G to be researched, obtaining a road topological structure of the road network, designing and installing detectors at the positions of an inlet road and an outlet road of each signalized intersection, and obtaining actual traffic flow data of each intersection in the road network, wherein the steps comprise: the number N of lanes between adjacent intersections, the traffic flow Q, the number h of traffic flow branches and the average travel time T;

1.2) calculating the traffic association degree between adjacent intersections based on the obtained road topological structure and the actual traffic flow data, wherein the calculation formula of the traffic association degree between the adjacent intersection i and the intersection j is as follows:

wherein, w_ijIs the traffic association degree between the adjacent intersection i and the adjacent intersection j; i is_fThe road section flow imbalance coefficient is obtained; h is the number of branches for the traffic flow from the upstream intersection to enter;the sum of the traffic flow reaching the downstream intersection; q_maxThe maximum traffic flow in the branch for the traffic flow from the upstream intersection is Q_sMaximum value of (1); t is the average travel time of the vehicle between two adjacent intersections; n is the number of lanes from the upstream intersection to the downstream intersection;

1.3) calculating a density Laplace matrix L of a road network G according to the information of the relevance between adjacent signalized intersections, constructing a clustering feature matrix X, carrying out rapid global K-means clustering on the clustering feature matrix X, and dividing signalized intersections belonging to the same cluster into the same control subarea;

2) drawing a macroscopic basic graph MFD of the road network subareas divided in the step 1) according to the vehicle data collected by the detector in the road network:

2.1) collecting vehicle data of a region to be researched from a low peak state to a high peak state, obtaining the number of vehicles of main roads in a road network, meeting the drawing condition of a macroscopic basic diagram, and counting the number N of vehicles in the region at the time t;

2.2) a data processing stage, processing the acquired data, removing data points with large obvious errors and acquiring effective macroscopic basic pattern cost values;

2.3) drawing a road network MFD scatter diagram by taking the total number N of vehicles driving away from the road network in the sampling time interval delta t in the road network as an abscissa and taking the total number N of the vehicles driving away from the road network in the sampling time interval delta t in the road network as an ordinate, and fitting the MFD curve of the road network according to the scatter data to obtain the MFD curve of an area to be researched and the critical cumulative number N of the vehicles in the area_cr。

3) And constructing a vehicle balance equation after considering the disturbance:

3.1), as shown in fig. 2, assuming that an urban road network can be divided into a plurality of well-defined macroscopic basic map sub-areas according to traffic flow density or other correlation parameters, inflow and outflow of traffic flow between the sub-areas are relatively stable, sub-area 1 is an area in an oversaturated state or very prone to traffic congestion, and sub-area 2 is a peripheral area thereof, the vehicle dynamic balance equation can be expressed as:

wherein x is_ij(t) the cumulative number of vehicles in the sub-area i at time t and with the sub-area j as the destination; q. q.s₁₁(t) is the internal traffic demand of sub-area 1 at time t; q. q.s_ij(t) traffic flow rate from sub-zone i to sub-zone j at time t; u. of_ij(t) is the boundary control rate from sub-area i to sub-area j at time t; x is the number of₁(t) is the accumulated number of vehicles in the sub-area 1 at the time t; g₁(x₁(t)) is the cumulative number of vehicles in sub-area 1 to reach x₁(t) amount of travel vehicle completion.

3.2) as shown in FIG. 3, the dotted line is the boundary of the control area, and the boundary control gate is arranged at the upstream intersection, the invention considers three disturbance forms which are easy to cause influence on the accumulated vehicle number in the actual urban traffic system operation: (a) vehicles enter and exit the parking lot, (b) pedestrians cross the road randomly, and (c) sudden traffic accidents; the influence of vehicles entering and exiting the parking lot on vehicles normally running in a road network and the accumulated number of the vehicles is more serious, particularly in a large-scale business center or a residential area; (ii) a When part of pedestrians do not follow the traffic rules and cross the road, vehicles in the road network which normally run can react to the pedestrians and start to decelerate, and the randomness of the traffic environment is increased; traffic accidents easily cause congestion of road sections, so that the traffic capacity of the road sections is rapidly reduced in a short time, delay of other vehicles in a road network is increased, and the original congestion is further aggravated particularly by the traffic accidents occurring in peak hours.

3.3) because the phase green light duration of each signal period is kept unchanged in the current control period T by the boundary control method adopted by the invention until the next control period, the green light duration obtained by calculation of the previous control period is adjusted, the discretization of the vehicle balance equation after disturbance is added is obtained:

wherein T is a sampling period and is disturbance influencing the running state of the vehicle or the accumulated number of vehicles in a road network; eta (k) is the error produced by the output of the data acquisition equipment in the subarea.

4) Judging the running state of the road network, and if the running state is the congestion state, implementing boundary control; otherwise, continuously monitoring the running state of the road network:

4.1), according to the characteristics of the road network MFD: the accumulated number of vehicles in the road network has a critical value, and when the accumulated number of vehicles reaches the vicinity of the critical value, the finished quantity of the traveling vehicles in the road network is maximum. When the actual accumulated vehicle number is smaller than the critical value, the road network is in a smooth state, and the finished quantity of the traveling vehicles is increased along with the increase of the accumulated vehicle number; if the actual cumulative number of vehicles is greater than the threshold, the completion amount of the traveling vehicles in the road network decreases, and the traffic in the area may be congested.

4.2) if the accumulated number of vehicles in the road network is larger than the critical accumulated number of vehicles N obtained in the step 2)_crIf the road network is in a congestion state, regarding signalized intersections at the outermost periphery of the area as boundary nodes, and performing boundary control on the area; if the accumulated number of vehicles in the road network is less than the critical accumulationCounting the number of vehicles N_crAnd continuing to monitor the running state of the road network.

5) Implementing boundary control on the road network under the action of disturbance of different degrees, introducing an iterative control scheme by using the repetitive characteristics of urban traffic flow, and setting the green light time length of each phase of a boundary intersection according to the green light time length increment obtained by iterative learning control and an asymmetric signal timing strategy in a combined manner, so that the error value between the actual vehicle number and the expected vehicle number in the road network is gradually reduced until an ideal state is reached:

5.1), in order to meet the influence of factors such as pedestrian crossing time, signal period and the like, the green light duration of each phase of the traffic signal must be set within a reasonable range, namely, the control input after iterative learning control can meet the actual traffic condition: u. of_n(k)∈[u_min(k),u_max(k)]；

5.2) in order to reduce the influence of disturbance on the performance of the control system and enhance the robustness of the system, an open-closed loop PD type iteration is adopted in an iteration mode, the schematic diagram of the principle is shown in FIG. 4, and the iterative learning control law is as follows:

u_n+1(k)＝sat[u_n(k)]+β₁(e_n(k+1)-e_n(k))+β₂e_n+1(k)

wherein, sat [ u ]_n(t)]A saturation function that is a control input; beta is a₁For iterative learning gain, beta₂The gain is feedback controlled.

5.3), when the road network does not enter the congestion state, setting the expected curve at the time as the accumulated vehicle number at the fixed time; when the road network accumulated vehicle number exceeds the critical accumulated vehicle number, setting the expected curves as the critical accumulated vehicle number;

5.4) and setting each boundary intersection as two phases, namely: east-west going straight + left-turn phase, north-south going straight + left-turn phase, as shown in fig. 5. The original fixed timing scheme is adopted in the first iteration, the error between the first iteration and the accumulated vehicle number collected last time can be made from the second iteration, the control input of the current moment is calculated, the green light duration of each phase of the boundary intersection is set in combination with the asymmetric signal timing strategy, as shown in figure 6, namely, the current sub-intersection is startedWhen the zone is in an oversaturated state, the green light time increment g calculated according to iterative learning control_k+1(t)＝g_k(t) + Δ u limiting the vehicle from entering the sub-zone by turning off the green light entering the sub-zone in advance, and reserving the phase of the sub-zone allowing the vehicle to drive away as a transition phase until the next phase starts;

5.5) in the current control period, the phase green signal ratio of each signal period is kept unchanged, and when the next control period is reached, the green signal ratio is adjusted according to the green signal ratio calculated in the next control period.

While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.