A coordinated control method for urban regional road traffic based on road network sub-division

Abstract

The present invention discloses a coordinated control method for urban regional road traffic based on road network subdivision. The method comprises setting a periodic duration for detecting congestion on urban regional roads and acquiring traffic data for the urban regional roads according to the periodic duration; acquiring traffic data for the urban regional roads once and dividing the data into k road network subdivisions; determining the congestion level of each road network subdivision based on the traffic data; constructing a multi-objective function and multi-constraint function model for each road network subdivision; and sequentially generating initial coordinated control schemes for the k road network subdivisions to form a coordinated control scheme for urban regional road traffic in this round. The method can disperse congested traffic flows within the region as much as possible and evacuate them out of the region, thereby avoiding or delaying the occurrence of urban regional road congestion. The mitigation effect is superior to that of timed control and coordinated control without road network subdivision.

Description

Technical Field

[0001] The present invention relates to a traffic coordination control method, and in particular to an urban area road traffic coordination control method based on road network sub-area division. Background Art

[0002] Urban road traffic signal control is categorized into single-point signal control, arterial signal control, and regional signal control. For single-point urban road traffic signal control, the signal cycle and signal-to-green ratio need only be adjusted based on traffic flow. For coordinated control of regional urban road traffic, unified signal management of all intersections is required to alleviate congestion. Therefore, coordinated control of regional urban road traffic can be viewed as a multi-objective function optimization problem. Multi-objective function optimization solutions can be categorized into two main types: traditional optimization algorithms and intelligent optimization algorithms. Traditional optimization algorithms include weighted methods, constraint methods, and linear programming.

[0003] However, these methods all have some drawbacks, such as:

[0004] (1) The coordinated control optimization problem of urban area roads requires the simultaneous optimization of control parameters, which is difficult and challenging to solve using traditional methods.

[0005] (2) Coordinated control of urban road traffic based on multi-objective function optimization or genetic algorithms does not take into account the congestion distribution within urban road traffic and the division of road network sub-areas, or is limited to the optimization of genetic algorithms and ignores the essential purpose of coordinated control to alleviate urban road traffic congestion.

[0006] Due to the complex distribution of congestion within urban road traffic, it is necessary to consider not only the congestion relief problem in local areas, but also the traffic flow evacuation in the global area. This requires the formulation of different coordinated control strategies for sub-areas with different congested road conditions within urban road traffic.

[0007] Name explanation:

[0008] Intersection entry: The set of roads from which traffic flows toward an intersection.

[0009] Exit intersection: The set of roads where traffic flows out of the intersection.

[0010] Intersection: includes the collection of all entrances and exits at the intersection;

[0011] Boundary intersection: An intersection located at the boundary of a road network sub-area.

[0012] Internal intersection: An intersection located inside a road network sub-area.

[0013] Traffic Control Scheme: A coordinated control scheme for traffic flows within a road network, typically implemented by an urban traffic control system. An urban traffic control system connects all intersection traffic signal controllers, variable lane signals, pedestrian crossing signals, variable traffic signs, and traffic guidance information display screens within a city or regional road network to an urban traffic control center via data transmission links. Computers at the control center coordinate and control traffic flows within the road network from the perspective of overall system optimization. In the present invention, when no regulation is performed on a road network sub-area, the existing urban traffic control system's traffic control scheme, also referred to as the existing traffic control scheme, is used. Summary of the Invention

[0014] The purpose of the present invention is to provide a method for coordinating and controlling urban area road traffic based on road network sub-area division to solve the above problems, which not only considers the congestion relief problem in local areas, but also takes into account the traffic flow evacuation in local areas.

[0015] To achieve the above-mentioned object, the technical solution adopted by the present invention is as follows: a method for coordinated control of urban area road traffic based on road network sub-area division, comprising the following steps:

[0016] (1) Setting a cycle duration for detecting congestion status of urban area roads and obtaining traffic data of urban area roads according to the cycle duration;

[0017] (2) Obtain traffic data of urban area roads and divide them into road network sub-areas to obtain k road network sub-areas, which are marked as V1~V k ;

[0018] (3) determining the traffic congestion level of each road network sub-area based on the traffic data thereof, wherein the traffic congestion level includes at least unimpeded traffic, light congestion, moderate congestion, and heavy congestion;

[0019] (4) Construct a coordinated control model for the road network sub-area. The coordinated control model is a multi-objective function and multi-constraint function model with two objective functions: maximizing the regional road traffic output capacity and minimizing the regional road traffic delay, and with signal cycle time, green light time, and signal light phase difference as constraints.

[0020] (5) Generate initial coordinated control schemes for the k road network sub-areas in turn, and combine them to form an initial urban area road traffic coordinated control scheme, including steps (51)-(53);

[0021] (51) Obtain the traffic congestion level of V1;

[0022] If the traffic is smooth, the original traffic control plan of the road network sub-area is maintained and used as the initial coordinated control plan x1 of V1;

[0023] If it is other levels, the traffic data of V1 is input into the coordinated control model, and the optimal solution is obtained by using the genetic algorithm with non-inferior solution sorting, which is used as the initial coordinated control solution x1 of V1;

[0024] (52) Process V2~V in sequence according to step (51) k , and obtain the corresponding initial coordinated control scheme x2~x k ;

[0025] (53) Construct the initial urban area road traffic coordination control plan x = {x1, x2, ..., x k}, and control each network sub-area accordingly;

[0026] (6) After the cycle duration is reached, the traffic data of the urban area roads is obtained again, and the road network sub-areas are re-divided. For each road network sub-area, a coordinated control plan for this round is generated, and the coordinated control plan for the urban area road traffic in this round is combined to form the coordinated control plan for this round, including steps (61)-(63);

[0027] (61) Obtain the traffic congestion level of V1;

[0028] If it is unobstructed, the coordinated control scheme of the previous round V1 is used as the coordinated control scheme x'1 of this round V1;

[0029] If it is other levels, determine whether the traffic congestion level has changed;

[0030] If there is no change, the coordinated control scheme of the previous round V1 is used as the coordinated control scheme x'1 of this round V1;

[0031] Otherwise, the traffic data of this round V1 is input into the coordinated control model, and the optimal solution is obtained by using the genetic algorithm with non-inferior solution sorting as the coordinated control solution x'1 of this round V1;

[0032] (62) Process V2~V in sequence according to step (61) k , get the corresponding control scheme x'2~x' k ;

[0033] (63) The coordinated control scheme for urban area road traffic in this round is x' = {x'1, x'2, ... x' k}, and regulate each network sub-area accordingly;

[0034] (7) Repeat step (6) to obtain the urban area road traffic coordination control plan corresponding to each cycle duration.

[0035] Preferably, in step (4), the objective function of the model is:

[0036] Maximize regional road traffic output capacity:

[0037]

[0038] Minimize regional road traffic delays:

[0039]

[0040] In formula (1), Q is the traffic output capacity of the road network sub-area, n_merge is the number of intersections at the boundary of the road network sub-area, and ψ_merge represents the number of phases of traffic flow output from the intersections at the boundary of the road network sub-area. represents the vehicle saturation at intersection j in phase i; g i represents the effective green light time of phase i at intersection j; C j represents the signal period of intersection j; Indicates the right of way at intersection j at phase i. It is 1 when passage is allowed and 0 when passage is prohibited. j is the intersection index and i is the phase index.

[0041] In formula (2), represents the average delay in the direction of phase i of the intersection at the boundary of the road network sub-area, is the traffic flow corresponding to ψ, represents the average delay in the j direction of the intersection ψ phase within the road network sub-area, is the corresponding traffic flow, n represents the number of intersections at the boundary of the road network sub-area, m represents the number of intersections within the road network sub-area, x and y are variables, x = 1 ~ n, y = 1 ~ m;

[0042] The constraints satisfied by the model are:

[0043] Signal cycle time constraint: The intersection traffic lights within the road network sub-area are synchronized with a common signal cycle length, satisfying the following equations (3) and (4);

[0044]

[0045] C min ≤C pub ≤C max (4)

[0046] In formula (3), C pub Indicates the duration of the common signal cycle, C ψ represents the cycle duration of the intersection ψ, represents the green light time of intersection ψ at phase i, L ψrepresents the total loss time within the cycle of the intersection, n1 represents the number of phases of the intersection ψ, in formula (4), C min and C max The minimum and maximum values ​​of the public signal cycle duration preset according to the road network sub-area;

[0047] Green light time constraint: satisfy the following formula (5);

[0048]

[0049] g min and g max The minimum and maximum values ​​for the preset green light time;

[0050] Phase difference constraint, satisfying the following formula (6);

[0051] θ ψ1,ψ2 +θ ψ2,ψ1 =aC (6)

[0052] Among them, θ ψ1,ψ2 represents the downlink phase difference from intersection ψ1 to intersection ψ2, θ ψ2,ψ1 It represents the uplink phase difference from intersection ψ1 to intersection ψ2, where a is an integer. When a=1, the two intersections are adjacent. In this case, to ensure timely passage of vehicles, the phase difference should be less than the signal period.

[0053] Preferably, step (2) is specifically as follows:

[0054] (2-1) The urban area road traffic network is abstracted into a directed graph G = (V, E), where the intersection is abstracted into a node set V = {v1, v2, ..., v l ,…,v H}, the interval road is abstracted into a set of directed edges E = {e lh}, 1≤l, h≤H, e lh represents a directed edge from node l to node h;

[0055] (2-2) Calculate the congestion index of each directed edge and construct the congestion index set E = {ec lh}, 1≤l, h≤H, ec lh For directed edge e lh Congestion index;

[0056] (2-3) Calculate the congestion index of each node based on the congestion index of the directed edge, and construct the node congestion index set VC = {vc1, vc2, ..., vc l ,…,vc H}, vc lis the congestion index of node l;

[0057] (2-4) Rearrange the nodes in the node set V and the node congestion index set VC in descending order of the node congestion index to obtain a new node set V' = {v'1, v'2, ..., v' l ,…,v' H}, and the node congestion index set VC'={vc'1,vc'2,…,vc' l ,…,vc' H};

[0058] (2-5) Initialize the number of road network sub-areas k = 0;

[0059] (2-6) Let k = k + 1, create two empty sets, which are partition sets V k , temporary set V k ';

[0060] (2-7) The kth node v' in V' k Divide into set V k In, as V k The source point and initialize V k ';

[0061] (2-8) Create a set S containing v' k For each node in S, determine whether it belongs to V′. If it belongs to V′, add it to V k and V k ′ and subtract it from V′;

[0062] (2-9) Process V in sequence k ′, the processing method of one node p is:

[0063] (a1) Create a set S' containing all one-hop neighbor nodes of p. For each node in S', determine whether it belongs to V'. If it belongs to V' and its congestion index is not lower than that of the source node, add it to V. k and V k ', and subtract the node from V', from V k 'Subtract node p from it;

[0064] (a2) Determine V k ' and V' are empty sets, if V k If V′ is not an empty set, repeat step (91). If V′ is not an empty set, repeat step (6).

[0065] (2-10) The division is completed, and the set V of k road network sub-areas is obtained. p ={V1,V2,...,Vk}.

[0066] Preferably, the traffic congestion level of the road network sub-area in step (3) is manually calibrated according to the on-site conditions.

[0067] Preferably, the traffic congestion level of the road network sub-area in step (3) is determined by the source point of the road network sub-area: a congestion index of [0, 0.4) indicates smooth traffic, a congestion index of [0.4, 0.6) indicates mild congestion, a congestion index of [0.6, 0.8) indicates moderate congestion, and a congestion index of [0.8, 1] indicates severe congestion.

[0068] Regarding the coordinated control model, maximizing the regional road traffic output capacity and minimizing the regional road traffic delay are adopted as two objective functions, with signal cycle time, green light time, and signal light phase difference as constraints.

[0069] Maximizing the output capacity of regional road traffic refers to the number of vehicles that drive out of the regional road traffic boundary within a specified time range. When the demand for road traffic in an urban area gradually increases, it is necessary to promptly increase the traffic capacity of the intersections where traffic bottlenecks first appear or may appear. For regional roads or intersections where congestion occurs for the first time, if they are not promptly controlled and relieved, the traffic state will gradually become unbalanced. When the number of vehicles on the congested section reaches a certain threshold, if the traffic flow at the upstream intersection continues to be released, it will inevitably lead to queuing, affecting the traffic flow at the upstream intersection, causing the unstable traffic flow state to spread to the upstream adjacent intersections, thereby causing the urban area road traffic to fall into more serious traffic congestion. Therefore, when urban area road traffic is about to become congested or has already experienced mild congestion, it is necessary to first ensure that the output capacity of the regional road traffic is maximized. The setting for maximizing the output capacity of regional road traffic is shown in formula (1).

[0070] Minimizing regional road traffic delay refers to the sum of the time it takes for all vehicles in regional road traffic to arrive at and leave an intersection within a specified time range. Under congested urban regional road traffic conditions, due to the different and limited traffic resources available at each intersection, the phase difference between intersections affects the release order of same-direction vehicles between intersections. It is the most direct influence on the changes in traffic flow density between sections in urban regional road traffic and is also a key factor in achieving regional coordinated control between intersections in urban regional road traffic. Therefore, when urban regional road traffic is congested, it is also necessary to ensure that the traffic delay of the regional road traffic is minimized. In addition, urban regional road traffic delay includes regional boundary entrance delay and regional internal delay. Therefore, the setting for minimizing regional road traffic delay is shown in formula (2).

[0071] Signal cycle time: Under traffic congestion in urban areas, the coordinated control scheme requires unified management of the traffic lights at all intersections in the area. Generally, it is required that the traffic lights at intersections within the area be synchronized with a common signal cycle time C. pub , the unit is seconds, the public signal cycle length should be set within a reasonable range, C min ≤C pub ≤C max , C min and C max is a constant and is set according to the road traffic status of the road network. The signal cycle time is set as shown in formula (3) (4).

[0072] Green light time: When setting the signal cycle time, the effective green light time of each signal light in the intersection is also set within the reasonable effective green light time range. In the present invention, g min and g max It is also a constant and is set according to the road traffic status of the road network. The specific setting is shown in formula (5).

[0073] Signal light phase difference: Phase difference is used to link independent intersection signal lights in urban road traffic, ensuring that intersections on the same main road form a road traffic green wave belt, so that vehicles starting from a certain traffic direction can pass smoothly and continuously. The green light display at each intersection on the same main road has a certain time offset, which is recorded as phase difference θ. In the coordinated control of urban road traffic, the phase difference is closely related to the spacing between intersections. The phase difference constraint needs to consider the upward and downward coordination phase. The sum of the coordinated phase differences in the upward and downward directions of the intersection should be a multiple of the signal period. The specific setting of the signal light phase difference is shown in formula (6).

[0074] The basic ideas of the present invention mainly include the following three points:

[0075] (1) The urban area road traffic conditions are divided into road network sub-areas according to the congestion level, and coordinated control optimization is performed for each road network sub-area. The coordinated control strategies of each road network sub-area are combined to form a coordinated control strategy for the global area and iterative optimization is performed.

[0076] (2) Maximizing the output capacity of regional road traffic and minimizing regional road traffic delays are the two goals of coordinated control of urban regional road traffic. The purpose is to disperse the congested traffic flow within the urban regional road traffic within the region and evacuate it to the outside of the region as much as possible.

[0077] (3) The coordinated control of urban area road traffic is regarded as a multi-objective function model, which is optimized and solved using genetic algorithms. At the same time, different weights are assigned to the objective functions and applied to different road traffic control decisions.

[0078] Compared with the prior art, the advantages of the present invention are:

[0079] (1) A new coordinated control method for urban regional road traffic based on road network sub-area division is proposed. According to the congestion distribution within the urban regional road traffic, the coordinated control strategies of the road network sub-areas with different congested road conditions are unified into a coordinated control strategy for the global region and iteratively optimized. The coordinated control of each congested road network sub-area is regarded as a multi-objective function model, and the maximization of regional road traffic output capacity and the minimization of regional road traffic delay are taken as the two objectives of the coordinated control multi-objective function. The genetic algorithm is used to optimize and solve it, and the congested traffic flow within the urban regional road traffic is dispersed within the region as much as possible, and evacuated to the outside of the region, so that the traffic flow is evacuated outward layer by layer from the road network sub-area to the global region, so as to achieve the purpose of avoiding or delaying the occurrence of urban regional road traffic congestion.

[0080] (2) A simulation experiment comparing the effectiveness of the proposed method was conducted. The experimental results show that, under the premise of road network subdivision, the proposed method has a better congestion relief effect than timed control and coordinated control without road network subdivision. In addition, the congestion relief effect under mild congestion conditions in urban areas is more effective than that under moderate and severe congestion conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Figure 1 Flowchart of the present invention;

[0082] Figure 2 Schematic diagram of finding the optimal solution for road network sub-areas using a genetic algorithm that sorts non-inferior solutions;

[0083] Figure 3 After abstracting the urban area roads into a directed graph in Example 2, a congestion index diagram of each directed edge is obtained;

[0084] Figure 4 Based on Figure 3 Schematic diagram of the traffic congestion level classification of road network sub-areas;

[0085] Figure 5 This is a comparison chart of average delay times at urban area road traffic control intersections in Example 2;

[0086] Figure 6 This is a comparison chart of average queue lengths at urban area road traffic control intersections in Example 2;

[0087] Figure 7 This is the typical regional road traffic network in Example 3;

[0088] Figure 8 The simulated data is entered over time Figure 7 The number of vehicles on the road network;

[0089] Figure 9 Comparison of urban area road traffic output capacity under the two methods in Example 3;

[0090] Figure 10 Comparison of average road delay times in urban areas under the two methods in Example 3. DETAILED DESCRIPTION

[0091] The present invention will be further described below with reference to the accompanying drawings.

[0092] Example 1: See Figure 1 、 Figure 2 A method for coordinated control of urban area road traffic based on road network sub-area division comprises the following steps:

[0093] (1) Setting a cycle duration for detecting congestion status of urban area roads and obtaining traffic data of urban area roads according to the cycle duration;

[0094] (2) Obtain traffic data of urban area roads and divide them into road network sub-areas to obtain k road network sub-areas, which are marked as V1~V k ;

[0095] (3) determining the traffic congestion level of each road network sub-area based on the traffic data thereof, wherein the traffic congestion level includes at least unimpeded traffic, light congestion, moderate congestion, and heavy congestion;

[0096] (4) Construct a coordinated control model for the road network sub-area. The coordinated control model is a multi-objective function and multi-constraint function model with two objective functions: maximizing the regional road traffic output capacity and minimizing the regional road traffic delay, and with signal cycle time, green light time, and signal light phase difference as constraints.

[0097] (5) Generate initial coordinated control schemes for the k road network sub-areas in turn, and combine them to form an initial urban area road traffic coordinated control scheme, including steps (51)-(53);

[0098] (51) Obtain the traffic congestion level of V1;

[0099] If the traffic is smooth, the original traffic control plan of the road network sub-area is maintained and used as the initial coordinated control plan x1 of V1;

[0100] If it is other levels, the traffic data of V1 is input into the coordinated control model, and the optimal solution is obtained by using the genetic algorithm with non-inferior solution sorting, which is used as the initial coordinated control solution x1 of V1;

[0101] (52) Process V2~V in sequence according to step (51) k , and obtain the corresponding initial coordinated control scheme x2~x k ;

[0102] (53) Construct the initial urban area road traffic coordination control plan x = {x1, x2, ..., x k}, and control each network sub-area accordingly;

[0103] (6) After the cycle duration is reached, the traffic data of the urban area roads is obtained again, and the road network sub-areas are re-divided. For each road network sub-area, a coordinated control plan for this round is generated, and the coordinated control plan for the urban area road traffic in this round is combined to form the coordinated control plan for this round, including steps (61)-(63);

[0104] (61) Obtain the traffic congestion level of V1;

[0105] If it is unobstructed, the coordinated control scheme of the previous round V1 is used as the coordinated control scheme x'1 of this round V1;

[0106] If it is other levels, determine whether the traffic congestion level has changed;

[0107] If there is no change, the coordinated control scheme of the previous round V1 is used as the coordinated control scheme x'1 of this round V1;

[0108] Otherwise, the traffic data of this round V1 is input into the coordinated control model, and the optimal solution is obtained by using the genetic algorithm with non-inferior solution sorting as the coordinated control solution x'1 of this round V1;

[0109] (62) Process V2~V in sequence according to step (61) k , get the corresponding control scheme x'2~x' k ;

[0110] (63) The coordinated control scheme for urban area road traffic in this round is x' = {x'1, x'2, ... x' k}, and regulate each network sub-area accordingly;

[0111] (7) Repeat step (6) to obtain the urban area road traffic coordination control plan corresponding to each cycle duration.

[0112] In step (4) of this embodiment, the objective function of the model is:

[0113] Maximize regional road traffic output capacity:

[0114]

[0115] Minimize regional road traffic delays:

[0116]

[0117] In formula (1), Q is the traffic output capacity of the road network sub-area, n_merge is the number of intersections at the boundary of the road network sub-area, and ψ_merge represents the number of phases of traffic flow output from the intersections at the boundary of the road network sub-area. represents the vehicle saturation at intersection j in phase i; g i represents the effective green light time of phase i at intersection j; C j represents the signal period of intersection j; Indicates the right of way at intersection j at phase i. It is 1 when passage is allowed and 0 when passage is prohibited. j is the intersection index and i is the phase index.

[0118] In formula (2), represents the average delay in the direction of phase i of the intersection at the boundary of the road network sub-area, is the traffic flow corresponding to ψ, represents the average delay in the j direction of the intersection ψ phase within the road network sub-area, is the corresponding traffic flow, n represents the number of intersections at the boundary of the road network sub-area, m represents the number of intersections within the road network sub-area, x and y are variables, x = 1 ~ n, y = 1 ~ m;

[0119] The constraints satisfied by the model are:

[0120] Signal cycle time constraint: The intersection traffic lights within the road network sub-area are synchronized with a common signal cycle length, satisfying the following equations (3) and (4);

[0121]

[0122] C min ≤C pub ≤C max (4)

[0123] In formula (3), C pub Indicates the duration of the common signal cycle, C ψ represents the cycle duration of the intersection ψ, represents the green light time of intersection ψ at phase i, L ψ represents the total loss time within the cycle of the intersection, n1 represents the number of phases of the intersection ψ, in formula (4), C min and C maxThe minimum and maximum values ​​of the public signal cycle duration preset according to the road network sub-area;

[0124] Green light time constraint: satisfy the following formula (5);

[0125]

[0126] g min and g max The minimum and maximum values ​​for the preset green light time;

[0127] Phase difference constraint, satisfying the following formula (6);

[0128] θ ψ1,ψ2 +θ ψ2,ψ1 =aC (6)

[0129] Among them, θ ψ1,ψ2 represents the downlink phase difference from intersection ψ1 to intersection ψ2, θ ψ2,ψ1 It represents the uplink phase difference from intersection ψ1 to intersection ψ2, where a is an integer. When a=1, the two intersections are adjacent. In this case, to ensure timely passage of vehicles, the phase difference should be less than the signal period.

[0130] Step (2) is specifically as follows:

[0131] (2-1) The urban area road traffic network is abstracted into a directed graph G = (V, E), where the intersection is abstracted into a node set V = {v1, v2, ..., v l ,…,v H}, the interval road is abstracted into a set of directed edges E = {e lh}, 1≤l, h≤H, e lh represents a directed edge from node l to node h;

[0132] (2-2) Calculate the congestion index of each directed edge and construct the congestion index set E = {ec lh}, 1≤l, h≤H, ec lh For directed edge e lh Congestion index;

[0133] (2-3) Calculate the congestion index of each node based on the congestion index of the directed edge, and construct the node congestion index set VC = {vc1, vc2, ..., vc l ,…,vc H}, vc l is the congestion index of node l;

[0134] (2-4) Rearrange the nodes in the node set V and the node congestion index set VC in descending order of the node congestion index to obtain a new node set V' = {v'1, v'2, ..., v' l ,…,v' H}, and the node congestion index set VC'={vc'1,vc'2,…,vc' l ,…,vc' H};

[0135] (2-5) Initialize the number of road network sub-areas k = 0;

[0136] (2-6) Let k = k + 1, create two empty sets, which are partition sets V k , temporary set V′ k ;

[0137] (2-7) The kth node v' in V' k Divide into set V k In, as V k The source point and initialize V′ k ;

[0138] (2-8) Create a set S containing v' k For each node in S, determine whether it belongs to V′. If it belongs to V′, add it to V k and V′ k and subtract it from V′;

[0139] (2-9) Process V′ in sequence k Among the nodes, the processing method of one node p is:

[0140] (a1) Create a set S' containing all one-hop neighbor nodes of p. For each node in S', determine whether it belongs to V'. If it belongs to V' and its congestion index is not lower than that of the source node, add it to V. k and V′ k and subtract the node from V′, from V′ k Subtract node p from

[0141] (a2) Determine V′ k and V′ is an empty set, if V′ k If V′ is not an empty set, repeat step (91). If V′ is not an empty set, repeat step (6).

[0142] (2-10) The division is completed, and the set V of k road network sub-areas is obtained. p ={V1,V2,...,V k}.

[0143] In step (3), the traffic congestion level of the middle road network sub-area is manually calibrated according to the on-site conditions.

[0144] The traffic congestion level of the road network sub-area in step (3) is determined by the source point of the road network sub-area: a congestion index of [0, 0.4) indicates smooth traffic, a congestion index of [0.4, 0.6) indicates mild congestion, a congestion index of [0.6, 0.8) indicates moderate congestion, and a congestion index of [0.8, 1] indicates severe congestion.

[0145] In the present invention, regarding the use of a genetic algorithm with non-inferior solution sorting to solve the optimal solution: the urban area road traffic coordinated control designed by the present invention is actually a problem of multi-objective optimization of the signal control of all intersections in the urban area road traffic.

[0146] In multi-objective optimization problems, improving one objective may lead to performance degradation of another. This is particularly true in coordinated urban traffic control, where signal cycle duration, green light timing, and phase difference, as key decision variables, are highly correlated. Solving this problem requires coordinating these various objectives. To simultaneously and coordinatively optimize the overall control parameters of urban traffic, find a better multi-objective optimization model, and obtain a comprehensive optimal solution, it is necessary to define the superiority of the multi-objective solution and clearly identify the corresponding dominant objective.

[0147] For general multi-objective optimization problems, Let be the objective function of the multi-objective optimization problem, is the feasible domain of the model, for the solution x1∈X f and x2∈X f , if and only if If the multi-objective optimization problem has a solution x * ∈X f , and x * Better than feasible region X f For all other solutions in , we consider x * is the optimal solution to the problem, also known as the Pareto optimal solution.

[0148] In multi-objective optimization problems, achieving optimal solutions for all objectives simultaneously is often difficult. Therefore, the Pareto optimal solution is often used as the optimal solution. During the specific solution process, the top-ranked non-inferior solutions in each generation are memorized. When the number of saved non-inferior solutions reaches a threshold during the iterative solution process, all non-inferior solutions are sorted again according to their fitness values ​​to obtain the final optimal solution.

[0149] In the coordinated control algorithm for urban regional road traffic based on road network subdivision, each solution consists of three parts: common cycle time, green signal ratio of each intersection, and phase of each intersection. In this paper, a genetic algorithm is used to optimize and solve the multi-objective function of coordinated control for urban regional road traffic.

[0150] In the process of solving multi-objective function optimization using genetic algorithm, the parameters of adaptive adjustment are mainly aimed at the crossover probability P c and mutation probability P m The genetic algorithm selects individuals with high fitness through the selection algorithm and passes them on to the next generation. The selection algorithm in this paper adopts the roulette method, so the fitness function value is F k The probability of an individual (x) being selected is shown in formula (7).

[0151]

[0152] For the crossover and mutation operators, the crossover operation adopts the single-point crossover method, and the mutation operation is to eliminate the poor individuals as much as possible. The mutation operation adopts the basic bit mutation method, and after determining the node according to the mutation probability, the gene transformation is performed to form a new individual, as shown in formula (8) and formula (9).

[0153] Among them, F max is the maximum fitness in the contemporary population; F′ is the larger fitness of the two crossovers; F is the contemporary average fitness; F(x k ) is the current individual fitness; N is the chromosome length; K1, K2, K3, and K4 are adjustment coefficients. To avoid falling into a local optimal solution during the solution process, K1 = 1, K2 = 1, K3 = 0.5, and K4 = 0.5 are generally used.

[0154]

[0155]

[0156] To speed up the iteration process of the optimization solution, after calculating the sub-objective function values ​​for all individuals in the population, this paper sorts the individuals by fitness, recording the top N individuals with the highest fitness values. Crossover and mutation are then performed to obtain a new population of M individuals. The top N individuals are then merged with the new M individuals to obtain a population of N+M individuals. Finally, the new population is sorted by fitness, and the bottom N individuals are eliminated. The resulting new population is then iterated and optimized. The main process is as follows:

[0157] Step 1: Obtain urban area road traffic network data, initialize and generate a first generation population with a population size of M according to the coordinated control timing plan of each intersection, set the population memory size N and the calculation parameters K of the crossover probability and mutation probabilityi (i=1,2,3,4).

[0158] Step 2: Initialize the evolutionary generation Gen=1 and generate the initial generation population P(Gen).

[0159] Step 3: Calculate the fitness function of the current algebraic population P(Gen) and remember the first N individuals.

[0160] Step 4: Perform a roulette wheel selection operation on the population, select individuals with higher fitness function values ​​and pass them on to the next generation to obtain the population P1(Gen).

[0161] Step 5: Use the single-point crossover method to perform a crossover operation on any two individuals to generate new individuals and obtain the population P2(Gen).

[0162] Step 6: Use the basic bit mutation method to mutate the population and update the population to obtain P3(Gen).

[0163] Step 7: Merge the population P3(Gen) and the first N individuals in memory into a new population P4(Gen). The population size is now N+M.

[0164] Step 8: Determine whether the optimization conditions (convergence condition, whether the maximum number of generations of population evolution has been reached) are met. If the conditions are met, proceed to the next step. If not, take the top M individuals in the population P4(Gen) to form the next generation of evolutionary population P(Gen), Gen = Gen + 1, and proceed to step 3.

[0165] Step 9: Obtain the current optimal timing plan, including the phase time, signal cycle and phase difference of each intersection.

[0166] Example 2: See Figures 1 to 6 ,Based on Example 1, we use SUMO traffic simulation tool to build a 4×4 urban area road traffic network, such as Figure 3 As shown in Figure 1, the urban area road traffic coordination control network has a total of 16 intersections, that is, H = 16, consisting of 16 nodes v1~v 16 The diagram shows the components, marked sequentially with circles and numbers within them. In the diagram, the line between two nodes is a directed edge, and the arrow indicates the direction of the directed edge, which also represents the direction of vehicle travel on that road segment. Each directed edge has been assigned a different congestion index, specifically the number marked on the directed edge. This congestion index is based on existing technology.

[0167] Then, the road network sub-area division is performed according to step (2) of the present invention, and a set of 6 road network sub-areas V is obtained. p= {V1, V2, V3, V4, V5, V6}. In this embodiment, the division of the road network sub-areas refers to the steps (2-1) to (2-10). The 6 road network sub-areas are: V1 = {v6, v2, v5, v7, v 10 ,v9}, V2={v1}, V3={v4,v3,v8}, V4={v 11 ,v 12 ,v 15}、V5={v 13 ,v 14}、V6={v 16}.

[0168] Then, the traffic congestion level of each road network sub-area is determined according to step (3) of the present invention. In this embodiment, it is determined based on the source point of the road network sub-area. The source point of V1 is v6. In the figure, the congestion index of v6 is 0.84, which is a severe congestion. The source point of V2 is v1. In the figure, the congestion index of v1 is 0.66, which is a moderate congestion. By analogy, we can obtain the traffic congestion level of each road network sub-area, V1 = {v6, v2, v5, v7, v 10 ,v9} is a heavy congestion area, V2={v1} and V3={v4,v3,v8} are moderate congestion areas, V4={v 11 ,v 12 ,v 15} is a lightly congested area, V5={v 13 ,v 14} and V6={v 16} is a smooth area.

[0169] Then, a coordinated control model is constructed according to step (4) of the present invention, and an initial coordinated control scheme is generated according to step (5) of the present invention.

[0170] In order to illustrate how the initial coordinated control scheme is generated in step (5) of the present invention, we Figure 3 The urban roads shown are simulated using traffic data from the urban roads. This data includes, but is not limited to, hourly traffic flow tables for each direction at each intersection and initial phase settings for each intersection. The hourly traffic flow tables for each direction at each intersection are based on the traffic flow at the entrance and exit lanes of each intersection during the transition period from off-peak to peak traffic hours. See Table 1 for details:

[0171] Table 1. Hourly traffic flow in each direction at each intersection of urban roads

[0172]

[0173] The signal cycle of each intersection in the urban area road traffic network is set to 144 seconds. The initial phase data settings of each intersection are shown in Table 2, and Table 2 can be regarded as the original traffic control plan for the urban area roads.

[0174] Table 2: Initial phase settings for each intersection of urban roads

[0175]

[0176]

[0177] First, according to steps (1) to (3) in Example 2, the traffic congestion level of the road network sub-area and each road network sub-area is obtained. Then, after constructing the coordinated control model according to step (4), combined with the data in Table 1 and Table 2, the initial coordinated control scheme can be generated according to step (5) of the present invention. Specifically:

[0178] Except for V5 and V6, which are unobstructed and do not require regulation, the rest of the road network sub-areas need to be adjusted. That is to say, for V1-V4, the corresponding traffic data must be input into the coordinated control model respectively, and the optimal solution is solved by the genetic algorithm with non-inferior solution sorting. In the present invention, in the process of optimizing and solving the multi-objective function of regional road traffic, the basic parameters of the genetic algorithm are uniformly set, where M=100, N=10; the termination iteration condition of the genetic algorithm is Gen=50 or the optimal individual fitness loss obtained in 20 consecutive generations is less than ε=0.001; the traffic lane saturation rate is 1500pcu / h; the signal phase switching time is 5s, the yellow light time is 3s, and the red light time is 2s; the intersection signal cycle is set to 120s≤C pub ≤230s; Phase 1 setting follows g min =45s,g max =75s, Phase 1 setting follows g min =15s,g max =40s, Phase 3 setting follows g min =45s,g max =75s; Phase 4 setting follows g min =15s,g max =40s.

[0179] Finally, each coordinated control scheme is obtained and merged into the urban area road traffic coordinated control scheme, as shown in Table 3 below:

[0180] Table 3 Urban area road traffic coordination control plan

[0181]

[0182]

[0183] Based on the method of the present invention, the overall effect of the output capacity, traffic delay and queue length of regional road traffic is compared with that of single-point timing control and urban regional road traffic coordinated control without road network sub-division, and the results are shown in Table 4:

[0184] Table 4 Comparison of overall effects of urban area road traffic control schemes

[0185] Control scheme Output capacity Traffic delays queue length Single point timing control 10543 421450 6280 Coordinated control without road network subdivision 12822 392530 5458 The present invention 14269 364590 4846

[0186] As can be seen from Table 4, under the same urban area road traffic conditions, the method of the present invention is superior to single-point timing control and urban area road traffic coordinated control without road network sub-division in terms of the three indicators of regional road traffic output capacity, traffic delay and queue length.

[0187] In addition, we compared the average delay time and average queue length of each intersection of urban area road traffic using the method of the present invention with single-point timing control and urban area road traffic coordinated control without road network sub-division. The simulation results of 16 intersections are shown in Tables 5, 6, Figure 5 、 Figure 6 shown.

[0188] Table 5 Comparison of average delay time at intersections of urban area road traffic control schemes

[0189]

[0190] Table 6 Comparison of average queue lengths at intersections of urban area road traffic control schemes

[0191]

[0192] From Table 5, Table 6, Figure 5 、 Figure 6 As can be seen from the figure, for each intersection in the urban road network, the regional coordinated control method achieves improved performance in terms of average delay time and average queue length compared to the single-point timing control method. For coordinated control based on network subdivision, intersections 2, 5, 7, and 9 achieve the best congestion relief results, with significant reductions in average delay time and average queue length, significantly alleviating congestion. Although some lightly congested areas carry some traffic flow from moderately and severely congested areas, resulting in reduced delay time and queue length in these areas compared to coordinated control without network subdivision, the coordinated control of urban road traffic based on network subdivision further refines the road network, resulting in longer signal cycle times and effective green light times in moderately and severely congested areas, thereby achieving better congestion relief for the entire urban area.

[0193] Example 3: See Figure 7-10 We conducted simulation experiments on the effects of alleviating congestion in different areas. We used the SUMO traffic simulation tool to construct a typical regional road traffic network consisting of four intersection nodes, such as Figure 7 As shown in the figure, the direction of the arrows indicates the driving direction of vehicles in each road section.

[0194] We regard the typical regional road traffic network as a region, which has a total of 8 traffic input lanes and 8 traffic output lanes. The length of the edge roads and internal roads is set to 150 meters. The vehicle generation speed at the edge intersection is 1500 vehicles / hour. The OD matrix of the traffic flow entering the regional road is evenly distributed to ensure that each road and each intersection in the region is at the same congestion level. The initial congestion level of the region is smooth, and all surrounding intersections and roads are heavily congested, that is, the traffic flow entering the regional road is greater than the traffic flow leaving. The regional road traffic will experience a change process from smooth to heavily congested, such as Figure 8 shown.

[0195] because Figure 7 The area shown is relatively simple. We no longer divide it into sub-areas, but directly input the traffic data of the area and use the coordinated control model of the present invention to obtain a control solution.

[0196] Taking the regional road traffic output capacity and average delay time as evaluation indicators, the method of the present invention and the single-point timing control method are used to evaluate the Figure 8 The traffic conditions shown in the figure are regulated, and the comparison of urban area road traffic output capacity under the two methods is obtained as shown in the figure below. Figure 9 As shown in the figure below; and the comparison with the average delay time of urban area roads is shown in the figure below Figure 10 As shown. Figure 9 and Figure 10 It can be seen that in terms of the two evaluation indicators of regional road traffic output capacity and average waiting time, the regional road traffic coordination method proposed in the present invention is significantly better than the single-point timing control method.

[0197] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for coordinated control of urban area road traffic based on road network sub-area division, characterized in that: The following steps are involved: (1) Setting a cycle duration for detecting congestion status of urban area roads and obtaining traffic data of urban area roads according to the cycle duration; (2) Obtain traffic data of urban area roads and divide them into road network sub-areas to obtain k road network sub-areas, which are marked as V1~V k ; (3) determining the traffic congestion level of each road network sub-area based on the traffic data thereof, wherein the traffic congestion level includes at least unimpeded traffic, light congestion, moderate congestion, and heavy congestion; (4) Construct a coordinated control model for the road network sub-area. The coordinated control model is a multi-objective function and multi-constraint function model with two objective functions: maximizing the regional road traffic output capacity and minimizing the regional road traffic delay, and with signal cycle time, green light time, and signal light phase difference as constraints. (5) Generate initial coordinated control schemes for the k road network sub-areas in turn, and combine them to form an initial urban area road traffic coordinated control scheme, including steps (51)-(53); (51) Obtain the traffic congestion level of V1; If the traffic is smooth, the original traffic control plan of the road network sub-area is maintained and used as the initial coordinated control plan x1 of V1; If it is other levels, the traffic data of V1 is input into the coordinated control model, and the optimal solution is obtained by using the genetic algorithm with non-inferior solution sorting, which is used as the initial coordinated control solution x1 of V1; (52) Process V2~V in sequence according to step (51) k , and obtain the corresponding initial coordinated control scheme x2~x k ; (53) Construct the initial urban area road traffic coordination control plan x = {x1, x2, ..., x k }, and control each network sub-area accordingly; (6) After the cycle duration is reached, the traffic data of the urban area roads is obtained again, and the road network sub-areas are re-divided. For each road network sub-area, a coordinated control plan for this round is generated, and the coordinated control plan for the urban area road traffic in this round is combined to form the coordinated control plan for this round, including steps (61)-(63); (61) Obtain the traffic congestion level of V1; If it is unobstructed, the coordinated control scheme of the previous round V1 is used as the coordinated control scheme x'1 of this round V1; If it is other levels, determine whether the traffic congestion level has changed; If there is no change, the coordinated control scheme of the previous round V1 is used as the coordinated control scheme x'1 of this round V1; Otherwise, the traffic data of this round V1 is input into the coordinated control model, and the optimal solution is obtained by using the genetic algorithm with non-inferior solution sorting as the coordinated control solution x'1 of this round V1; (62) Process V2~V in sequence according to step (61) k , get the corresponding control scheme x'2~x' k ; (63) The coordinated control scheme for urban area road traffic in this round is x' = {x'1, x'2, ... x' k }, and regulate each network sub-area accordingly; (7) Repeat step (6) to obtain the urban area road traffic coordination control plan corresponding to each cycle duration.

2. The urban area road traffic coordination control method based on road network sub-division according to claim 1 is characterized in that: In step (4), the objective function of the model is: Maximize regional road traffic output capacity: Minimize regional road traffic delays: In formula (1), Q is the traffic output capacity of the road network sub-area, n_merge is the number of intersections at the boundary of the road network sub-area, and ψ_merge represents the number of phases of traffic flow output from the intersections at the boundary of the road network sub-area. represents the vehicle saturation at intersection j in phase i; g i represents the effective green light time of phase i at intersection j; C j represents the signal period of intersection j; φ i j Indicates the right of way at intersection j at phase i. It is 1 when passage is allowed and 0 when passage is prohibited. j is the intersection index and i is the phase index. In formula (2), represents the average delay in the direction of phase i of the intersection at the boundary of the road network sub-area, is the traffic flow corresponding to ψ, represents the average delay in the j direction of the intersection ψ phase within the road network sub-area, is the corresponding traffic flow, n represents the number of intersections at the boundary of the road network sub-area, m represents the number of intersections within the road network sub-area, x and y are variables, x = 1 ~ n, y = 1 ~ m; The constraints satisfied by the model are: Signal cycle time constraint: The intersection traffic lights within the road network sub-area are synchronized with a common signal cycle length, satisfying the following equations (3) and (4); C min ≤C pub ≤C max (4) In formula (3), C pub Indicates the duration of the common signal cycle, C ψ represents the cycle duration of the intersection ψ, represents the green light time of intersection ψ at phase i, L ψ represents the total loss time within the cycle of the intersection, n1 represents the number of phases of the intersection ψ, in formula (4), C min and C max The minimum and maximum values ​​of the public signal cycle duration preset according to the road network sub-area; Green light time constraint: satisfy the following formula (5); g min and g max The minimum and maximum values ​​for the preset green light time; Phase difference constraint, satisfying the following formula (6); i ψ1,ψ2 +θ ψ2,ψ1 =aC (6) Among them, θ ψ1,ψ2 represents the downlink phase difference from intersection ψ1 to intersection ψ2, θ ψ2,ψ1 It represents the uplink phase difference from intersection ψ1 to intersection ψ2, where a is an integer. When a=1, the two intersections are adjacent. In this case, to ensure timely passage of vehicles, the phase difference should be less than the signal period.

3. The urban area road traffic coordination control method based on road network sub-division according to claim 1 is characterized in that: Step (2) is specifically as follows: (2-1) The urban area road traffic network is abstracted into a directed graph G = (V, E), where the intersection is abstracted into a node set V = {v1, v2, ..., v l ,…,v H }, the interval road is abstracted into a set of directed edges E = {e lh }, 1≤l, h≤H, e lh represents a directed edge from node l to node h; (2-2) Calculate the congestion index of each directed edge and construct the congestion index set E = {ec lh }, 1≤l, h≤H, ec lh For directed edge e lh Congestion index; (2-3) Calculate the congestion index of each node based on the congestion index of the directed edge, and construct the node congestion index set VC = {vc1, vc2, ..., vc l ,…,vc H }, vc l is the congestion index of node l; (2-4) Rearrange the nodes in the node set V and the node congestion index set VC in descending order of the node congestion index to obtain a new node set V' = {v'1, v'2, ..., v' l ,…,v' H }, and the node congestion index set VC'={vc'1,vc'2,…,vc' l ,…,vc' H }; (2-5) Initialize the number of road network sub-areas k = 0; (2-6) Let k = k + 1, create two empty sets, which are partition sets V k , temporary set V′ k ; (2-7) The kth node v' in V' k Divide into set V k In, as V k The source point and initialize V′ k ; (2-8) Create a set S containing v' k For each node in S, determine whether it belongs to V′. If it belongs to V′, add it to V k and V′ k and subtract it from V′; (2-9) Process V′ in sequence k Among the nodes, the processing method of one node p is: (a1) Create a set S' containing all one-hop neighbor nodes of p. For each node in P', determine whether it belongs to V'. If it belongs to V' and its congestion index is not lower than that of the source node, add it to V. k and V′ k and subtract the node from V′, from V′ k Subtract node p from (a2) Determine V′ k and V′ is an empty set, if V′ k If V′ is not an empty set, repeat step (91). If V′ is not an empty set, repeat step (6). (2-10) The division is completed, and the set V of k road network sub-areas is obtained. p ={V1,V2,...,V k }.

4. The urban area road traffic coordination control method based on road network subdivision according to claim 1 is characterized in that: In step (3), the traffic congestion level of the middle road network sub-area is manually calibrated according to the on-site conditions.

5. The urban area road traffic coordination control method based on road network subdivision according to claim 3 is characterized in that: The traffic congestion level of the road network sub-area in step (3) is determined by the source point of the road network sub-area: a congestion index of [0, 0.4) indicates smooth traffic, a congestion index of [0.4, 0.6) indicates mild congestion, a congestion index of [0.6, 0.8) indicates moderate congestion, and a congestion index of [0.8, 1] indicates severe congestion.

Images