CN113192318A - Data drive control regional traffic signal dynamic optimization method and system - Google Patents

Data drive control regional traffic signal dynamic optimization method and system Download PDF

Info

Publication number
CN113192318A
CN113192318A CN202110128223.6A CN202110128223A CN113192318A CN 113192318 A CN113192318 A CN 113192318A CN 202110128223 A CN202110128223 A CN 202110128223A CN 113192318 A CN113192318 A CN 113192318A
Authority
CN
China
Prior art keywords
control
intersection
traffic
traffic flow
period
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110128223.6A
Other languages
Chinese (zh)
Other versions
CN113192318B (en
Inventor
陈家旭
齐行知
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Anhui Changtongxing Traffic Information Service Co ltd
Traffic Management Research Institute of Ministry of Public Security
Anhui Keli Information Industry Co Ltd
Original Assignee
Anhui Changtongxing Traffic Information Service Co ltd
Traffic Management Research Institute of Ministry of Public Security
Anhui Keli Information Industry Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Anhui Changtongxing Traffic Information Service Co ltd, Traffic Management Research Institute of Ministry of Public Security, Anhui Keli Information Industry Co Ltd filed Critical Anhui Changtongxing Traffic Information Service Co ltd
Priority to CN202110128223.6A priority Critical patent/CN113192318B/en
Publication of CN113192318A publication Critical patent/CN113192318A/en
Application granted granted Critical
Publication of CN113192318B publication Critical patent/CN113192318B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/07—Controlling traffic signals
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Abstract

The invention discloses a data-driven control regional traffic signal dynamic optimization method and a system, wherein the method comprises the following steps: analyzing a road network topological structure to obtain intersection attribute parameters in an intersection control area, wherein the intersection attribute parameters comprise traffic flow analysis and control channels; obtaining traffic monitoring state indexes of analyzed traffic flow and control channels based on lane level state monitoring and the intersection attribute parameters; constructing an intersection storage and forwarding state space equation meeting the signal control adaptation requirement based on the traffic detection state index and the signal control scheme; an iterative learning control model meeting the requirements of balance control and convergence is constructed, iterative optimization is carried out on the intersection storage and forwarding state space equation, and real-time traffic signal optimization control is achieved; the regional traffic signal dynamic optimization method can adapt to different signal control design schemes and demand change based on steering, realizes fine control of intersections, has high automation degree, and is not easy to overflow.

Description

Data drive control regional traffic signal dynamic optimization method and system
Technical Field
The invention relates to the technical field of regional traffic signal dynamic optimization, in particular to a data-driven control regional traffic signal dynamic optimization method and system.
Background
The dynamic optimization control of the traffic signals is a hot technology concerned by various large signal control manufacturers and researchers in China in recent years, and can adjust a signal control scheme in real time according to the traffic flow change trend, adapt to the traffic flow change condition in a control range, improve the traffic efficiency and reduce the driving delay. However, the dynamic optimization control technology suitable for regional traffic signals has shortcomings in research and application.
The realization forms of signal control systems in various regions in China are different, but the actual control effect is difficult to meet the control requirement in scene application by using a regional dynamic optimization control technology taking a traffic model and intelligent calculation control as a core due to model applicability and calculation capability constraint; the data driving control method which is not based on the traditional mechanism mathematical modeling, is not traffic prediction control and directly takes actual traffic flow input and output data as a timing optimization basis has lower requirements on model calibration and calculation capacity, has certain advantages on system stability and response capacity, and can meet the requirements on large-scale traffic signal dynamic optimization control.
However, the technical scheme of the existing application data driving control method has the defects of insufficient practicability and applicability in the field application: (1) only the optimization control of a given intersection and a signal control scheme is supported, the application scene is single, and the automation degree of the system is low; (2) the phase or the road section is used as a state index carrier, so that the fine control requirement of the intersection cannot be met, and the overflow problem is easy to generate in practical application.
Disclosure of Invention
Based on the technical problems in the background art, the invention provides a data-driven control regional traffic signal dynamic optimization method and system, which have high automation degree and are not easy to overflow.
The invention provides a dynamic optimization method of data-driven control regional traffic signals, which comprises the following steps:
analyzing a road network topological structure to obtain intersection attribute parameters in an intersection control area, wherein the intersection attribute parameters comprise traffic flow analysis and control channels;
obtaining a traffic detection state index of the intersection attribute parameters based on the lane level state monitoring and the intersection attribute parameters;
according to the traffic detection state index and the signal control scheme, constructing an intersection storage and forwarding state space equation meeting the signal control adaptation requirement;
and constructing an iterative learning control model meeting the requirements of balance control and convergence, and performing iterative optimization on the intersection storage and forwarding state space equation to realize the real-time optimization control of the traffic signals.
Further, the traffic detection state index for obtaining the intersection attribute parameter based on the lane-level state monitoring and the intersection attribute parameter includes:
converting traffic flow information of a lane level inside the intersection into traffic flow taking the traffic flow as a carrier, taking the traffic flow as the carrier as an input parameter under the control of a traffic signal, and acquiring the traffic flow information in real time through traffic detection equipment;
and obtaining the number of the queued vehicles of the control channel according to the lane detection queuing length, wherein the number of the queued vehicles is obtained in real time through traffic detection equipment.
Further, in the construction of an intersection store-and-forward state space equation meeting the signal control adaptation requirement according to the traffic detection state index of the intersection attribute parameter, the method comprises the following steps:
according to the store-and-forward model on the road section steering, the steering state equation belonging to the same control channel is accumulated to obtain a store-and-forward model based on the intersection control channel;
defining signal control elements by an intersection signal control scheme based on convergence constraint to obtain an intersection store-and-forward state space equation with an indefinite period;
and (4) considering the coordination control requirement, carrying out derivation and conversion of the given period and the green time loss time on the crossing store-and-forward state space equation with the indefinite period to obtain the crossing store-and-forward state space equation with the definite period.
Further, an iterative learning control model meeting the requirements of balance control and convergence is constructed, iterative optimization is carried out on the intersection storage and forwarding state space equation, and real-time traffic signal optimization control is realized, and the method comprises the following steps:
establishing an evaluation function taking the phase-to-phase queuing balance as a target, constructing a balance target function, and obtaining a P-type learning law based on ILC (constant-period balance control);
and constructing an iterative learning control model and a control flow based on a P-type learning law, a balanced objective function and a fixed-period intersection storage and forwarding state space equation.
Further, in the process of converting the traffic flow information of the lane level inside the intersection into the traffic flow taking the traffic flow as a carrier, the traffic flow of the intersection comprises an inlet traffic flow IAAnd exit traffic flow IBThe concrete formula is as follows;
for import traffic flow IA：
Figure RE-GDA0003120550820000031
Wherein q isXRepresenting the traffic volume of the respective steered traffic stream, X representing a steering attribute, qY,ZThe detected traffic volume of the Z-th lane of the guide Y is shown, Y is the lane guide attribute, alpha is the steering proportion of the composite guide straight to the left, beta is the steering proportion of the composite guide straight to the right,
Figure RE-GDA0003120550820000032
the steering proportion of the composite steering to the left and the right is shown, and gamma is the steering proportion of a three-way lane;
outlet traffic flow IBThe traffic flow is from the inlet traffic flow IAAnd (3) statistical acquisition:
(1) according to the intersection traffic flow data, a vehicle flow convergence relation matrix R ═ { R ═ R can be obtainedij|i∈[1,m],j∈[1,n]M denotes the set IBN represents the set IAThe elements of the matrix R satisfy:
Figure RE-GDA0003120550820000033
wherein pi (i) represents an import traffic flow set of an export traffic flow i;
(2) outlet traffic flow IBTraffic qBAnd import traffic flow IATraffic qASatisfies the following matrix relationship:
qB＝R×qA
wherein q isB＝{qn+1,qn+2,...,qn+m}T，qA＝{q1,q2,...,qn}T。
Further, in the storage and forwarding model based on the intersection control channel obtained by accumulating the steering state equations of the same control channel according to the storage and forwarding model on the road section steering, the storage and forwarding model on the road section steering and the storage and forwarding model based on the intersection control channel are obtained by the following formulas respectively:
the store-and-forward model on the road section steering corresponds to the following formula:
xz,l(k+1)＝xz,l(k)+λz,lT[qz(k)+εz(k)]-Tuz,l(k)
εz(k)＝dz(k)-sz(k)
wherein x isz,l(k +1) represents the number of queued vehicles on the road segment z out of turn l at the beginning of the (k +1) th time period, xz,l(k) Representing the number of queued vehicles on the road segment z out of turn/at the beginning of the kth time period; q. q.sz(k) Represents [ kT, (k +1) T]Traffic volume, e, during which the vehicle travels into the stretch zz(k) Represents [ kT, (k +1) T]Disturbance of the state of the section z itself during the period, λz,lIs the proportion of the steering that the vehicle on the observable stretch z is driving out of the steering; u. ofz,l(k) Is [ kT, (k +1) T]Traffic volume during which a section z is driven from a turn l, T is a statistical time interval, dz(k) Represents [ kT, (k +1) T]The demanded flow, s, generated by the section z itself during the periodz(k) Represents [ kT, (k +1) T]The periodic roadDissipation flow for section z;
the storage forwarding model based on the intersection control channel corresponds to the following formula:
Figure RE-GDA0003120550820000041
qz,l(k)＝λz,l(1+δz)qz(k)
Figure RE-GDA0003120550820000042
wherein x isN,f(k +1) represents the number of queued vehicles exiting the control lane N, f to turn to l at the beginning of the (k +1) th time period, U (N, f) represents the traffic flow at the upstream intersection of the control lane N, f, qz,l(k) Represents [ kT, (k +1) T]During which the demanded flow, u, of the traffic flow is driven out of the turn l of the stretch zz,l(k) Represents [ kT, (k +1) T]During which the dissipated flow, delta, of the traffic flow is driven out of the turn/of the stretch zzRepresenting the proportion of disturbance of the demanded flow on the road section z, qN,f(k) Represents [ kT, (k +1) T]During which the required flow, lambda, of the channel N, f is controlledz,lIs the proportion of the steering, δ, at which the vehicle on the stretch z can be seen to be driven out of the steering lU(N,f)Representing the disturbance proportion q generated by the transmission of the upstream intersection traffic flow of the control channels N and f through the road sectionU(N,f)(k) Indicating the required flow generated by the traffic flow at the upstream intersection of the control channels N, f, and psi (N, f) indicating the steering traffic flow set contained in the control channels N, f.
Furthermore, defining signal control elements in an intersection signal control scheme based on convergence constraint to obtain an indefinite-period intersection store-and-forward state space equation,
the constraint includes: (A) the right turn control is normally green by default, only the motor vehicle straight-going and left turn control in the signal control is considered, and (B) the motor vehicle straight-going and left turn control channel is controlled by only one phase stage;
defining the signal control element includes: (C) a control set f, f represents a straight-going and left-turning control channel set of an intersection, (D) a timing scheme p,p represents a control scheme of the intersection; wherein p isj＝(pfj,gj) Pf represents a phase release channel set, g represents a phase green time;
the indefinite-period intersection store-and-forward state space equation is expressed by a matrix as follows:
Figure RE-GDA0003120550820000051
wherein the content of the first and second substances,
Figure RE-GDA0003120550820000052
s＝diag{s1,s2,...,sn}，
Figure RE-GDA0003120550820000053
ω is the matching matrix of the control channel and the phase scheme, ω ═ ωij|i∈[1,m],j∈[1,n]}，gj(k) Represents [ kT, (k +1) T]Green time of period phase j, c is signal control period duration, siIs a control channel fiSaturated flow rate of qi(k) Represents [ kT, (k +1) T]Duration control channel fiThe required flow of (2);
the intersection storage forwarding state space equation with fixed period is as follows:
Figure RE-GDA0003120550820000054
wherein the content of the first and second substances,
Figure RE-GDA0003120550820000055
sω＝[s1ω11(c-l),s2ω21(c-l),...,snωn1(c-l)]Tand s ω is a constant matrix generated at the time of the fixed-period conversion.
Further, establishing an evaluation function taking the queuing balance among the phases as a target, establishing a balance target function, and obtaining a P-type learning law based on the ILC fixed period balance control, wherein the P-type learning law comprises the steps of;
the construction formula of the equalization objective function is as follows:
y(k)＝[y2,1(k),y3,1(k),...,ym,1(k)]T
＝[z2(k)-z1(k),z3(k)-z1(k),...,zm(k)-z1(k)]T
wherein the phase queuing length
Figure RE-GDA0003120550820000061
Figure RE-GDA0003120550820000062
Indicating the predicted queue length, ym,l(k) Representing the difference between the phase m queue length and the reference phase queue length, ym,l(k)＝zm(k)-z1(k)；
The corresponding formula of the P-type learning law is as follows:
Figure RE-GDA0003120550820000063
ek(k+1)＝yd-yk(k+1)
where, k represents the number of iterations,
Figure RE-GDA0003120550820000064
is the green time produced after the kth iteration in the kth period, Γ is the iterative learning gain matrix, Γ satisfies
Figure RE-GDA0003120550820000065
I is an identity matrix of (m-1) × (m-1), C is y (k) and
Figure RE-GDA0003120550820000066
s is diag { s ═ d1,s2,...,sn}，yd＝0，
Figure RE-GDA0003120550820000067
Is omega at the time of conversion of the fixed period modelIn the form of a variation matrix of (a);
Figure RE-GDA0003120550820000068
the following constraints are satisfied:
Figure RE-GDA0003120550820000069
a data-driven control regional traffic signal dynamic optimization system comprises an analysis module, a traffic data analysis module, an intersection store-and-forward state space equation construction module and an iterative learning control model construction module;
the analysis module is used for analyzing the road network topological structure to obtain intersection attribute parameters in an intersection control area, wherein the intersection attribute parameters comprise traffic flow analysis and control channel analysis;
the traffic data analysis module is used for obtaining a traffic detection state index of the intersection attribute parameter based on the lane level state monitoring and the intersection attribute parameter;
the intersection storage and forwarding state space equation building module is used for building an intersection storage and forwarding state space equation meeting the signal control adaptation requirement according to the traffic detection state index of the intersection attribute parameters;
the iterative learning control model building module is used for building an iterative learning control model meeting the requirements of balance control and convergence, and performing iterative optimization on the intersection storage and forwarding state space equation to realize real-time optimization control of traffic signals.
A computer readable storage medium having stored thereon a number of acquisition and classification procedures for being invoked by a processor and performing the method for dynamic optimization of regional traffic signals according to claim 1.
The data-driven control regional traffic signal dynamic optimization method and the system provided by the invention have the advantages that: the dynamic optimization method and the system for the data-driven control regional traffic signals can adapt to different signal control design schemes and the change of the demand based on steering, realize fine control of intersections, have wide application scenes and high system automation degree, and are not easy to overflow; calculating through a road network topological structure to obtain a crossing storage and forwarding state space equation; the intersection store-and-forward state space equation is subjected to iterative optimization through the iterative learning control model, so that the control of the intersection store-and-forward state space equation on the intersection traffic flow is improved, and the overflow prevention of the intersection traffic path is realized.
Drawings
FIG. 1 is a schematic structural view of the present invention;
FIG. 2(a) is a schematic view of traffic flow inside an intersection with a road network topology structure;
FIG. 2(b) is a schematic diagram of a road network topology structure of intersection control channel;
FIG. 3 is a schematic illustration of an intersection in which a store-and-forward model of out-of-turn is considered;
FIG. 4 is a schematic diagram of an iterative learning process;
FIG. 5(a) is a cross connection diagram;
FIG. 5(b) is a schematic view of traffic flow inside the intersection;
fig. 6 is a recursive relationship between lane-traffic-control lanes.
Detailed Description
The present invention is described in detail below with reference to specific embodiments, and in the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein, but rather should be construed as broadly as the present invention is capable of modification in various respects, all without departing from the spirit and scope of the present invention.
As shown in fig. 1 to 6, the present invention provides a data-driven controlled regional traffic signal dynamic optimization method, including:
s1: analyzing a road network topological structure to obtain intersection attribute parameters in an intersection control area, wherein the intersection attribute parameters comprise traffic flow analysis and control channels;
s11: according to the design scheme of the intersection ditch division, intersection traffic flow data and control channel data can be obtained by analysis, and the analysis result of the traffic flow and the control channel of the intersection ditch division is shown in figure 2;
the traffic flow divides the traffic flow in the intersection into an inlet traffic flow and an outlet traffic flow (figure 2a) which are respectively turned into different directions according to the driving direction and the track of the vehicle, and sets I are respectively usedAAnd IB. A data structure as in table 1 is established to store traffic flow information.
TABLE 1 traffic flow data Structure
Crossing Numbering of intersections in system road network
Branch of Numbering of branches in intersections
Traffic flow Numbering of traffic flows in intersections
Type (B) Travel directions and turn types of traffic flows, including: straight, left-turn, right-turn and exit
Afflux traffic flow Import traffic flow set of import traffic flow merging into export traffic flow, type effective when export
In actual intersection channeling, the number of control lanes is less than the number of incoming traffic flows, and there are cases where one control lane contains multiple incoming traffic flows (fig. 2b), represented by set F. A data structure as in table 2 is established to store the control channel information.
TABLE 2 control channel data Structure
Crossing Numbering of intersections in system road network
Control channel Control channel numbering inside intersections
Controlling traffic flow Inlet traffic flow set of current intersection contained by control channel
S12: analyzing the topology of the road network, and determining the upstream traffic flow information of the intersection control channel according to the communication relation of adjacent intersections in the road network. And expanding the data structure on the basis of the table 2 to obtain a road network data structure based on the control channels, and referring to the table 3.
Table 3 control channel-based road network connection data structure
Crossing Numbering of intersections in system road network
Control channel Control channel numbering inside intersections
Upstream crossing Upstream crossing for controlling direction of passage
Upstream traffic flow Outlet traffic flow for generating traffic demand for control channel at upstream intersection
Through S11 and S12, technical data analysis of the road network is realized, and the control traffic flow and the control channel are numbered.
S2: obtaining a traffic detection state index of the intersection attribute parameters based on the lane level state monitoring and the intersection attribute parameters;
the traffic detection equipment can acquire traffic flow information of the road surface in real time, such as key information of flow, queue length and the like.
S21: in order to support subsequent modeling, the traffic detection result of the lane level inside the intersection needs to be converted into the traffic volume taking the traffic flow as a carrier. For import traffic flow IAIn other words, its value is obtained from the associated lane traffic statistics:
Figure RE-GDA0003120550820000091
wherein q isXRepresenting the traffic volume of the respective steered traffic stream, X representing a steering attribute, qY,ZThe detected traffic volume of the Z-th lane of the guide Y is shown, Y is the lane guide attribute, alpha is the steering proportion of the composite guide straight to the left, beta is the steering proportion of the composite guide straight to the right,
Figure RE-GDA0003120550820000092
the steering ratio of the composite steering to the left and right is shown, and γ is the steering ratio of the three-way lane.
Outlet traffic flow IBThe traffic flow is from the inlet traffic flow IAAnd (3) statistical acquisition:
(1) according to the intersection traffic flow data, the traffic flow sink can be obtainedIn relation matrix R ═ Rij|i∈[1,m],j∈[1,n]Is set I for m and n respectivelyBAnd IAThe elements of the matrix R satisfy:
Figure RE-GDA0003120550820000093
wherein Π (i) is an incoming traffic flow set of the outgoing traffic flow i.
(2) Outlet traffic flow IBTraffic qBAnd import traffic flow IATraffic qASatisfies the following matrix relationship:
qB＝R×qA (2)
wherein q isB＝{qn+1,qn+2,...,qn+m}T，qA＝{q1,q2,...,qn}T
S22: and counting according to the lane detection queuing length to obtain the data of the number of the queued vehicles of the control channel. According to the control channel data structure in table 2, the queuing length of the covered lane is counted as the index of the number of queued vehicles of the control lane by the corresponding relationship of the control channel, the traffic flow and the lane, and the calculation formula of the queuing length corresponds to the following formula:
Figure RE-GDA0003120550820000101
wherein, qNumfIs the number of queued vehicles, qLen, for control channel flIndicates the length of the queue of the lane L, LvehRepresenting the average vehicle occupancy length, lane i is controlled by lane f.
The control traffic flow on the control passage is calculated through steps S21 to S22 as the input of the intersection store-and-forward state space equation of step S3.
S3: according to the traffic detection state index and the signal control scheme, constructing an intersection storage and forwarding state space equation meeting the signal control adaptation requirement;
s31: for one directional branch z connecting the intersections M and N as shown in fig. 3, the vehicle is driven from the intersection M to the intersection N. Considering a store-and-forward model in which a branch z turns at an intersection N, the dynamics of the branch can be described as:
xz,l(k+1)＝xz,l(k)+λz,lT[qz(k)+εz(k)]-Tuz,l(k) (4)
εz(k)＝dz(k)-sz(k)
wherein x isz,l(k +1) represents the number of queued vehicles on the road segment z out of turn l at the beginning of the (k +1) th time period, xz,l(k) Representing the number of queued vehicles on the road segment z out of turn/at the beginning of the kth time period; q. q.sz(k) Represents [ kT, (k +1) T]Traffic volume, e, during which the vehicle travels into the stretch zz(k) Represents [ kT, (k +1) T]Disturbance of the state of the section z itself during the period, λz,lIs the proportion of the steering that the vehicle on the observable stretch z is driving out of the steering; u. ofz,l(k) Is [ kT, (k +1) T]Traffic volume during which a section z is driven from a turn l, T is a statistical time interval, dz(k) Represents [ kT, (k +1) T]The demanded flow, s, generated by the section z itself during the periodz(k) Represents [ kT, (k +1) T]The dissipated flow of the section z during.
Taking into account disturbances epsilonz(k) Can be expressed in terms of demand and disturbance factor, i.e. εz(k)＝δzqz(k) And delta is the observable disturbance coefficient, equation (4) can be further simplified to
xz,l(k+1)＝xz,l(k)+Tqz,l(k)-Tuz,l(k) (5)
Wherein q isz,l(k)＝λz,l(1+δz)qz(k) And the lambda and the delta can be popularized as time change functions in a long-time operation system, and specific calibrated parameters are determined in practical application.
S32: the formula (5) is a store-and-forward model established on road section steering, the road section steering and the intersection steering have a corresponding relation, and the steering state equation belonging to the same control channel can obtain the store-and-forward model based on the intersection control channel through accumulation
Figure RE-GDA0003120550820000111
qz,l(k)＝λz,l(1+δz)qz(k)
Based on the upstream intersection and traffic flow information of the control lane in Table 3, q in equation (6)N,f(k) Can be represented by the traffic volume of the upstream intersection and the traffic flow
Figure RE-GDA0003120550820000112
Wherein x isN,f(k +1) represents the number of queued vehicles exiting the control lane N, f to turn to l at the beginning of the (k +1) th time period, U (N, f) represents the traffic flow at the upstream intersection of the control lane N, f, qz,l(k) Represents [ kT, (k +1) T]During which the demanded flow, u, of the traffic flow is driven out of the turn l of the stretch zz,l(k) Represents [ kT, (k +1) T]During which the dissipated flow, delta, of the traffic flow is driven out of the turn/of the stretch zzRepresenting the proportion of disturbance of the demanded flow on the road section z, qN,f(k) Represents [ kT, (k +1) T]During which the required flow, lambda, of the channel N, f is controlledz,lIs the proportion of the steering, δ, at which the vehicle on the stretch z can be seen to be driven out of the steering lU(N,f)Representing the disturbance proportion q generated by the transmission of the upstream intersection traffic flow of the control channels N and f through the road sectionU(N,f)(k) Indicating the required flow generated by the traffic flow at the upstream intersection of the control channels N, f, and psi (N, f) indicating the steering traffic flow set contained in the control channels N, f.
S33: and (4) considering the actual signal control scheme of the intersection. To ensure final convergence, the following two constraints (1) and (2) are imposed on the interface signal control:
(1) the default of the right-turn control is evergreen, and only the motor vehicle straight-going and left-turn control in the signal control scheme is considered in the model;
(2) the motor vehicle straight and left-turn control channels are controlled in only one phase, i.e. no overlapping phase of the straight and left-turn channels occurs.
Defining signal control elements:
A. control set f, f ═ f1,f2,...,fn]Representing a straight-going and left-turning control channel set of a crossing;
B. timing scheme p, p ═ p [ p ]1,p2,...,pm]Indicating a control scheme of the intersection; wherein p isj＝(pfj,gj) Pf is a phase release channel set, g is a phase green time;
a matching matrix ω ═ ω { ω } exists between the control set and the phase schemeij|i∈[1,m],j∈[1,n]And expressing the controlled relationship between the control channels and the phase scheme in the set, wherein the matrix element values satisfy the following relationship:
Figure RE-GDA0003120550820000121
establishing a storage forwarding state space equation of elements in the intersection control set f based on the formula (6):
xi(k+1)＝xi(k)+Tqi(k)-Tui(k) (7)
wherein the index i denotes the ith control channel f in the control set fi，qi(k) And ui(k) Respectively a control channel fiIs in a near-saturation state, the dissipation flow u is considered to bei(k) Satisfy the requirement of
Figure RE-GDA0003120550820000122
Wherein, gj(k) Represents [ kT, (k +1) T]Green time of period phase j, c is signal control period duration, siIs a control channel fiThe saturation flow rate of (c). If the statistical period length T is set to the signal control period c, equation (7) can be converted to
Figure RE-GDA0003120550820000123
At this time, the original output ui(k) Is replaced bygj(k) The expressed output capability, so the state space transition result is a predicted value.
Therefore, the intersection store-and-forward state space equation of the whole intersection with the indefinite period is expressed by a matrix as follows:
Figure RE-GDA0003120550820000124
wherein the content of the first and second substances,
Figure RE-GDA0003120550820000125
s＝diag{s1,s2,...,sn}，
Figure RE-GDA0003120550820000126
ω is the matching matrix of the control channel and the phase scheme, ω ═ ωij|i∈[1,m],j∈[1,n]}，gj(k) Represents [ kT, (k +1) T]Green time of period phase j, c is signal control period duration, siIs a control channel fiSaturated flow rate of qi(k) Represents [ kT, (k +1) T]Duration control channel fiThe required flow rate of (c).
S34: in order to meet basic regional coordination control, intersection signal control can be required to keep the same period duration within a period of time.
According to
Figure RE-GDA0003120550820000131
With period c and loss l known, the green time for phase 1 is represented by the other phase green time:
Figure RE-GDA0003120550820000132
the intersection store-and-forward state space equation with a fixed period can be obtained by substituting the equation (9):
Figure RE-GDA0003120550820000133
wherein the content of the first and second substances,
Figure RE-GDA0003120550820000134
sω＝[s1ω11(c-l),s2ω21(c-l),...,snωn1(c-l)]Tand is a constant matrix generated during the fixed period conversion.
Through steps S31 to S34, the traffic flow of the intersection control channel directly entering and exiting is converted into the traffic flow controlled by the signal, and whether the control cycle duration is set for the traffic flow controlled by the signal is set, so as to obtain the intersection store-and-forward state space equation with the indefinite cycle and the intersection store-and-forward state space equation with the definite cycle.
S4: and (3) an iterative learning control model based on balance control and convergence requirements is established, iterative optimization is carried out on the intersection storage and forwarding state space equation, and real-time traffic signal optimization control is realized.
S41: and establishing an evaluation function taking the queuing balance among the phases as a target. In each determined signal period, the phase (x) is selectedi(k)+c·qi(k))/siPredicted queue length for maximum lane
Figure RE-GDA0003120550820000135
As the phase queuing length zj(k) And constructing a queuing balance objective function among phases. Taking phase 1 as a reference phase, and queuing difference y between other phases and the reference phasej,1(k)＝zj(k)-z1(k) And (3) forming an equalization objective function:
y(k)＝[y2,1(k),y3,1(k),...,ym,1(k)]T
＝[z2(k)-z1(k),z3(k)-z1(k),...,zm(k)-z1(k)]T
wherein the content of the first and second substances,
Figure RE-GDA0003120550820000141
Figure RE-GDA0003120550820000142
indicating the predicted queue length, ym,l(k) Representing the difference between the phase m queue length and the reference phase (phase 1) queue length, i.e. ym,l(k)＝zm(k)-z1(k) Due to zj(k) Is determined by (x)i(k) And q isi(k) Is deterministic) and thus y (k) can be written about
Figure RE-GDA0003120550820000143
The linear relation of (1):
Figure RE-GDA0003120550820000144
wherein C is an m-1 × n matrix with the element value { -1,0,1}, and the queuing balance means that the queuing lengths in all directions of the intersection are equal during the control period, so the desired target is ydThe difference between the desired target and the actual target is defined as the deviation, 0:
ek(k+1)＝yd-yk(k+1)
s41: through the modeling process, the P-type learning law of the fixed-period balance control based on the ILC can be obtained:
Figure RE-GDA0003120550820000145
where, k represents the number of iterations,
Figure RE-GDA0003120550820000146
is the green time produced after the kth iteration in the kth period, Γ is the iterative learning gain matrix, Γ needs to meet
Figure RE-GDA0003120550820000147
The method specifically comprises the following steps: the value is given by
Figure RE-GDA0003120550820000148
The value of (2) is determined; according to the prior artResults of the study when satisfied
Figure RE-GDA0003120550820000149
The gain matrix values of (a) are such that convergence of the model is guaranteed, wherein I is a matrix of (m-1) × (m-1), and C is y (k) and
Figure RE-GDA00031205508200001410
s is diag { s ═ d1,s2,...,sn}，
Figure RE-GDA00031205508200001411
Is the variation matrix form of omega in the fixed period model conversion. Will control in practical application
Figure RE-GDA00031205508200001412
By selecting a value level of (e.g. selecting)
Figure RE-GDA00031205508200001413
The gain matrix of (a) takes values.
The iterative learning input quantity is the green light time of each phase of the intersection, is limited by the duration of the minimum green, the maximum green and the fixed period, and is balanced and controlled in the fixed period, the green light time of the first phase is represented by other phases, periods and loss time, so that
Figure RE-GDA00031205508200001414
The following constraints are to be satisfied:
Figure RE-GDA00031205508200001415
in the iterative learning allocation process, changes exceeding the maximum green, the minimum green and the fixed period limit occur, and the change value exceeding the limit is uniformly allocated and balanced.
S42: an iterative learning process. Each signal period generates the initial state x of the current intersection according to the signal control scheme of the previous period0(k)、
Figure RE-GDA0003120550820000151
And demand q generated by adjacent crossing0(k) Continuously iteratively updating according to the learning rate in equation (11)
Figure RE-GDA0003120550820000152
Until an iteration ceiling or convergence threshold is met, i.e. | | eκ(k+1)||2Sigma is less than or equal to. The iterative process from the iteration start time and the initial state is shown in fig. 4, where the current state quantity and requirement of the k-th iteration are strict repetitions of the initial state, i.e. x, in the same signal periodκ(k)＝x0(k)，qκ(k)＝q0(k)。
And performing iterative optimization on the intersection store-and-forward state space equation through S41 to S42, and improving the optimization degree of the intersection store-and-forward state space equation.
Through steps S1-S4, the application scenarios are wide, the system automation degree is high, and overflow is not easy to generate.
As an example:
road network data analysis based on road network topological structure and intersection and channel planning design
Fig. 5a shows a simple road network structure, in which intersections 1-4 are all crossroads, intersection 5 is a T-shaped intersection with east-west connectivity, and the inside of the intersection is not restricted or restricted. Firstly, traffic flow analysis is set according to the type of the intersection and the channel division to obtain the traffic flow structure of each intersection in fig. 5b, and part of the traffic flow analysis content of the intersection 1 is given in table 4.
Table 4 example intersection traffic flow analysis results
Figure RE-GDA0003120550820000153
Assuming that the internal channeling of intersection 1 is consistent with the channeling scheme in fig. 2a, and the communication relationship between adjacent intersections is combined, the analysis result of the control channel data of intersection 1 is shown in table 5.
Table 5 illustrates the intersection control channel resolution results
Figure RE-GDA0003120550820000161
And secondly, constructing a detection data analysis module based on the traffic flow and the control channel. And taking the lane-level detection data as input, and obtaining the dynamic data output required by the technical application embodiment according to the progressive attribution relation of the lane, the traffic flow and the control channel. The data processing flow is shown in fig. 6. And taking the lane-level detection data as input, and obtaining the dynamic data output required by the technical application embodiment according to the progressive attribution relation of the lane, the traffic flow and the control channel. The data processing flow is shown in fig. 6.
1. Traffic volume statistics
(1) And (3) statistics of the import traffic volume: on the basis of matching of the inlet traffic flow and the lane, the lane detection flow is counted according to the formula (1) to obtain the traffic volume counting result of all inlet traffic flows in the intersection;
(2) and (4) outlet traffic volume statistics: determining a traffic flow influx relation matrix according to intersection traffic flow data, taking intersection traffic flow data in table 4 as an example, IB＝{13,14,15,16}，IA1,2, 12, a relationship matrix
Figure RE-GDA0003120550820000162
Further carrying out an inlet traffic flow I according to the formula (2)ATraffic flow to Exit IBTraffic volume conversion of
qB＝R×qA
Wherein q isB＝{q13,q14,q15,q16}T，qA＝{q1,q2,...,q12}T。
(3) Export traffic application: the exit traffic volume inside the intersection can be used as the demand traffic volume of the associated branch of the downstream intersection.
2. Queue length statistics
And on the basis of matching of the control channel and the lane, converting the output lane queuing length data into the queuing vehicle number data of the control channel according to the formula (3) and outputting the data, wherein the data can be used as the input of the formula (4).
Thirdly, intersection store-and-forward state space equation
1. And determining an intersection signal control scheme according to the actual traffic demand, and analyzing and obtaining signal control elements required by the state space equation by taking the signal control scheme of the intersection 1 given in the table 6 as an example.
Table 6 signal control scheme example
Figure RE-GDA0003120550820000171
Note: the release channel information only comprises a direct-turning control channel and a left-turning control channel
(1) A control channel: f ═ f1,f2,...,f7}＝{2,3,5,6,8,10,11}；
(2) The timing scheme is as follows:
p＝{p1,p2,p3,p4}＝{[pf1,27],[pf2,27],[pf3,27],[pf4,27]}; wherein, pf1＝{f1,f2}＝{2,3}，pf2＝{f5}＝{8}，pf3＝{f3,f6}＝{5,10}， pf4＝{f4,f7}＝{6,11}；
(3) Matching a matrix:
Figure RE-GDA0003120550820000172
2. determining the state space equation of the intersection 1 with the statistical interval of one signal cycle c according to the equation form given by the formula (10):
Figure RE-GDA0003120550820000173
wherein the content of the first and second substances,
Figure RE-GDA0003120550820000181
sω＝[108s1,108s2,0,0,0,0,0,0]Tq (k) is the traffic demand generated upstream of the control channel, q (k) is the first control channel f1For example, the traffic demand of (A) is satisfied
q1(k)＝λ1(1+δ2→1)q2→1(k)＝λ1(1+δ2→1)q2,15(k)
Wherein λ is1Is f1Steering ratio of delta2→1Is the traffic flow disturbance coefficient of intersection 2 to 1;
Figure RE-GDA0003120550820000182
the minimum green time, maximum green time constraints in table 6 are met:
Figure RE-GDA0003120550820000183
fourthly, constructing an iterative learning control model
1. An equalization control target is established based on the phase. Taking intersection 1 as an example, the number of queued vehicles in the first phase is taken as a reference, and the difference value between the number of queued vehicles in other phases and the reference is taken as a balance control target
y(k)＝[y2,1(k),y3,1(k),y4,1(k)]T
＝[z2(k)-z1(k),z3(k)-z1(k),z4(k)-z1(k)]T
Selecting (x) in phasei(k)+c·qi(k))/siThe state prediction value of the larger control channel is taken as the phase-queued vehicle number, and the equalization control target is further converted into the form y (k) ═ cx (k) with respect to the control channel. Under the signal control conditions given by the example, C has eight values, each of which is a different choice of control channel in the phase: to be provided with
Figure RE-GDA0003120550820000184
For the purpose of example only,
Figure RE-GDA0003120550820000185
2. establishing iterative learning input
Figure RE-GDA0003120550820000186
And output yd-y (k +1) P-type learning rate, with the k +1 th iteration as follows:
Figure RE-GDA0003120550820000191
to ensure the convergence of the learning rate, the gain matrix Γ should satisfy
Figure RE-GDA0003120550820000192
And according to the eight values of C, searching and obtaining corresponding eight gain matrixes meeting the convergence condition through a particle swarm optimization algorithm, and selecting the determined gamma to participate in iterative computation in each signal period.
3. An iterative learning process. Each signal period is based on the current intersection initial state x generated by the previous period signal control scheme0(k)、
Figure RE-GDA0003120550820000193
And demand q generated by adjacent crossing0(k) According to x0(k) And q is0(k) Determining a coefficient matrix C (i.e. a linear relation matrix) and a gain matrix gamma of the balance target of the current signal period, and continuously and iteratively updating the learning rate in the formula (12)
Figure RE-GDA0003120550820000194
Until an iteration ceiling or convergence threshold is met, i.e. | | eκ(k+1)||2Sigma is less than or equal to. The iterative process from the iteration start time and the initial state is shown in fig. 4, where the current state quantity and requirement of the k-th iteration are strict repetitions of the initial state, i.e. x, in the same signal periodκ(k)＝x0(k)，qκ(k)＝q0(k)。
It should be noted that the above turning ratios are all determined by observation.
A data-driven control regional traffic signal dynamic optimization system is characterized by comprising an analysis module, a traffic data analysis module, an intersection storage and forwarding state space equation construction module and an iterative learning control model construction module;
the analysis module is used for analyzing the road network topological structure to obtain intersection attribute parameters in an intersection control area, wherein the intersection attribute parameters comprise traffic flow analysis and control channel analysis;
the traffic data analysis module is used for obtaining a traffic detection state index of the intersection attribute parameter based on the lane level state monitoring and the intersection attribute parameter;
the intersection storage and forwarding state space equation building module is used for building an intersection storage and forwarding state space equation meeting the signal control adaptation requirement according to the traffic detection state index of the intersection attribute parameters;
the iterative learning control model building module is used for building an iterative learning control model meeting the requirements of balance control and convergence, and performing iterative optimization on the intersection storage and forwarding state space equation to realize real-time optimization control of traffic signals.
A computer-readable storage medium, wherein the computer-readable storage medium has stored thereon a plurality of acquisition and classification programs, and the plurality of acquisition and classification programs are used for being called by a processor and executing the regional traffic signal dynamic optimization method according to the present application.
Those of ordinary skill in the art will understand that: all or part of the steps for implementing the method embodiments may be implemented by hardware related to program instructions, and the program may be stored in a computer readable storage medium, and when executed, the program performs the steps including the method embodiments; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.

Claims (10)

1. A data-driven control regional traffic signal dynamic optimization method is characterized by comprising the following steps:
analyzing a road network topological structure to obtain intersection attribute parameters in an intersection control area, wherein the intersection attribute parameters comprise traffic flow analysis and control channels;
obtaining a traffic detection state index of the intersection attribute parameter based on the lane level state monitoring and the intersection attribute parameter;
constructing an intersection storage and forwarding state space equation meeting the signal control adaptation requirement according to the traffic detection state index and the signal control scheme;
and constructing an iterative learning control model meeting the requirements of balance control and convergence, and performing iterative optimization on the intersection storage and forwarding state space equation to realize real-time optimization control of the traffic signals.
2. The data-driven regional traffic signal dynamic optimization method of claim 1, wherein the obtaining of the traffic detection state index of the intersection attribute parameter based on the lane-level state monitoring and the intersection attribute parameter comprises:
converting traffic flow information of a lane level inside the intersection into traffic flow taking the traffic flow as a carrier, taking the traffic flow as the carrier as an input parameter under the control of a traffic signal, and acquiring the traffic flow information in real time through traffic detection equipment;
and obtaining the number of the queued vehicles of the control channel according to the lane detection queuing length, wherein the number of the queued vehicles is obtained in real time through traffic detection equipment.
3. The dynamic optimization method of regional traffic signals under data-driven control according to claim 1, wherein the construction of the intersection store-and-forward state space equation meeting the signal control adaptation requirement according to the traffic detection state index and the signal control scheme comprises:
according to the store-and-forward model on the road section steering, the steering state equation belonging to the same control channel is accumulated to obtain a store-and-forward model based on the intersection control channel;
defining signal control elements by an intersection signal control scheme based on convergence constraint to obtain an intersection store-and-forward state space equation with an indefinite period;
and (4) considering the coordination control requirement, carrying out derivation and conversion of the given period and the green time loss time on the crossing store-and-forward state space equation with the indefinite period to obtain the crossing store-and-forward state space equation with the definite period.
4. The dynamic optimization method of traffic signals in data-driven controlled areas according to claim 1, wherein in the construction of an iterative learning control model satisfying requirements for equalization control and convergence, an iterative optimization is performed on a spatial equation of a store-and-forward state of a road junction, and a real-time traffic signal optimization control is realized, the method comprises the following steps:
establishing an evaluation function taking the phase-to-phase queuing balance as a target, constructing a balance target function, and obtaining a P-type learning law based on ILC (constant-period balance control);
and constructing an iterative learning control model and a control flow based on a P-type learning law, a balanced objective function and a fixed-period intersection storage and forwarding state space equation.
5. The method according to claim 2, wherein in converting the traffic flow information at the lane level inside the intersection into the traffic flow using the traffic flow as the carrier, the traffic flow at the intersection includes an inlet traffic flow IAAnd exit traffic flow IBThe concrete formula is as follows;
for import traffic flow IA：
Figure FDA0002924202390000021
Wherein q isXRepresenting the traffic volume of the respective steered traffic stream, X representing a steering attribute, qY,ZThe detected traffic volume of the Z-th lane of the guide Y is shown, Y is the lane guide attribute, alpha is the steering proportion of the composite guide straight to the left, beta is the steering proportion of the composite guide straight to the right,
Figure FDA0002924202390000022
the steering proportion of the composite steering to the left and the right is shown, and gamma is the steering proportion of a three-way lane;
outlet traffic flow IBThe traffic flow is from the inlet traffic flow IAAnd (3) statistical acquisition:
(1) according to the intersection traffic flow data, a vehicle flow convergence relation matrix R ═ { R ═ R can be obtainedij|i∈[1,m],j∈[1,n]M denotes the set IBN represents the set IAThe elements of the matrix R satisfy:
Figure FDA0002924202390000031
wherein pi (i) represents an import traffic flow set of an export traffic flow i;
(2) outlet traffic flow IBTraffic qBAnd import traffic flow IATraffic qASatisfies the following matrix relationship:
qB＝R×qA
wherein q isB＝{qn+1,qn+2,...,qn+m}T，qA＝{q1,q2,...,qn}T。
6. The dynamic optimization method of regional traffic signals under data-driven control according to claim 3, wherein in the step of obtaining the store-and-forward model based on the intersection control channel by adding the steering state equations of the same control channel according to the store-and-forward model on the road section steering, the store-and-forward model on the road section steering and the store-and-forward model based on the intersection control channel are obtained by the following formulas respectively:
the store-and-forward model on the road section steering corresponds to the following formula:
xz,l(k+1)＝xz,l(k)+λz,lT[qz(k)+εz(k)]-Tuz,l(k)
εz(k)＝dz(k)-sz(k)
wherein x isz,l(k +1) represents the number of queued vehicles on the road segment z out of turn l at the beginning of the (k +1) th time period, xz,l(k) Representing the number of queued vehicles on the road segment z out of turn/at the beginning of the kth time period; q. q.sz(k) Represents [ kT, (k +1) T]Traffic volume, e, during which the vehicle travels into the stretch zz(k) Represents [ kT, (k +1) T]Disturbance of the state of the section z itself during the period, λz,lIs the proportion of the steering that the vehicle on the observable stretch z is driving out of the steering; u. ofz,l(k) Is [ kT, (k +1) T]Traffic volume during which a section z is driven from a turn l, T is a statistical time interval, dz(k) Represents [ kT, (k +1) T]The demanded flow, s, generated by the section z itself during the periodz(k) Represents [ kT, (k +1) T]Dissipation flow of the period section z;
the storage forwarding model based on the intersection control channel corresponds to the following formula:
Figure FDA0002924202390000032
qz,l(k)＝λz,l(1+δz)qz(k)
Figure FDA0002924202390000041
wherein x isN,f(k +1) represents the number of queued vehicles exiting the control lane N, f to turn to l at the beginning of the (k +1) th time period, U (N, f) represents the traffic flow at the upstream intersection of the control lane N, f, qz,l(k) Represents [ kT, (k +1))T]During which the demanded flow, u, of the traffic flow is driven out of the turn l of the stretch zz,l(k) Represents [ kT, (k +1) T]During which the dissipated flow, delta, of the traffic flow is driven out of the turn/of the stretch zzRepresenting the proportion of disturbance of the demanded flow on the road section z, qN,f(k) Represents [ kT, (k +1) T]During which the required flow, lambda, of the channel N, f is controlledz,lIs the proportion of the steering, δ, at which the vehicle on the stretch z can be seen to be driven out of the steering lU(N,f)Representing the disturbance proportion q generated by the transmission of the upstream intersection traffic flow of the control channels N and f through the road sectionU(N,f)(k) Indicating the required flow generated by the traffic flow at the upstream intersection of the control channels N, f, and psi (N, f) indicating the steering traffic flow set contained in the control channels N, f.
7. The data-driven regional traffic signal dynamic optimization method according to claim 3, wherein signal control elements are defined in an intersection signal control scheme based on convergence constraints, and an indefinite-period intersection store-and-forward state space equation is obtained in which:
the constraint includes: (A) the right turn control is normally green by default, only the motor vehicle straight-going and left turn control in the signal control is considered, and (B) the motor vehicle straight-going and left turn control channel is controlled by only one phase stage;
defining the signal control element includes: (C) a control set f represents a straight-going and left-turning control channel set of an intersection, and (D) a timing scheme p represents a control scheme of the intersection; wherein p isj＝(pfj,gj) Pf represents a phase release channel set, g represents a phase green time;
the indefinite-period intersection store-and-forward state space equation is expressed by a matrix as follows:
Figure FDA0002924202390000042
wherein x (k) ═ x1(k),x2(k),...,xn(k)]T，q(k)＝[q1(k),q2(k),...,qn(k)]T，s＝diag{s1,s2,...,sn}，g(k)＝[g1(k),g2(k),...,gm(k)]TWhere ω is the matching matrix for the control channel and phase scheme, ω ═ ωij|i∈[1,m],j∈[1,n]}，gj(k) Represents [ kT, (k +1) T]Green time of period phase j, c is signal control period duration, siIs a control channel fiSaturated flow rate of qi(k) Represents [ kT, (k +1) T]Duration control channel fiThe required flow of (2);
the intersection storage forwarding state space equation with fixed period is as follows:
Figure FDA0002924202390000043
wherein the content of the first and second substances,
Figure FDA0002924202390000051
sω＝[s1ω11(c-l),s2ω21(c-l),...,snωn1(c-l)]Tand is a constant matrix generated during the fixed period conversion.
8. The dynamic optimization method of regional traffic signals under data-driven control according to claim 4, wherein in establishing an evaluation function with the aim of phase-to-phase queuing equalization, constructing an equalization objective function, and obtaining a P-type learning law of ILC-based fixed-period equalization control, the method comprises the following steps;
the construction formula of the equalization objective function is as follows:
y(k)＝[y2,1(k),y3,1(k),...,ym,1(k)]T
＝[z2(k)-z1(k),z3(k)-z1(k),...,zm(k)-z1(k)]T
wherein the phase queuing length
Figure FDA0002924202390000052
Figure FDA0002924202390000053
Indicating the predicted queue length, ym,l(k) Representing the difference between the phase m queue length and the reference phase queue length, ym,l(k)＝zm(k)-z1(k)；
The corresponding formula of the P-type learning law is as follows:
Figure FDA0002924202390000054
ek(k+1)＝yd-yk(k+1)
where, k represents the number of iterations,
Figure FDA0002924202390000055
is the green time produced after the kth iteration in the kth period, Γ is the iterative learning gain matrix, Γ satisfies
Figure FDA0002924202390000056
I is an identity matrix of (m-1) × (m-1), C is y (k) and
Figure FDA0002924202390000057
s is diag { s ═ d1,s2,...,sn}，yd＝0，
Figure FDA0002924202390000058
Is a variation matrix form of omega in the fixed period model conversion;
Figure FDA0002924202390000059
the following constraints are satisfied:
Figure FDA00029242023900000510
9. a data-driven control regional traffic signal dynamic optimization system is characterized by comprising an analysis module, a traffic data analysis module, an intersection storage and forwarding state space equation construction module and an iterative learning control model construction module;
the analysis module is used for analyzing the road network topological structure to obtain intersection attribute parameters in an intersection control area, wherein the intersection attribute parameters comprise traffic flow analysis and control channel analysis;
the traffic data analysis module is used for obtaining a traffic detection state index of the intersection attribute parameter based on lane level state monitoring and the intersection attribute parameter;
the intersection store-and-forward state space equation construction module is used for constructing an intersection store-and-forward state space equation meeting the signal control adaptation requirement according to the traffic detection state index;
the iterative learning control model building module is used for building an iterative learning control model meeting the requirements of balance control and convergence, and performing iterative optimization on the intersection store-and-forward state space equation to realize the real-time optimization control of the traffic signals.
10. A computer-readable storage medium having stored thereon a plurality of acquisition and classification procedures for being invoked by a processor and performing the method for dynamic optimization of regional traffic signals according to claim 1.
CN202110128223.6A 2021-01-29 2021-01-29 Data drive control regional traffic signal dynamic optimization method and system Active CN113192318B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110128223.6A CN113192318B (en) 2021-01-29 2021-01-29 Data drive control regional traffic signal dynamic optimization method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110128223.6A CN113192318B (en) 2021-01-29 2021-01-29 Data drive control regional traffic signal dynamic optimization method and system

Publications (2)

Publication Number Publication Date
CN113192318A true CN113192318A (en) 2021-07-30
CN113192318B CN113192318B (en) 2022-09-02

Family

ID=76972745

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110128223.6A Active CN113192318B (en) 2021-01-29 2021-01-29 Data drive control regional traffic signal dynamic optimization method and system

Country Status (1)

Country Link
CN (1) CN113192318B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113380044A (en) * 2021-08-12 2021-09-10 深圳市城市交通规划设计研究中心股份有限公司 Overflow control signal optimization method and device and storage medium

Citations (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100864451B1 (en) * 2007-04-19 2008-10-22 학교법인 영광학원 Vehicle detection system based vehicle to roadside communication and traffic signal control method using the vehicle detection system in crossroad
DE102008049568A1 (en) * 2008-09-30 2010-04-08 Siemens Aktiengesellschaft A method of optimizing traffic control at a traffic signal controlled node in a road traffic network
CN103942968A (en) * 2014-04-22 2014-07-23 南京大学镇江高新技术研究院 Intelligent traffic light dispatching system and method based on traffic flow sensing
CN104575021A (en) * 2014-12-17 2015-04-29 浙江工业大学 Distributed model predictive control method for urban road network system based on neighborhood optimization
CN104882006A (en) * 2014-07-03 2015-09-02 中国科学院沈阳自动化研究所 Message-based complex network traffic signal optimization control method
CN105006158A (en) * 2015-06-25 2015-10-28 广东工业大学 Single road intersection traffic signal control method based on real-time traffic information
CN105279982A (en) * 2015-11-13 2016-01-27 北京航空航天大学 Single intersection dynamic traffic signal control method based on data driving
US20160379486A1 (en) * 2015-03-24 2016-12-29 Donald Warren Taylor Apparatus and system to manage monitored vehicular flow rate
CN106408957A (en) * 2016-11-21 2017-02-15 华南理工大学 Intersection phase time distribution method based on pass demand balance
CN109767630A (en) * 2019-04-12 2019-05-17 江苏广宇科技产业发展有限公司 A kind of traffic signal control system based on bus or train route collaboration
CN109993982A (en) * 2019-05-08 2019-07-09 公安部交通管理科学研究所 Unbalance intersection signal phase optimization design method based on queueing theory
CN110570651A (en) * 2019-07-15 2019-12-13 浙江工业大学 Road network traffic situation prediction method and system based on deep learning
CN111583675A (en) * 2020-05-14 2020-08-25 吴钢 Regional road network traffic signal lamp coordination control system and method
CN111710167A (en) * 2020-08-24 2020-09-25 深圳市城市交通规划设计研究中心股份有限公司 Single-point optimization control method and device based on online traffic simulation
CN112133086A (en) * 2020-08-10 2020-12-25 北方工业大学 Regional traffic signal data driving control method based on multi-agent network
CN112258856A (en) * 2020-08-10 2021-01-22 北方工业大学 Method for establishing regional traffic signal data drive control model

Patent Citations (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100864451B1 (en) * 2007-04-19 2008-10-22 학교법인 영광학원 Vehicle detection system based vehicle to roadside communication and traffic signal control method using the vehicle detection system in crossroad
DE102008049568A1 (en) * 2008-09-30 2010-04-08 Siemens Aktiengesellschaft A method of optimizing traffic control at a traffic signal controlled node in a road traffic network
CN103942968A (en) * 2014-04-22 2014-07-23 南京大学镇江高新技术研究院 Intelligent traffic light dispatching system and method based on traffic flow sensing
CN104882006A (en) * 2014-07-03 2015-09-02 中国科学院沈阳自动化研究所 Message-based complex network traffic signal optimization control method
CN104575021A (en) * 2014-12-17 2015-04-29 浙江工业大学 Distributed model predictive control method for urban road network system based on neighborhood optimization
US20160379486A1 (en) * 2015-03-24 2016-12-29 Donald Warren Taylor Apparatus and system to manage monitored vehicular flow rate
CN105006158A (en) * 2015-06-25 2015-10-28 广东工业大学 Single road intersection traffic signal control method based on real-time traffic information
CN105279982A (en) * 2015-11-13 2016-01-27 北京航空航天大学 Single intersection dynamic traffic signal control method based on data driving
CN106408957A (en) * 2016-11-21 2017-02-15 华南理工大学 Intersection phase time distribution method based on pass demand balance
CN109767630A (en) * 2019-04-12 2019-05-17 江苏广宇科技产业发展有限公司 A kind of traffic signal control system based on bus or train route collaboration
CN109993982A (en) * 2019-05-08 2019-07-09 公安部交通管理科学研究所 Unbalance intersection signal phase optimization design method based on queueing theory
CN110570651A (en) * 2019-07-15 2019-12-13 浙江工业大学 Road network traffic situation prediction method and system based on deep learning
CN111583675A (en) * 2020-05-14 2020-08-25 吴钢 Regional road network traffic signal lamp coordination control system and method
CN112133086A (en) * 2020-08-10 2020-12-25 北方工业大学 Regional traffic signal data driving control method based on multi-agent network
CN112258856A (en) * 2020-08-10 2021-01-22 北方工业大学 Method for establishing regional traffic signal data drive control model
CN111710167A (en) * 2020-08-24 2020-09-25 深圳市城市交通规划设计研究中心股份有限公司 Single-point optimization control method and device based on online traffic simulation

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
路婷等: "基于交叉口重要度深度搜索的区域信号协调控制方法", 《交通运输系统工程与信息》 *
陈家旭，赵永进: "基于关联交通流的区域交通信号协调控制方法", 《工业技术创新》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113380044A (en) * 2021-08-12 2021-09-10 深圳市城市交通规划设计研究中心股份有限公司 Overflow control signal optimization method and device and storage medium
CN113380044B (en) * 2021-08-12 2022-01-07 深圳市城市交通规划设计研究中心股份有限公司 Overflow control signal optimization method and device and storage medium

Also Published As

Publication number Publication date
CN113192318B (en) 2022-09-02

Similar Documents

Publication Publication Date Title
CN112466117A (en) Road network short-term traffic flow prediction method based on deep space-time residual error network
CN107591011B (en) Intersection traffic signal self-adaptive control method considering supply side constraint
CN108648446B (en) Road network traffic signal iterative learning control method based on MFD
CN110782663B (en) Road network traffic flow short-time prediction method combining time-space characteristics
CN109887297B (en) Method for dividing urban traffic control subareas based on rapid global K-means spectral clustering
CN111932914B (en) Double-layer boundary control method for road network in urban congestion area
CN112629533B (en) Fine path planning method based on road network rasterization road traffic prediction
CN103996289B (en) A kind of flow-speeds match model and Travel Time Estimation Method and system
CN102033994B (en) Steering engine reliability simulation sampling method based on Markova chain Monte Carlo
CN109410574B (en) Phase-phase signal control scheme-oriented timing parameter optimization method
CN113538910B (en) Self-adaptive full-chain urban area network signal control optimization method
CN106910350B (en) A method of finding signalized crossing group critical path
CN113192318B (en) Data drive control regional traffic signal dynamic optimization method and system
CN110889967B (en) Overflow risk balance signal control optimization method based on main road segmentation
CN104700634B (en) The Adjacent Intersections arterial road coordinate control method of improved adaptive GA-IAGA is clustered based on minimum spanning tree
CN111192457A (en) Method and device for controlling urban-level integral traffic signals
CN108629970A (en) Intersection signal parameter optimization method based on the search of Monte Carlo tree
CN111554118B (en) Dynamic prediction method and system for bus arrival time
CN113838296B (en) Traffic signal control method, device, equipment and storage medium
CN114694382A (en) Dynamic one-way traffic control system based on Internet of vehicles environment
CN112258856A (en) Method for establishing regional traffic signal data drive control model
CN116681176A (en) Traffic flow prediction method based on clustering and heterogeneous graph neural network
CN116673947A (en) Mobile robot travel path point prediction method
CN114141013B (en) Intersection signal timing method based on low-permeability network connection track data
CN115240418A (en) Short-term traffic flow prediction method based on causal gating-low-pass graph convolution network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant