CN105809958A - Traffic control method and system based on intersection group - Google Patents

Abstract

The invention relates to a traffic control method and system based on an intersection group.The method comprises the steps that 1, 360-degree panoramic video of intersections is dynamically collected in real time through an intelligent robot, an intersection operation model is established according to video data, and traffic characteristics of the intersection group are analyzed according to the intersection operation model; 2, intersection index evaluation and online simulation analysis are conducted according to the traffic characteristics, and the traffic operation state of the intersection group is identified; 3, supersaturation state intersection signal timing control scheme optimization is conducted on critical paths of a supersaturation state intersection group, and a supersaturation state intersection group traffic signal control strategy is adjusted; 4, the adjusted intersection group traffic signal control strategy is operated, and steady-state operation of an intersection control signal timing optimization scheme and linkage command of the intelligent robot are achieved.By means of the traffic control method and system based on the intersection group, traffic efficiency and service level of intersection single point control can be improved, and therefore the operation efficiency of an urban traffic system is greatly improved, and urban traffic jams are relieved.

Description

A kind of traffic control method based on intersection group and system

Technical field

The invention belongs to technical field of traffic control, particularly relate to a kind of traffic control method based on intersection group and system.

Background technology

Intersection group refers in urban road network that geographical position is adjacent and exists gather compared with some crossings of High relevancy, and the impact of road network traffic circulation state is notable, is core node and the key point of urban traffic blocking and traffic safety.Intersection group existence form in urban road network is divided into the crossing of urban central zone, the crossing at two ends, urban inner tunnel, crossing that viaduct is adjacent, expressway to import and export the signalized crossing etc. of ring road and urban road street joint intersection, city expressway import and export.Intersection group constitutes the emphasis traffic zone of urban road network, is an up the key of urban traffic control performance, solves intersection group congestion problems and the traffic jam issue making whole section, city road network is obtained very big alleviation.Intersection group relatedness be mainly manifested in crossing spacing compared with short, critical path flow compared with big, wagon flow discreteness is little, the wagon flow of downstream intersection arrives distribution and presents wagon flow and form a team state, and upstream intersection situation can be subject to the impact of downstream queuing vehicle under certain condition.The proposition of intersection group concept, it is based on the needs that crossing coordinates to control the earliest, when Tongji University's state's urban highway traffic real-time adaptive control under study for action and management system, it is proposed to intersection group is coordinated to control and induction management, and it can be used as a function of system.Intersection group key definition includes:

Crossing hypersaturated state: when crossing both direction flow and crossing saturation volume ratio and during more than 1, when namely transport need exceedes its traffic capacity, the state of crossing is defined as hypersaturated state.

Intersection group hypersaturated state: when transport need is more than the traffic capacity of intersection group road network in intersection group, it is believed that intersection group is in hypersaturated state.The ratio (V/C ratio) utilizing intersection group entirety transport need and the traffic capacity judges whether crossing or intersection group block up.Equally, delay queuing can also be applied to define hypersaturated state, namely there is vehicle can not (green light be being queued up before starting by the situation of crossing within a green light cycle, still fail when green time terminates to pass through crossing), just this state of definable is hypersaturated state, and extend correlative factor: the degree (queue length) of hypersaturated state, the pace of change (queuing rate of increase) of hypersaturated state, hypersaturated state (stops overflow in the impact within intersection group, green light sky such as puts at the negative effect), the persistent period (persistent period) etc. of hypersaturated state.

Intersection group critical path: in the signal of intersection group controls, path is the sequence of a crossing in intersection group so that have a section to arrive the next crossing of this sequence from its each crossing.In view of in intersection group, crossing is limited, all paths in intersection group are finite path, all there is starting point crossing and terminal crossing in every paths, corresponding crossing flows to and is defined as starting point and flows to and terminal flows to, and the crossing of path process is defined as in path crossing.Intersection group critical path refers to that the volume of traffic in intersection group is maximum and determines the path of intersection group overall operation efficiency, in intersection group critical path, the change of any road section traffic volume service level all can externally within the scope of intersection group other paths produce impact, it is easy to produce to block up.

For a long time, urban traffic blocking, intersection group run the traffic control problems based on intersection group of hypersaturated state, do not think deeply from sight macroscopical, middle, microcosmic integrated aspect in actual applications and solve the name of the game origin cause of formation.Prior art controls for urban traffic safety, the crossing that blocks up, is only by single technological means and promotes, as: crossing modeling analysis, or intersection simulation, or crossing optimization etc..Based on this, through studying discovery for a long period of time: run in intersection group in hypersaturated state, first urban road intersection single-point runs and does not reach optimum state；Secondly, crossing morning, evening peak and the flat peak period that urban transportation hypersaturated state is run, floor manager majority adopts artificial mode, is directly affected by conditions such as weather, season, time, air, healths.Like this, can not meet for a long time, continuously Science and coordination commander integrative design intersection, improve crossing operational efficiency, owing to traffic pollution reason damages again the physical and mental health of crossing commanding, this problem should be solved.Enter the generation information Age of Technology such as the big data of traffic, cloud computing, according to domestic and international correlation technique current situation, run all many-sides all in the urgent need to perfect and lifting obtaining commander crossing after intersection group scope, hypersaturated state generation time, intersection group critical path, traffic flow parameter saturation characteristics, control signal machine and intelligent robot link.

At present, traffic administration under hypersaturated state and control method had correlational study, but identified that crossing hypersaturated state is few, in the traffic administration strategy of existing hypersaturated state, generally assume that crossing obtains big flow it is known that thus obtain the traffic behavior of crossing；And under hypersaturated state, it is impossible to provide enough effective data to identify the traffic behavior of crossing.

nullCurrently，The urban traffic signal control system major part of domestic and international main flow adopts hierarchical control structure by different level，SCOOT (SplitCyeleOffsetOptimizationTechnique) such as Britain、The SCATS (SydneyCoordinatedAdaptiveTrafficSystem) of Australia、The STREAM (StrategicReal-timeControlforMegalopolis-traffic) of Japan、The MOTION (MethodfortheOptimizationofTrafficSignalsInOn-LineControl ledNetwork) etc. of Germany.The control structure of Subsystem Based on Layered Structure Model stepwise is generally divided into organized layer, cooperation layer, key-course, and wherein cooperation layer is region class control.The SCATS of the SCOOT of Britain and Australia broadly falls into static partition control strategy, the difference of the two be mainly subregion after adjacent subarea territory merge different from the strategy separated, the SCOOT of Britain can not merge, the SCATS of Australia can merge, the shortcoming of the two is static partition control strategy, it is impossible to adapt to the dynamic change of urban road traffic network traffic flow OD distribution.The domestic not unalterable ground of other pattern is complete to be introduced.

In sum, urban road intersection group or similar concept carry out much research both at home and abroad discuss, the main contents discussed include intersection group concept, scope defines, traffic associate feature, traffic coordinating and controlling method, and the control strategy and technique study for hypersaturated state intersection group also rests on the primary stage.Particularly under the big data of traffic, cloud computing environment, both made to be discussed about traffic control under hypersaturated state and traffic are modeled, but major part research work is only absorbed in and how to detect delay that hypersaturated state can bring or in hypersaturated state drag effect, it is thus achieved that the formula of similar HCM form or workflow.The most important thing is overlength produced by hypersaturated state is queued up the degree of supersaturation (V/C than) being managed or controlling whole road network for the operation of traffic flow under management hypersaturated state；The strategy that the existing queuing that hypersaturated state is produced is managed, major part is that the queue length that the senior detector according to downstream intersection can detect is estimated, but queuing exceedes test point seldom to have model to predict, the situation of even whole road section length.Existing research can realize not utilizing outlet detector just can estimate hypersaturated state queuing situation；Great majority are based on adaptive control system development for the research of hypersaturated state traffic control；These systems can must work under hypersaturated state, or at least can effectively run under hypersaturated state.But being because existed system great majority is commercial system, the rarely found periodical of document of method detailed about adaptive traffic signal control system hypersaturated state estimated and control.

For the supersaturation characteristic of traffic flow, the theoretical optimization strategy of some traffic signalization optimizations or algorithm are suggested.The off-line traffic signalization that part is commonly used optimizes software (such as PASSER and TRANSYT) and have developed the optimization method of the signal period of hypersaturated state, split, phase contrast.These theoretical optimization strategies and algorithm main disadvantage is that the flow requiring road when algorithm is applied must be known, namely the traffic parameter such as hypersaturated state down-off must be measurable, but traffic parameter not easily obtains under hypersaturated state.Theoretical algorithm does not discuss the algorithm suitability in zones of different, it is impossible to simple directly application.Only part research proposes the hypersaturated state traffic control strategy that can carry out practical engineering application, because the restriction of downstream intersection, recent study is paid close attention to the signal of single-point intersection more and is controlled timing designing, or identify and queue up and method of adjustment to signal timing dial parameter for removing added turning lane, these methods have positive role for being undertaken classifying by signal control strategy, but must carry out the definition of system.Traffic control strategy is proposed the strategy such as application " Reverse Commuting control " or " secondary road application blue flash lamp ", but effect produced by these strategies and mechanism is not thoroughly discussed and analyzed.Adaptive control system SCOOT, SCATS and RHODES also only discuss the control strategy of macroscopic view.Such as the closure in the upperreaches control strategy of SCOOT, it is only discuss control principle in existing document, but does not provide the restriction being applied in actual control system.Therefore, to representative having in hypersaturated state traffic control theoretical research:

1. strategy (the RealTime/InternalMeteringPolicytoOptimizeSignalTiming that dams in real time of traffic signalization optimization, RT/IMPOST) oversaturated arterial highway network it is primarily adapted for use in, RT/IMPOST increases by restricting the flow of the traffic volume control oversaturated intersection import department in section, upstream, and this method is sufficiently used the storage capacity of road network.

2. maximum current strategy is mainly adjusted by different signal timing plans to make being open to traffic of oversaturated intersection several maximum.That applies this type strategy mainly has Zhoucheng city, Texas signal control strategy (TexasUrbanDiamondSignalControl), Arlington's control strategy (ArlingtonApproach), KimMesser control strategy.

3. the phase optimization method preventing overflow can be applicable to trellis state urban road network, and this control strategy is once in the CBD certain applications of New York, United States Manhattan, and makes total travel time reduce by 20%.

Chinese scholars proposes the sub-district of several traffic control/intersection group scope and defines algorithm, hypersaturated state and recognizer, bottleneck road distinguished number and traffic control strategy and intelligent algorithm, alleviates the traffic congestion of urban road network.But the intrinsic propesties such as intersection group traffic characteristics and capacity under hypersaturated state mostly is qualitative analysis by existing research, fails to see its feature of announcement of quantification；The traffic control strategy proposed mostly is the solution for practical problem, does not possess universality；Corresponding signal Controlling model and algorithm mostly also are theory study, fail to be verified application in real road network.

In sum, the shortcoming of prior art is mainly manifested in the following aspects:

1. intersection group is coordinated span of control and is failed to embody the Real-time and Dynamic change of its traffic relatedness；The existing research of prior art has realized that the linked character of intersection group is not limited only by the impact of crossing spacing, also relevant with the traffic circulation characteristic of the intersection groups such as wagon flow distribution characteristics, signal timing plan.In Practical Project runs and uses, the scope of intersection group and traffic coordinated control is all dynamically change, and conventional cross mouth group's range determining method intelligence degree is not high, only according to historical data static division, and do not consider the topological relation of road network, it is necessary to the judgement of intersection group Relating Characteristic and intersection group scope is carried out from new knowledge.

2. intersection group hypersaturated state is difficult to；

In hypersaturated state intersection group, transport need is more than its traffic capacity, and the queuing of crossing is long even overflows, and makes convention traffic detection method can not accurately detect real-time traffic service data.Because the traffic control strategy of hypersaturated state and the traffic control strategy of stable state are different, if hypersaturated state initial time can not be accurately identified, will affect traffic control optimized algorithm application effect.

3. the method lacking quantitative analysis hypersaturated state critical path；

Intersection group is carried out signal coordinated control as entirety and obtains approval and the concern of scholar, but existing traffic control strategy is generally based on global optimization or the regulation of crucial crossing, the collaborative path chosen in optimization process is generally artificial appointment, fails the identification to the critical path within the scope of intersection group and classification and carries out systematic study and application.

4. traffic coordinated control algorithm fails according to hypersaturated state intersection group traffic characteristics optimization；

Intersection group requires that traffic signal control system must take into account the harmony between Adjacent Intersections, optimizes the signal timing plan of all signalized intersections in high density road network；Additionally due to intersection group Adjacent Intersections spacing is little, between Adjacent Intersections, traffic flow influences each other bigger.

Summary of the invention

The invention provides a kind of traffic control method based on intersection group and system, by introducing the big data of traffic and cloud computing technology, set up the traffic behavior optimized needed for hypersaturated state intersection group traffic control, set up the city traffic control intelligent robot based on intersection group, thus solving the problems referred to above of the prior art at least to a certain extent.

Implementation of the present invention is as follows, a kind of traffic control method based on intersection group, comprises the following steps:

Step a: gather 360 °, crossing panoramic video by intelligent robot Real-time and Dynamic, sets up crossing moving model according to video data, and analyzes intersection group traffic characteristics according to crossing moving model；

Step b: carry out crossing assessment index and in-circuit emulation analysis according to traffic characteristics, identifies intersection group traffic circulation state；

Step c: the critical path of hypersaturated state intersection group is carried out oversaturated intersection signal timing dial design approach, adjusts hypersaturated state intersection group traffic signal control strategy；

Step d: the intersection group traffic signal control strategy after combustion adjustment, it is achieved crossing control signal timing designing scheme steady-state operation and intelligent robot interlinked command.

The technical scheme that the embodiment of the present invention is taked also includes: described step a also includes: crossing moving model is carried out operation situation monitoring；Described operation situation monitoring method includes: analysis intersection group blocks up and forms and evacuate collection and the process of mechanism and traffic circulation parameter；Described traffic circulation parameter acquisition and processing method specifically include: video encoder server and the modeling of traffic relatedness index.

The technical scheme that the embodiment of the present invention is taked also includes: in described step b, and described identification intersection group traffic circulation state specifically includes: intersection group scope defines, intersection group hypersaturated state identification, intersection group critical path detection and traffic parameter short-term prediction modeling and simulation.

The technical scheme that the embodiment of the present invention is taked also includes: in described step c, and the described critical path to hypersaturated state intersection group carries out oversaturated intersection signal timing dial design approach mode and specifically includes: intersection signal timing controls prioritization scheme static optimization；Dynamic cooperation traffic signalization intersection group；The traffic control strategy of hierarchical screening hypersaturated state intersection group；Coordinate timing scheme based on non-dominated sorted genetic algorithm optimization, control the benchmark timing scheme of dynamic optimization as signal；Traffic parameter real time dynamic optimization algorithm.

The technical scheme that the embodiment of the present invention is taked also includes: in described step d, and the method for described integrative design intersection prioritization scheme and intelligent robot interlinked command includes: the quiet collaborative traffic control of urban road oversaturated intersection group motion；Intersection group critical path coordinates the selection in control cycle；The phase contrast on-line optimization of oversaturated intersection group's critical path；Mixed traffic flow impact reasonable consideration in minimax green time and copper sulfate basic retrain on Split Optimization；Set up new intersection signal timing and control collaborative linkage and commander's operational mode.

The technical scheme that the embodiment of the present invention is taked also includes: described intersection group critical path is coordinated the Cycle Length computing formula of control cycle selection and is:

In above-mentioned formula, L is road section length；W is crossing, upstream width；Ga is the effective green time of downstream intersection；H for from when sailing vehicle headstock from；L is lost time；Lu is average traffic effective wagon degree；RL is shock wave dissipation place；C1 is the Cycle Length preventing overflow；SF is vehicle safety coefficient when emptying；U is from the velocity of wave sailing shock wave；V is the speed of next wagon flow first car；ω is the velocity of wave of parking shock wave；Δ is for coordinating to control phase contrast.

Another technical scheme that the embodiment of the present invention provides is: a kind of traffic control system based on intersection group, including intelligent robot, described intelligent robot includes the first video camera module, the second video camera module and data processor module；Described first video camera module and the second video camera module are connected with data processor module respectively；nullDescribed first video camera module and the second video camera module gather 360 °, crossing panoramic video for Real-time and Dynamic，And by the video data transmission of shooting to data processor module，Described data processor module is for setting up crossing moving model according to video data，Intersection group traffic characteristics is analyzed according to crossing moving model，Crossing assessment index and in-circuit emulation analysis is carried out according to intersection group traffic characteristics，Identify intersection group traffic circulation state，Thus the critical path of hypersaturated state intersection group is carried out oversaturated intersection signal timing dial design approach，Adjust hypersaturated state intersection group traffic signal control strategy，And control the intersection group traffic signal control strategy after intelligent robot combustion adjustment，Realize crossing control signal timing designing scheme steady-state operation and intelligent robot interlinked command.

The technical scheme that the embodiment of the present invention is taked also includes: described first video camera module is the 360 ° of panoramic high-definition video cameras that highly can stretch, it is located at the above-head of intelligent robot, described second video camera module is HD video video camera, is located at the eye of intelligent robot.

The technical scheme that the embodiment of the present invention is taked also includes: described data processor module includes model and sets up unit, traffic characteristics analytic unit, traffic circulation state recognition unit, policy optimization unit and scheme running unit；

Model sets up unit for receiving the first video camera module and the video data of the second video camera module transfer, and carry out video data sorting out after screening, image recognition and feature extraction etc. process and generate crossing real-time dynamic information environment, set up that picture is clear, the crossing moving model of broad view；

Traffic characteristics analytic unit for carrying out operation situation monitoring to crossing moving model, and analyzes intersection group traffic characteristics according to crossing moving model；

Traffic circulation state recognition unit, for carrying out crossing assessment index and in-circuit emulation analysis according to traffic characteristics, identifies intersection group traffic circulation state；

Policy optimization unit, for the critical path of hypersaturated state intersection group is carried out oversaturated intersection signal timing dial design approach and induction, adjusts hypersaturated state intersection group traffic signal control strategy；

Scheme running unit is for the intersection group traffic signal control strategy after combustion adjustment, it is achieved crossing control signal timing designing scheme steady-state operation and intelligent robot interlinked command.

The technical scheme that the embodiment of the present invention is taked also includes: described intelligent robot also includes display module, described display module is for touching display screen, it is positioned at the body part of intelligent robot, described first video camera module and the second video camera module are connected with display module respectively, described first video camera module and the second video camera module are by the video data transmission of shooting to display module, and described display module is for showing the first video camera module and the video data of the second video camera module photograph.

nullThe traffic control method based on intersection group of the embodiment of the present invention and system are monitored in real time by building 360 ° of crossing panoramic videos and are modeled、Crossing assessment index and in-circuit emulation analysis、Oversaturated intersection critical path and control strategy optimization、Integrative design intersection optimizes " the four step methods " method with intelligent robot interlinked command，Set up the city traffic control intelligent robot based on intersection group，Solving urban road oversaturated intersection single-point operation optimization problem，Form intellectuality commander's urban road intersection、Oversaturated intersection、Oversaturated intersection group's traffic control and prioritization scheme，And adopt the traffic control intelligent robot based on intersection group，Realize intelligent robot and traffic signal controlling machine linkage，Set up integrative design intersection robot service mode，Promote traffic efficiency and service level that crossing single-point controls，Be conducive to scientifically and rationally the traffic flow of urban road network being carried out dynamic monitoring and optimizing tissue，Thus increasing substantially the operational efficiency of Traffic Systems，Alleviate urban traffic blocking.

Accompanying drawing explanation

Fig. 1 is the flow chart of the traffic control method based on intersection group of the embodiment of the present invention；

Fig. 2 is the flow chart of the method identifying intersection group traffic circulation state of the embodiment of the present invention；

Fig. 3 is the traffic signalization Optimizing Flow figure of the hypersaturated state of the embodiment of the present invention；

Fig. 4 is the oversaturated intersection critical path schematic flow sheet with control strategy optimization method of the embodiment of the present invention；

Fig. 5 is the intersection group traffic control dynamic optimization method frame diagram of the embodiment of the present invention；

Fig. 6 is the Cycle Length computational methods schematic diagram preventing overflow of the embodiment of the present invention；

Fig. 7 is the structural representation of the traffic control system based on intersection group of the embodiment of the present invention；

Fig. 8 is the structural representation of the data processor module of the embodiment of the present invention；

City Road Network and associated cross mouth group static models figure centered by Fig. 9；

Figure 10 is that intersection group dynamic traffic control Current Situation analyzes schematic diagram；

Figure 11 is that Lianhua Road signalized crossing carries out dynamic traffic control optimization schematic diagram；

Figure 12 is that Hong Li road signalized crossing carries out dynamic traffic control optimization schematic diagram.

Detailed description of the invention

In order to make the purpose of the present invention, technical scheme and advantage clearly understand, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein is only in order to explain the present invention, is not intended to limit the present invention.

Refer to Fig. 1, be the flow chart of the traffic control method based on intersection group of the embodiment of the present invention.The traffic control method based on intersection group of the embodiment of the present invention comprises the following steps:

Step 100: gather 360 °, crossing panoramic video by intelligent robot Real-time and Dynamic, and set up crossing moving model according to video data；

In step 100, the above-head of intelligent robot is provided with the 360 ° of panoramic high-definition video cameras that highly can stretch, the eye of intelligent robot is HD video video camera, 360 °, crossing panoramic video is gathered by 360 ° of panoramic high-definition video cameras and HD video video camera Real-time and Dynamic, panoramic video for gathering carries out sorting out after screening, image recognition, feature extraction etc. process and generates crossing real-time dynamic information environment, set up picture clearly, the crossing moving model of broad view.

Step 200: crossing moving model is carried out operation situation monitoring, and analyzes intersection group traffic characteristics according to crossing moving model；

In step 200, carry out the big data cross mouth moving model of traffic to demarcate and operation situation monitoring, and according to the intersection group traffic characteristics such as traffic flow discrete feature and traffic signalization characteristic between crossing global model analysis intersection group geometry topological structure characteristic, path space characteristic, crossing, analyze traffic flow operation characteristic and traffic circulation data acquisition and processing method in intersection group, as the basis of traffic status identification and traffic signalization.Specifically, the method that crossing moving model carries out operation situation monitoring comprises the following steps:

Step 201: analysis intersection group blocks up and formed and evacuate mechanism；

Specifically include: analyze the risk factor that intersection group blocks up, it is determined that crossing overflow, green light sky are put, are detained the harmful effects such as the queuing impact for intersection group traffic congestion, it is determined that the process that hypersaturated state is formed；Judging traffic flow running rate when traffic flow bottleneck dissipates, application transportation network load balancing theoretical description congestion status evacuates the traffic flow character of process, for analyzing the traffic behavior based theoretical of hypersaturated state intersection group.

Step 202: the collection of traffic circulation parameter and process；

Specifically include: determine the traffic circulation parameter analyzed required for the logical running status of urban road intersection flock-mate, the pluses and minuses of the various traffic circulation parameter collecting method of comparative analysis and the adaptability to hypersaturated state traffic signalization, it is preferable that the Data Source needed for intersection group traffic status identification and traffic control；Set up traffic circulation parameter cleaning treatment method, it is determined that the algorithm of Data-parallel language, the differentiation of traffic flow wrong data, correction and traffic flow Reduction for Redundant Data is lost in traffic flow, lays the foundation for traffic state analysis.The traffic circulation parameter acquisition of the embodiment of the present invention and processing method specifically include video encoder server and the modeling of traffic relatedness index；Wherein,

The concrete mode of video encoder server is:

1) moving target candidate region is extracted, it is determined that vehicle region that may be present；

2) goal verification, the candidate region that the upper stage was produced confirms, it is judged that be vehicle or background；

3) Target Segmentation, by identifying the pixel meeting vehicle characteristics in image, separates target to be identified from background；

4) target following, according to the vehicle in frame before and after the characteristic matching extracted, thus calculating traffic circulation parameter；

5) target classification, classifies to different types of vehicle according to geometric shape, textural characteristics etc.；

6) later stage processes, and calculates traffic circulation parameter according to detection demand, such as vehicle flowrate, speed etc..

Traffic relatedness index includes discreteness coupling index and retardancy coupling index；

Discreteness coupling index is: by the impact of the discrete factor of wagon flow, downstream intersection all passes through crossing to the first bus and last bus ensureing fleet in same green time, then need to design the green wave band that broadens of a kind of diffusion type.But this design can make the green time of the crossing of most downstream look cannot be accepted, and is a kind of control mode that discreteness is thrown the reins to, often inadvisable in practical engineering application.The wide green ripple of the many employings of control method to discrete constraint, but the method can make the Some vehicles being positioned at wagon flow stem or afterbody can have certain delay at each crossing.Set discreteness relatedness index I1 rise as path in the signal control cycle, the settled isometric green time of the some ratio by vehicle, it may be assumed that

In formula (1): q0 (i) represents the wagon flow of a certain paths initial upstream intersection parking i period of line by number；The wagon flow of qd (i+T) delegated path terminal chiasma mouth the i-th+T period arrives number；T represents the running time from path starting point to the end；Tg represents the green wave duration in the signal period.Q0 (i) and qd (i+T) can adopt field observation value, it is also possible to is calculated by Robertson motorcade dispersion formula, it may be assumed that

In formula (2): the wagon flow of qd (j) delegated path terminal chiasma mouth jth period arrives number, t=β T=β (j-i), coefficient of dispersionα, β represent undetermined parameter, and Robertson advises value respectively 0.35 and 0.8.

Retardancy coupling index is: intersection group is formed to any section m on certain road, if there is N number of different flow direction in the crossing inlet road along this path direction of advance, calculates the functional areas length value of each flow direction Queue lengthField observation statistical value can be adopted, it is possible to use queue length computing formula is estimated, the present invention adopts the queue length computational methods of Synchro7, deceleration distanceWith perception-reaction distanceComputational methods, willIt is defined as section m along the crossing inlet road of path direction of advance, flows to the ratio of functional areas length maximum and path L, it may be assumed that

If this path is made up of M section, then its retardancy index I2 is:

Pass through above step, realize static/dynamically suitable, look down comprehensive 360 ° of modelings of the overall situation, intersections planning design objective, the currently running index in crossing, the crossing optimization generation other index of index three major types are compared and studied and judged, complete crossing Real-time and Dynamic operational monitoring, form round-the-clock continuously model optimization self adaptation periodic duty, gather and converge the round-the-clock three phases morning peak in access crossing, Ping Feng, evening peak unlike signal control model and combination control model.Real-time and Dynamic converges and accesses the big data of traffic; crossing line modeling is carried out in synchronization; the different crossings data that fusion structure, semi-structured, destructuring gather; optimize and improve crossing moving model; complete crossing dynamic model; build the information source pond of territory, the whole city big data of intersection traffic, by each season annual, each, each month, every day monitoring and model dynamic calibration, it is achieved crossing real-time dynamic monitoring modeling machines people.

The concrete analysis mode that intersection group traffic characteristics is analyzed of the embodiment of the present invention is: understand the traffic characteristics of intersection group respectively from aspects such as intersection group geometry topological property, path space characteristic, traffic stream characteristics, traffic signalization characteristics, find the variation characteristic of traffic flow in intersection group, provide foundation for application supersaturation traffic control strategy.Wherein, intersection group is classified by intersection group geometry topological property according to the road path number feature between in intersection group two crossings；Traffic flow can be run the impact produced by the design of path space specificity analysis road traffic facility；Traffic stream characteristics gives suitable in the descriptive model of cutout between hypersaturated state urban road, according to intersection group traffic stream characteristics, chooses suitable traffic circulation data acquisition means, sets up data cleansing and processing method.Traffic signalization specificity analysis basic control principle and control structure, lay a good foundation for setting up traffic control method.

Step 300: carry out crossing assessment index and in-circuit emulation analysis according to traffic characteristics, identifies intersection group traffic circulation state；

In step 300, in intersection group assessment index and reproduction simulation analysis, for the intersection group traffic status identification of hypersaturated state traffic signalization, mainly include intersection group scope define, intersection group hypersaturated state identification, the critical path detection of intersection group, four aspects of short-term traffic flow Parameters variation Predicting Performance Characteristics.The traffic circulation parameter required when the evaluation of crossing running status specifically includes that speed, flow, occupation rate etc., mainly includes exponential smoothing, California algorithm, McMaster algorithm, SND method, cross-correlation method, Kalman filtering method etc. for traffic congestion state automatic decision algorithm.Specifically, in order to know explanation step 300, see also Fig. 2, be the flow chart of the method identifying intersection group traffic circulation state of the embodiment of the present invention.The method identifying intersection group traffic circulation state of the embodiment of the present invention comprises the following steps:

Step 301: define intersection group scope；

Building intersection group in object road network, it is necessary to set up the elemental range analyzed and solve intersection group traffic problems, defining of intersection group scope is intersection group to be carried out to traffic status identification and the prerequisite of traffic control optimization.Coordinating in control at urban road network main signal, the crossing of intersection group aspect pair coordinates control, it is possible to obtain the traffic circulation shape improving within the scope of intersection group significantly.Traffic signaling equipment computing capability with calculate deficiency of time with the optimization traffic control scheme of the whole road network of direct solution; when being often absorbed in local optimum; whole road network is divided into several intersection groups by algorithm to utilize intersection group scope to determine, and then optimizes its traffic control strategy and be by the practical way of traffic coordinated control.The coordination of intersection group controls between single-point control and Region control, and its scope should meet the hsrdware requirements of traffic signaling equipment, and can select the traffic control strategy of optimum at short notice.The principle that intersection group scope defines is as follows:

1) crossing having relatively High relevancy should be divided into an intersection group, and the crossing that relatedness is not strong should be divided in different intersection groups；

2) in urban road network, the crossing number in each intersection group should be roughly equal, and meets the hsrdware requirements of traffic controlling machine；

3) time complexity of algorithm is low, and committed memory to lack；

4) the result reply traffic flow that scope defines runs the impact having front.

In embodiments of the present invention, the method defining intersection group scope specifically includes: judging on the basis of intersection group spatial character and internal association mechanism, analyze the traffic relatedness between crossing in intersection group, set up the intersection group scope confining method of feature based matrix and based on the intersection group scope confining method of self organizing neural network.Application vehicle queue length describes intersection group linked character with the ratio of line crossing space length and the Quality degree of effective utilization of green time respectively, the former is in conjunction with flow factor and distance factor, the latter takes into account flow factor and timing factor, the various characteristic analysis method of integrated application, defines the scope of intersection group.

Step 302: identify and assessment intersection group hypersaturated state；

Understand formation and the evacuation mechanism of intersection group hypersaturated state, provide theoretical foundation for hypersaturated state intersection group traffic signalization.In embodiments of the present invention, the recognition methods of intersection group hypersaturated state identification and assessment index is: based on the method analyzing intersection group degree of super saturation, propose the ratio applying invalid green time and the total green time caused by negative effect to define degree of super saturation index, and weigh intersection group degree of super saturation with this.Based on hypersaturated state intersection group characteristic of produced negative effect on Spatial Dimension and time dimension, calculate the degree of super saturation index of intersection group respectively in room and time dimension.By shock wave model and space-time diagram on Spatial Dimension, what produce when the shock wave produced when dissipating by queuing up and green ripple start calculates crossing maximum queue length from sailing shock wave, the parking shock wave produced when the shock wave produced when dissipating by queuing up and lower cycle red light start calculates the tributary queue length of crossing, calculates the degree of super saturation coefficient of Spatial Dimension with this.At time dimension, the upstream detector mainly through being produced by crossing queuing overflow occupies phenomenon for a long time to calculate the degree of super saturation coefficient of crossing.The degree of super saturation coefficient of general space dimension and time dimension, identifies the hypersaturated state of intersection group.

Hypersaturated state directly can not be measured by traffic parameter or calculate identification, indirectly obtains only by negative effects such as overflows produced by hypersaturated state.For the hypersaturated state of quantitative judge intersection group, the definition of intersection group hypersaturated state is extended, calculate supersaturation coefficient by the negative effect caused by hypersaturated state, so that it is determined that the hypersaturated state of intersection group.Hypersaturated state refers to the situation when a means of transportation generation transport need by traffic signalization is more than its traffic capacity state (green time maximum by number), its negative effect that by the delay queuing in certain cycle, negative effect or the upstream means of transportation in next cycle can be produced within a cycle because overflow goes out defines, and the ratio (supersaturation coefficient) applying invalid green time and total green time weighs degree of super saturation.

In embodiments of the present invention, adopting the hypersaturated state of induction coil Vehicle Detection data assessment intersection group, induction coil typical case lays mode and includes stop line detector and senior detector (laying in stop line upstream) two kinds.Under hypersaturated state, intersection group is queued up longer, all can not accurately detect, regardless of stop line detector or senior detector, the traffic organization identifying hypersaturated state crossing, it is necessary to identify the hypersaturated state of intersection group with method for parameter estimation.The negative effect that under application hypersaturated state, traffic control produces in space-time unique replaces traditional method of estimation to evaluate the state of means of transportation.Viscous flow queue length when the identified negative effect of algorithm mainly has the signal period to terminate and the overflow phenomena of crossing, upstream, two kinds of negative effects all can cause the effective green time of signalized intersections to reduce.Queue length is detained in the method estimation crossing adopting shock wave (Shockwave), according to the detector high occupation rate (QueueOverDetector that queuing vehicle extended stationary periods face on the detector is caused, QOD) overflow phenomena in phenomenon identification intersection group, identifies the hypersaturated state of intersection group further.

The calculating of shock wave velocity of wave, if velocity of wave (u2, u3, u4) is also used to calculate the maximum queue length in the cycle, because traffic arrives flow rate variance relatively greatly, queuing shock wave (u1) is not suitable for estimation queue length.Selecting from making shock wave (u2) and deviating from shock wave (u3) and estimate queue length, computing formula is:

In formula (4): flow rate when qm and km represents maximum flow respectively and density, kj represents jamming density,WithRepresent traffic arrival rate and corresponding density.WithRefer to the traffic flow modes through detector after time Tc, when solving u2 it is assumed herein that qm, km and kj are fixed value, compression shock ripple u4 and have identical velocity of wave from sailing shock wave u2.

High-resolution traffic data is used to estimation and includesQm, km at interior various traffic variations, wherein traffic flow rate data, asWith qm can directly by detector acquisition, butKm isodensity data must be estimated.The independent holding time can be provided, it is assumed that effectively vehicle commander is it is known that space mean speed can be obtained based on the traffic data of event；Now available AFR estimates density data divided by space mean speed.Estimating individual speed ui, space mean speed us, the method for flow rate q and density k is:

In formula (5) to formula (8): t0, i and tg, i represents detector holding time and the interval of vehicle i, ui and hi represents speed and the space headway of vehicle i, q, us and k represent AFR, space mean speed and density respectively, Le represents effective vehicle commander, and n represents the vehicle number of a fleet in same traffic behavior.It is detained queue length and degree of super saturation Index for Calculation, the maximum queue length in the n-th cycleWith the moment reaching maximum queue lengthFor:

In formula (9) and formula (10): Ld represents stop line to the distance between detector.

Step 303: the detection of intersection group critical path and classification；

The critical path of intersection group is the section occurred frequently of traffic congestion, also it is the bottleneck road of intersection group, the path grade of intersection group is analyzed according to real-time dynamic information, identify the critical path of intersection group, intersection group traffic control can be made the more efficient traffic flow to intersection group to be optimized, based on the feature that fleet's traffic relatedness in intersection group is strong, adopt the intersection group critical path identification way based on wavelet transformation and spectrum analysis, analyze and extract intersection group traffic flow variation characteristic in short-term, utilize the critical path of the method detection intersection group of data mining analysis, for intersection group path ranking.In conjunction with the characteristic that intersection group critical path upstream and downstream wagon flow dispersion degree is little, traffic signal are pressed different frequency and are decomposed by application wavelet transformation technique, retain the high-frequency signal of reflection traffic flow variation characteristic in short-term and the low frequency signal of reflection traffic flow basis variation characteristic, filtered traffic signal are reconstructed the new traffic signal highlighting traffic flow variation characteristic in short-term, as the input data of critical path identification and classification.Each import of intersection group of calculating wavelet transformation reconstruct flows to the cross-spectral density between the power spectral density of traffic signal and the flow direction.The degree of association of two traffic signal is determined by calculating the consistency coefficient of each cross spectrum, obtain correspondence and specify the criticality index in all paths of import, again through the phase place calculated between two signals, the travel time checking being aided with at 2 calculates effectiveness, the comprehensive significance level analyzing all import critical paths.

In intersection group, the power of intersection traffic relatedness is mainly manifested between crossing wagon flow dispersion degree size, the i.e. similarity arriving stream characteristics and upstream stream characteristics of downstream intersection.The performance in critical path of this similarity becomes apparent from, once related grouped intersections middle and upper reaches crossing causes the traffic flow parameter such as flow, speed to change because of traffic signalization or traffic congestion, High relevancy according to relatedness Adjacent Intersections, the variation characteristic in short-term of traffic flow parameter can keep to downstream intersection.Under hypersaturated state, because wagon flow passes through crossing with saturation volume rate always, when in section, the dispersion degree of traffic flow running parameter is than stable state less, crossing is respectively flowed to the variation characteristic in short-term of traffic flow parameter as foundation, the critical path of Model Identification hypersaturated state intersection group can be set up.Model is it needs to be determined that suitable traffic parameter is to describe wagon flow feature, and chooses the variation characteristic in short-term of appropriate data digging method extraction wagon flow.

The feature little in order to highlight intersection group critical path upstream and downstream wagon flow dispersion degree, use small wave converting method that traffic signal are pressed different frequency to decompose, retain the high-frequency signal of reflection traffic flow variation characteristic in short-term and the low frequency signal of reflection traffic flow basis variation characteristic, filtered traffic signal are reconstructed the new traffic signal highlighting traffic variation characteristic in short-term, as the input data of critical path identification and classification.Wavelet transformation (WaveletTransformation) is the localization analysis of time (space) frequency, signal (function) is progressively carried out multi-scale refinement by flexible shift operations by it, it is finally reached high frequency treatment time subdivision, the frequency segmentation of low frequency place, can automatically adapt to the requirement that time frequency signal is analyzed, thus any details of signal can be focused on, solve the difficult problem of Fourier transformation.Wavelet transformation is that a kind of window size is fixed and its shape is variable, the time frequency resolution that time window and frequency window can change, and HFS has higher temporal resolution and relatively low frequency resolution.

Wavelet transformation is inherited and has developed the thought of Short Time Fourier Transform localization, can overcome again window size simultaneously and can not provide a T/F window with frequency shift with shortcomings such as frequency changes, time carry out the ideal tools of signal time frequency analysis and process.It is mainly characterized by by convert can the feature of abundant some aspect of outstanding problem, be obtained for successful application in a lot of fields.

Wavelet transformation is and is analysed to signal and is launched into the weighted sum of family's small echo machine, and its implication is morther wavelet (MotherWavelet) functionAfter making displacement τ, then make inner product with signal f (t) to be analyzed under different scale α:

In formula (11): α represents scale factor, α > 0；τ represents displacement, and its value can just can be born；Represent wavelet function and displacement thereof and yardstick stretches.

For the degree of association of quantitative Analysis intersection group each path upstream and downstream traffic flow, the method adopting spectrum analysis, using traffic flow change as input signal, analyze its spectral change feature at different frequencies.By calculating the cross-spectral density of each crossing inlet traffic signal, analyze the consistency coefficient of its signal, to determine the degree of association of two traffic signal, and apply the phase contrast of two signals, with the effectiveness of evaluation algorithm.

Frequency spectrum refers to the signal of time domain representation under frequency domain, it is possible to carries out Fourier transformation for signal and obtains, and the conclusion of gained is respectively with amplitude or phase place for the longitudinal axis, and frequency is transverse axis.Representing, with amplitude frequency spectrum, the situation that amplitude changes with frequency, phase frequency spectrum represents the situation that phase place changes with frequency.Frequency spectrum can represent that a signal is made up of the string ripple of which frequency, it is also possible to the information such as size and phase place of finding out each frequency string ripple.Spectrum analysis is a kind of technology that sophisticated signal is decomposed into relatively simple signal, finds out the way of signal information (such as amplitude, power, intensity, phase place etc.) at different frequencies and ascends the throne spectrum analysis.

Power spectrum is the sign of sequence power distribution properties on a different frequency digit time, if time series auto-covariance function γ_kSatisfy conditionThen there is following corresponding relation between power spectral density f (μ) and γ k:In formula: f (μ) defines on [-π, π], is real-valued nonnegative function.

Step 304: traffic parameter short-term prediction modeling and simulation；

According to hypersaturated state traffic stream characteristics it can be seen that conventional traffic flow model can not calculate following traffic behavior either directly through model.The variation characteristic of the EXSMOOTH of application enhancements, state space neural network, EKF method, data fusion method prediction intersection group traffic parameter in short-term.By utilizing the traffic data of present period and historical period, the traffic data of subsequent period being predicted, model is not by the restriction of hypersaturated state.Traffic parameter short-term prediction has important effect in dynamic traffic control algorithm designs, it was predicted that precision have appreciable impact for the effectiveness of traffic control algorithm.The difference of the basic mode according to prediction, short-term traffic flow forecasting model is divided into data-driven and based on model two types.The method of the method mathematical statistics of data-driven or artificial intelligence processes, and such as historical traffic data such as traffic flow, traffic speed, hourages, and predicts the change of future time period traffic flow；Mainly apply traffic flow propagation model based on the method for model the traffic flow modes on Xue Ding path is estimated and predicts, according to the careful degree that traffic flow is described by model, model can be divided into macromodel, mesoscopic model, micromodel three kinds.The method form being applied to traffic parameter short-term prediction is various, and effect is different, adopts the short-sighted forecasting traffic flow model based on state space neural network (StateSpaceNeuralNetwork, SSNN) and EKF in this patent.With tradition, neutral net is different, state space neural network by the state layer of neuron state before adding a storage as impermanent memory layer, so that neutral net can predict output valve according to the Determines of the state of current time and previous moment, the spatio-temporal state of study complexity that can be more efficient.By the mathematical description of state space neural network it can be seen that the vectorial s (t) of hidden layer is input vector and deviation weighted sum, it can pass through to transmit functional expression and be calculated by input layer vector x (t):

In formula (12): sm represents the value of m-th hidden layer neuron,Represent the weight connecting i-th input layer and m-th hidden layer neuron,Represent and connect e hidden layer neuron and the neuronic weight of m-th state layer,Representing the deviation value weight with m-th hidden layer neuron, bm represents the deviation value of m-th hidden layer neuron, and its value is fixed as 1, h () and represents transmission function.

Step 400: the critical path of hypersaturated state intersection group is carried out oversaturated intersection signal timing dial design approach and induction, adjusts hypersaturated state intersection group traffic signal control strategy；

In step 400, the optimization of intersection group facility, control structure, traffic control strategy and model determine signal timing plan optimum ideals under hypersaturated state and control effect.Owing to currently not yet forming comparatively ripe hypersaturated state traffic control objective, when the target therefore controlled when convention traffic can make circulation of traffic run, comparatively ripe signal optimisation strategy should be adopted as far as possible rather than select new control strategy.

Control structure refers to as realizing the system structure that control strategy is taked, mainly include centralized, distributing, distributed three kinds.Traffic control system is owing to having the feature of typical information dispersion (subsystem is distributed in wide city space scope), along with the expansion of road network scale is difficult to centerized fusion, according to the differentiation to road net traffic state, realize controlling classification and the combination of parameter and control structure, be the core way solving control problem.

Characteristic according to the space of hypersaturated state intersection group, traffic flow and traffic control, when optimizing hypersaturated state intersection group traffic control scheme, the traffic control of intersection group should be divided into three layers: intersection group traffic administration layer, critical path coordinate key-course, single-point intersection optimization layer.Intersection group controls management level and in intersection group aspect, overall transport need is managed, it is ensured that under hypersaturated state, shared to periphery road network by traffic pressure；Critical path is coordinated key-course and is mainly optimized the coordination traffic signal control scheme of critical path, utilizes the storage capacity of intersection group road network, ensures the smooth and easy operation of traffic flow of intersection group critical path, rapid evacuation traffic congestion；Single-point intersection optimization layer is the signal time distributing conception optimizing each crossing according to enforcement dynamic traffic situation, it is ensured that critical path is minimum by vehicle queue length maximum, average, it is to avoid negative effect produces.

The three layers traffic control structure of corresponding intersection group, the traffic control strategy correspondence of hypersaturated state is divided into single spot optimization layer, critical path optimization layer, network optimization layer.Single spot optimization layer is focused mainly on the calculating of single crossing timing scheme, feed back initial signal time distributing conception (split, Cycle Length etc.) at critical path optimization layer and optimize initial timing scheme afterwards, and final signal time distributing conception is sent the control unit to crossing, each control unit needs energy phase interchangeable information, carry out short-time traffic flow forecast, complete the rolling optimization of control program.Critical path optimization layer is to detect data and critical path according to real-time dynamic traffic, takes into account traffic control optimisation strategy and optimization aim constraints, forms critical path Coordinated Control Scheme.This scheme reflects traffic controller and alleviates the decision thought of bottleneck road within the scope of intersection group, is the basis of Internet signal timing plan optimization, is also the core alleviating intersection group hypersaturated state.

When urban road intersection group under hypersaturated state coordinates control, should in conjunction with supersaturation control strategy, operation characteristic for intersection group road network traffic flow, between simplification strategic control parameter on interactional basis, the common signal period is adopted to wake up with a start control in critical path, and the dispersion degree of fleet between crossing is limited in tuneable threshold value, make full use of the space storage capacity on Heavenly Stems and Earthly Branches road, make the control output scheme after global optimization can better adapt to the real-time change of transport need situation within the scope of intersection group.

Under hypersaturated state, the traffic circulation state evaluation standard of urban road intersection group is different with the traffic circulation method for evaluating state of stable state, and its optimization aim is also different.The traffic control strategy of hypersaturated state intersection group needs comprehensively to determine according to the optimization aim (if crossing is by number, queue length etc.) under intersection group real-time traffic running status, the design characteristics of intersection group, hypersaturated state.The data gathered by detecting device should be passed through to process and calculate the demand that could meet traffic control and management.DSS is part most crucial in whole traffic control loop, real-time traffic service data and short-term prediction information that this system draws according to traffic information processing system determine traffic control strategy in real time, thus the control target realizing when disturbance presetting (as crossing is the shortest etc. by maximum number, queue length), for communications policy personnel's reference.Communications policy personnel determine final traffic control strategy by traffic on the spot and intersection group traffic circulation characteristic.The effectiveness of intersection group traffic control system is the effectiveness by control strategy and the dependency with practical situation determines, therefore when determining traffic control strategy, should the optimization method of sophisticated systems and select Theory of Automatic Control algorithm as far as possible, but not simply apply some specific algorithm and solve problem.

In urban road intersection group, the situation of traffic circulation can describe by multiple evaluation index, analyzes in order to convenient, and it is that standard is evaluated that this patent chooses the consumption of overall travel time in road network.Assume in period t, i-th crossing enter intersection group region vehicle number be Di (t) (i=1,2 ...), so total inlet of intersection group is:In like manner, if the vehicle fleet flowing out intersection group in period t is:Therefore, in period t, intersection group scope road network vehicle number is: N (t)=N (t-1)+D (t)-S (t).If the internal initial flow of intersection group is N (0), then:If the elapsed time of i-th car is ti in road network, then the total time-consuming Ts in road network is:The output flow that the wastage in bulk or weight minimal time of urban road intersection group is equivalent under time weighting is maximum, and namely under suitable traffic control measure, vehicle can leave intersection group more soon, and overall the consumed time is more short.

See also Fig. 3, be the traffic signalization Optimizing Flow figure of the hypersaturated state of the embodiment of the present invention.Oversaturated intersection critical path and control strategy optimization mode are particularly as follows: in intersection group scope, hypersaturated state, critical path, under the clear and definite premise of short-term traffic flow Parameters variation information, first the optimization aim of hypersaturated state traffic signalization is optimized, the traffic control strategy of traffic control structure and different aspects, realize hypersaturated state intersection group traffic signalization, choose critical path maximum minimum for optimization aim with queuing by vehicle number, application intersection group layer, critical path layer, three level Optimizing Modes of single-point intersection layer discuss traffic control optimisation strategy respectively；To prevent the negative effects such as intersection group produces that overflow, green light sky are put for boundary condition, determine the optimization range of intersection group traffic control parameter, and traffic control parameter optimization method is proposed, so that the smooth and easy operation of hypersaturated state intersection group traffic flow, the state of stable state traffic control optimization method can be applied in fast quick-recovery road.On the basis of Static reference timing scheme optimization, according to real-time dynamic traffic stream and short-time traffic flow forecast information, dynamically update Traffic Signal Timing scheme.In order to know explanation step 400, see also Fig. 4, be the oversaturated intersection critical path schematic flow sheet with control strategy optimization method of the embodiment of the present invention.The oversaturated intersection critical path of the embodiment of the present invention comprises the following steps with control strategy optimization and method of adjustment:

Step 401: intersection signal timing controls prioritization scheme static optimization；Under hypersaturated state, stable state traffic control so that traffic flow to run optimization aim smoothly no longer applicable.The current optimization aim such as car number is maximum, queue length is minimum in analysis of key path are in the suitability of hypersaturated state traffic control, and determine traffic control optimization aim, and the optimization for traffic control parameter lays the foundation.Need in conjunction with hypersaturated state intersection group to optimize the control target dredging bottleneck road traffic flow, select the traffic control structure of hierarchical when traffic control, and be divided into intersection group layer, critical path layer, single-point intersection layer.Intersection group layer is mainly through the method such as current limliting, Self Adaptive Control, and by intersection group internal transportation stream rapid evacuation, suitably restriction external traffic flows into simultaneously；Critical path layer pays close attention to the coordination signal time distributing conception in the most prominent path of intersection group traffic problems；Single-point intersection layer then optimizes timing parameter by the semaphore of crossing according to the Coordinated Control Scheme of real-time traffic parameter and critical path layer, finally determines that intersection signal timing controls prioritization scheme.

Step 402: dynamic cooperation traffic signalization intersection group；

Step 403: the traffic control strategy of hierarchical screening hypersaturated state intersection group；Three layers hierarchical optimal Controlling model according to intersection group, in existing control strategy, screening is applicable to the traffic control strategy of hypersaturated state.Wherein the traffic control strategy of single-point intersection layer have green light time delay, in advance termination phase, phase place to service again, dynamically turn left, left turn phase in advance/move after, short line crossing adopt identical timing scheme etc.；Critical path layer includes the phase contrast design etc. reversely coordinated control, synchronous transport control, blue flash and prevent overflow, green light sky from putting；The main current limiting of control strategy of intersection group layer, Self Adaptive Control etc..

Step 404: coordinate timing scheme based on non-dominated sorted genetic algorithm optimization, controls the benchmark timing scheme of dynamic optimization as signal；Based on the off-line data that intersection group runs, traffic control objective according to hypersaturated state, choose that the current vehicle number of weighting that critical path passes through is maximum and critical path is on average queued up minimum for optimization aim, with the green time of each crossing for input variable, timing scheme is coordinated in application second filial generation multiple target non-dominated sorted genetic algorithm optimization, controls the benchmark timing scheme of dynamic optimization as signal.

Step 405: traffic parameter real time dynamic optimization algorithm；

Specifically see also Fig. 5, be the intersection group traffic control dynamic optimization method frame diagram of the embodiment of the present invention.Based on traffic state information, Forecasting Short-term Traffic, key control parameter span, on the basis of benchmark control program, dynamically adjust the value of traffic control parameter consumption analysis when each step is carried out according to real time traffic data.For reaching to prevent hypersaturated state intersection group from producing the target of negative effect by traffic control, adjustment period length can be passed through, it is to avoid before the joint of discrete shock wave and queue clearance shock wave is positioned at crossing, upstream, thus reaching to avoid being detained the purpose queued up；By adjusting the phase contrast of two crossings, overflow and green light sky can be avoided too to put the generation of phenomenon.Adopting said method obtains the span of each traffic parameter, it is possible to as the span of traffic parameter dynamic optimization.

Step 500: the intersection group traffic signal control strategy after combustion adjustment, it is achieved crossing control signal timing designing scheme steady-state operation and intelligent robot interlinked command；

In step 500, existing intelligent robot has been able to competent accurate, repetitive work, but many times it can't carry out oneself's adjustment according to new task neatly, a unfamiliar or uncertain sight can not be dealt with, such as urban highway traffic intelligent robot interlinked command crossing operation etc., the present invention is by carrying out perception, cognition and Behavior-Based control to intelligent robot, it is achieved integrative design intersection optimization and intelligent robot interlinked command.Intelligent robot is realized to the perception of crossing, cognition by step 100 to step 400, enter stable state intersection signal timing and control prioritization scheme, intersection signal timing is controlled prioritization scheme and runs three cycles, intersection signal timing is controlled prioritization scheme properly functioning with intelligent robot interlinked command crossing, it is achieved integrative design intersection optimization and intelligent robot interlinked command simultaneously.The integrative design intersection prioritization scheme of the embodiment of the present invention and the method for intelligent robot interlinked command comprise the following steps:

(1) the quiet collaborative traffic control flow process of urban road oversaturated intersection group motion

According to the logical Controlling model structure of urban road intersection flock-mate, hypersaturated state intersection group is carried out traffic control, the hypersaturated state of intersection group in conjunction with intersection group state recognition algorithm, should be identified.When determining that intersection group is in hypersaturated state, and adjust traditional traffic signal control method when can not eliminate current congestion status, should first determine the reason that intersection group hypersaturated state is formed, if it is create overflow because of traffic design due to indivedual crossings or green light sky such as puts at the negative effect that intersection group produces hypersaturated state, corresponding traffic administration control measure should be adopted, to get rid of traffic congestion as early as possible；If the volume of traffic is excessive, then should carry out retaining or the method for current limliting at crossing bounds, evacuate the delay queuing vehicle within intersection group as early as possible, result in combination with short-term traffic flow prediction, based on static optimization scheme, traffic signal are carried out dynamic optimization by the bottleneck road for intersection group--critical path, to discongest the wagon flow in critical path as quickly as possible.When optimizing the traffic timing scheme of each crossing, it is necessary to make full use of the wagon flow storage capacity of road network, ensure that smooth flow runs, make to block up and dissipate as early as possible.If the formation of the hypersaturated state of intersection group normalization, then need in the entire scope of city, transport need to be analyzed, by improving supply and the traffic management measure of means of transportation, and in conjunction with modes such as traffic guidances, reduce the traffic flow of bottleneck road.

(2) selection in control cycle is coordinated

Intersection group critical path coordinates the mission critical that the selection in control cycle is hypersaturated state signal coordinated control, and choosing non-optimal signal period length will increase crossing queuing overflow and stop the probability occurred.Under stable traffic flow state, Cycle Length can be determined by parameters such as exterior traffic amount and road section capacities；And under hypersaturated state, the major influence factors coordinating to control Cycle Length is section storage capacity and the arrival rate of red time and green time vehicle.

The main target that hypersaturated state traffic coordinated control Cycle Length is chosen is in that queuing overflow phenomena occurs in the crucial crossing avoiding intersection group, application closure in the upperreaches strategy, avoids the generation of crossing overflow phenomena by coordinating the Cycle Length of crossing, upstream.The suggestion Cycle Length applying this strategy generating is guarantee the shock wave formed maximum cycle length of dissipation before arriving crossing, upstream of queuing up.

Specifically, see also Fig. 6, be the Cycle Length computational methods schematic diagram preventing overflow of the embodiment of the present invention.It is as follows that the present invention draws out the computing formula calculating the peak signal control cycle preventing queuing overflow by space-time diagram:

In formula (13): L-road section length；Crossing, W-upstream width；The effective green time of Ga-downstream intersection；H-from when sailing vehicle headstock from；L-lost time；Lu-average traffic effective wagon degree；RL-shock wave dissipation place；C1-prevents the Cycle Length of overflow；Safety coefficient when SF-vehicle empties；U-is from the velocity of wave sailing shock wave；The speed of next wagon flow first car of v-；The velocity of wave of ω-parking shock wave；Δ-coordination controls phase contrast.

Coordinate traffic control Cycle Length under hypersaturated state it is also contemplated that dissociate under critical path sail rate and road section length [5], therefore, calculate Cycle Length and should be:

Signal period length on the scope that critical path coordinates the control cycle, should be scanned in conjunction with actual traffic arrival rate by the Cycle Length of each crossing of intersection group according to the traffic control optimisation strategy of single-point intersection layer and signal control constraints.When section or short line intersection traffic flow are bigger, it should avoid using the short cycle；For avoiding short line crossing to produce queuing spillover, the method adjusting phase contrast can be adopted to reduce the arrival rate of red time when the short cycle can not be used.Same extend downstream intersection green time so that the effect that produces in crossing, upstream to dam also can be avoided producing queuing flooding problems.Short line crossing is as described below to the restriction of Cycle Length when the volume of traffic is higher.

1. the minimum Capacity Constraints at each crossing:

In formula (15):The time span of-the i-th each crossing phase place j；The total losses time of each crossing of Li-i-th.

2. the maximum Capacity Constraints of each crossing:

3. the maximum saturation constraint of each crossing:

In formula (17):The phase place maximum saturation of each crossing of-；The flow-rate ratio sum of Yi-i-th crossing, it is calculated as follows shown in formula:

In formula (18): the phase contrast in mono-cycle of j-；yj,y’_jThe flow-rate ratio of-jth phase place and design discharge ratio；Qd-design traffic volume, unit pcu/h；Sd-designs saturation volume, unit pcu/h.

Intersection group is coordinated traffic control reference cycle length and is taken above-mentioned conditionary periodic minima:

C_ref=min (C₁,C₂,C₃,C₄,C₅)(19)

(3) offset optimization computational methods

The optimization problem that it is parameters optimization with phase contrast that offset optimization can be counted as, the value that its target is certain complicated function is maximum or minimum, during oversaturated intersection faciation potential difference on-line optimization, should optimize the phase contrast of critical path.When optimizing phase contrast, section each in intersection group is divided into some paths, and according to the significance level of critical path, it is optimized.In the path comprising n crossing, it is understood that there may be phase contrast number be (C/r) n-1, C be Cycle Length (s), r is step-size in search (s).Therefore, the computational complexity solving phase contrast is the exponential depth growth of n, need to adopt efficient optimization method [6], adopt line-axle associated methods (Link-PivotingCombinationMethod, LPCM) to carry out the phase contrast of Optimizing City intersection group's critical path.

Line-axle combined techniques is equivalent to a section by the step of a series of search, combination road network, combine every time and be equivalent to the section that extra to be converted into the section identical with section before, with the link flow that section before directly utilizing optimizes, it is relatively specific for the main line type intersection group of inner city.It optimizes the phase contrast of traffic signalization network by the form that " series connection " and " parallel connection " combines.

The span assuming j is from jo to jmax:

Step one: the start position at optimized arterial road defines crossing Jo in fact；

Step 2: combine each crossing on Trunk Road Network according to procedure below successively；

1. make { Δ }=Δ jo, Δ jo+1 ..., Δ j-1} (sets jth crossing as crucial crossing, offset optimization is the highest with jth crossing optimization level)；

2. { Δ }={ Δ } U{ Δ j} (wherein Δ j is the phase contrast previously merged) is made；

3. assume that each cycle can be divided into B period, each Period Length to be ω, if δ=1,2 ..., (B-1), increased by the phase contrast that each crossing is current and the phase contrast combined before, set up network phase difference evaluation model:

4. select suitable δ-value to obtain best evaluation effect so that { Δ } ← { Δ } δ.

Step 3: for isolated blob, { Δ j} is to particular value to specify crossing phase contrast to reach requirement to may specify the adjustment set of phase contrast.

Queue up in the crossing that the heavy turning traffic flow of the restriction and other flow direction remittance critical paths that optimize the phase contrast especially needed consideration downstream intersection traffic capacity of hypersaturated state intersection group is formed.Optimizing of hypersaturated state intersection group phase contrast needs to consider two constraints on the basis of original scheme: namely designed phase difference prevents crossing generation overflow phenomena and green light sky from putting phenomenon.

(4) signal controls real-time adaptive renewal

Optimizing and revising of split is active, the most most frequent parameter during the big parameter of traffic signal control system four (cycle, phase place phase sequence, split, phase contrast) adjusts.Single-point intersection Split Optimization real-time adaptive control key content is as follows:

1. the defining of split

After traffic control signal cycle duration is determined, the ratio of the effective green time of one of them signal phase and cycle duration is defined as the filtering ratio of signal phase, namelyWherein λ is split, and C is signal period duration, and ge is effective green time, and ge=g (green time)+A (yellow time)-L (starts lost time)；After C determines in the signal period, the optimization of split λ is optimized effective green time ge exactly, and determines that ge after determining display green time g simultaneously, optimize ge herein and just determine that optimization g.

2. Split Optimization arrange purpose and premise

After the signal period duration of traffic control system has optimized and determined, for the actual change of dynamically corresponding traffic flow, it is necessary to the green time of each phase place is carried out reallocation and adjusts by each cycle, so that the desired value that whole intersection traffic stream runs reaches optimization.Ensure that the optimum results of signal period and phase contrast is carried out simultaneously.Set up and assume:

1) signal period is reasonably determined；

2) phase place phase sequence is reasonably selected to optimize；

3) upstream and downstream of each entrance driveway line in crossing has all buried wagon detector underground；

4) mixed traffic flow impact reasonable consideration in minimax green time and copper sulfate basic etc. retrain on Split Optimization.

3. the determination of split initial value

When whistle control system brings into operation, the green time of phase place can be determined by offline optimization, or is invoked at the scheme of close temporal proximity before this, along with system operation can constantly on-line optimization adjustment, progressively met the running status of actual traffic stream by optimized algorithm.The ratio of each phase place Optimal green signal ratio in unlike signal cycle is substantially directly proportional to the ratio of phase place saturation volume ratio substantially:

In formula (21): gi, gj represent the Optimal green signal ratio of phase place i, j；Yi, yj represent the saturation volume ratio of phase place i, j；Qi, qj represent the flow of phase place i, j, and si, sj represent the saturation volume of phase place i, j.Therefore, when the signal period optimized determine, it is possible to according to etc. the principle of saturated distribution, carry out the determination of the split initial value under single-point real-time adaptive control according to the ratio of the saturation volume ratio of each phase place.

4. the constraints of Split Optimization

The constraints of Split Optimization is mainly signal period constraint, the constraint of minimax green time, Capacity Constraints:

In formula (22), i represents number of phases；Qi represents the flow of phase place i, C representation signal cycle；Gi represents the split of phase place i；S represents the saturation volume of phase place i；Xp represents the saturated acceptable maximum critical saturation of each phase place, generally takes Xp=0.95；Gmin represents the minimum green time of phase place, and gmax represents the maximum green light persistent period of phase place, and gmin and gmax can determine according to the traffic off-line in city, to advantageously ensure that traffic safety and to improve efficiency.

5. Split Optimization principle and algorithm

There is maximum difference with signal period optimization and be in that in Split Optimization: split is multi-C vector, and its dimension is equal to number of phases.Therefore, the complexity simplifying hyperspace optimization when ensureing to optimize precision and the memory cost taken is must take into consideration how when Split Optimization.Under the signal period determines situation, the distribution of split is generally of following methods:

The saturated timing method such as a.: based on fair principle, be used for the foundation of Split Optimization by saturation volume, has a feature simple, quick, near-optimization, but be open to traffic efficiency and service level not as total delay minimizes timing.

B. total delay minimizes timing method: the principle based on efficiency carries out split distribution, and the efficiency that is open to traffic and service level are best, but calculates time length, model needs complexity.

C. average traffic delay equal timing method: the average traffic delay making each phase place wagon flow is equal.

D. queuing rate equal timing method: the queuing rate making each phase place wagon flow is equal.

Based on this, select based on etc. the optimization that minimizes of the total delay of saturated distribution, using etc. the split of saturated distribution as the initial split of system optimizing, then the split that Step wise approximation is best.

6. Split Optimization flow process

According to above-mentioned analysis, the computing flow process of Split Optimization can be divided into three phases:

A. the initial dispensing phase of split

Upstream detector is utilized to detect the cycle traffic spirogram formula of generation in real time, principle according to equisaturation, carrying out original allocation according to the comparison signal period duration of the saturation volume ratio of each phase place, the split sum of each phase place obeys signal period constraint and minimax long green light time, the constraint of maximum critical saturation:

In formula (23), m represents the number of phases of crossing；

In formula (24), qi represents the volume of traffic of the i-th phase place, and Si represents the saturation volume of the i-th phase place.

B. the double optimization of split

If increasing the delay of phase place green time minimizing and the total revenue of stop frequency, more than the total losses suffered by the vehicle incured loss through delay by red light, green light timing just should be increased；Otherwise then should reduce green time.Based on this, optimizing from the prolongation phase place split on the main road of crossing of split, use climbing method, before green light is opened, the split performed with a upper cycle compares, search+Δ gs, 0, the change of intersection delay size in-Δ gs situation, find and incur loss through delay minimum split trimming scheme, now optimize all of non-prolongation phase place in tentative calculation crossing according to equisaturation principle, according to the ratio distribution split arriving saturation volume ratio, it follows that optimize and show that crossing extends the split of phase place and the split of other all non-prolongation phase places.If there is arbitrary non-prolongation phase place to be unsatisfactory for the constraint of minimax long green light time, the constraint of maximum critical saturation, then the saturated distribution such as newly carry out after meeting constraints above again:

In formula (25),WhereinFor the split of this cycle stretch-out phase place,For the split of a upper cycle stretch-out phase place, its optimization object function is:

C. the execution adjusting and optimizing of split

Owing to system upstream and downstream is provided with detector, therefore can controlling to save the situation of green time according to sensing, carry out saving redistributing of green time, to obtain better benefit, the delay value of accelerating system reduces further.Set up three class phase places: extend phase place, induced phase, master phase；It is primarily intended to is easy to when sensing controls reasonably to adjust the surplus and deficiency of each phase place green time, green time priority allocation unnecessary for non-prolongation phase place to the big prolongation phase place of the volume of traffic.

7. the principle that phase contrast is arranged is extended

Extend phase place and be generally arranged at that the volume of traffic is big or saturation volume is than big main road, its final green time, just can only can determine that after the split of other phase places is determined, it deducts the remaining time after other all phase places equal to cycle duration, and the total number extending phase place should be generally less than the sum arranging induced phase.

Introduce after extending phase place, it is necessary to extending phase place immediately following, after being arranged on induced phase, when induced phase is skipped or when having unnecessary green light to save, extending phase place and can obtain the whole green times having more of induced phase.Otherwise be arranged on before induced phase finishes then inadvisable extend phase place, because when induced phase not yet reaches maximum green light, then save green time and cannot adjust to extending phase place to ensure that the cycle duration optimized is carried out.

One main road direction is generally up to about arranging one and extends phase place, it is not necessary to each coordination direction all needs to be provided with prolongation phase place, particularly in two phase place situation.The phase place that master phase is intended merely to direction when regulation performs adjustment and introduces, it is not necessary to each crossing must be provided with master phase, particularly in two phase place situation.Redirected in a upper cycle if sensing controls phase place, then minimum green time when general phase place being arranged when the Split Optimization in next cycle is assigned to the induced phase split initialization of initial optimization.

8. based on the Split Optimization of double; two prolongation phase places

Single Split Optimization extended under phase condition is mainly described by foregoing, but can there is the not unique situation of prolongation phase place, and the large-scale crossing such as intersected at two major trunk roads exists typical four phase place situations.Now there are two and extend phase place, it is possible to adopt twocouese climbing method to be optimized search, obtain the Optimal green signal ratio under total delay minimum.Now to all non-prolongation phase places according to etc. saturated carry out green light distribution, to all prolongation phase places also according to etc. saturated carry out green light distribution, but be not equal to all phasetophases calculated in initial split situation and complete etc. saturated, but similar phasetophase relative etc. saturated.For typical four phase condition, the split making g1, g3 be non-prolongation phase place 1 and 3, g2, g4 are the split extending phase place 2 and 4, the Optimizing Search of split is adopted twocouese climbing method, then has:

The Split Optimization of double; two prolongation phase places adopts twocouese climbing method, and its optimization object function is:

In formula (28), d (g1), d (g3), d (g4) represent along extending each non-prolongation phase place delay value that phase place g2 direction uses climbing method to obtain；D (g11), d (g33), d (g44) represent along extending each non-prolongation phase place delay value that phase place g4 direction uses climbing method to obtain；Δ g2 represents the step-size in search extending phase place 2；Δ g4 represents the step-size in search extending phase place 4；Represent and extend the split that in phase place 2, a signal performs；Represent and extend the split that in phase place 4, a signal performs.

9. the interval of Split Optimization

Minimize to finally realize the delay that the signal period determines in situation, it is necessary to real-time matching is in the traffic of each import line being continually changing.When the adjustment interval of split is oversize, then real-time is poor, and it is excessively delayed that system tackles each phase place transport need change.When split adjustment interval is too short, then frequently adjust the instability that system will be brought to run.Owing to being spaced apart for two cycles as the optimization of signal period of strategy principal parameter, and split is as pure tactics parameter, its adjust interval should lower than the signal period, therefore split be optimized for each cycle once.The optimization time of split optimizes the split in next cycle before being typically in this cycle signal ended.Consider the time required for system optimization computing and communication transfer, therefore before first group of phase place green light of next cycle is opened the split of the necessary various phase place of optimization, its pre-set time, T was made up of following two parts: first be system optimization computing needed for time T1: depend on the performance of algorithm, calculating scale, hardware configuration situation；Second is that system schema performs required time T2: determined by communication transfer time and semaphore decoding time.

(5) set up new intersection signal timing and control collaborative linkage and commander's operational mode

The hypersaturated state intersection group signal set up controls to optimize new departure, in crossing, actual control signal machine environment runs three cycles, run with intelligent robot interlinked command with integrative design intersection optimization, 48 actions realizing artificial traffic signalization commander are harmonious with integrative design intersection, giving intelligent robot to optimize the traffic signalization function of hypersaturated state intersection group, the new function of manual coordination intersection traffic running command exercised by intelligent robot.

Refer to Fig. 7, be the structural representation of the traffic control system based on intersection group of the embodiment of the present invention.The traffic control system based on intersection group of the embodiment of the present invention includes intelligent robot, and intelligent robot includes first video camera module the 1, second video camera module 2, data processor module (not shown) and display module 3；First video camera module 1 and the second video camera module 2 are connected with data processor module and display module 3 respectively；nullFirst video camera module 1 and the second video camera module 2 gather 360 °, crossing panoramic video for Real-time and Dynamic，And by the video data transmission of shooting to data processor module and display module 3，Data processor module is for setting up crossing moving model according to video data，Crossing moving model is carried out operation situation monitoring，And analyze intersection group traffic characteristics according to crossing moving model，Crossing assessment index and in-circuit emulation analysis is carried out according to intersection group traffic characteristics，Identify intersection group traffic circulation state，Thus the critical path of hypersaturated state intersection group is carried out oversaturated intersection signal timing dial design approach，Adjust hypersaturated state intersection group traffic signal control strategy，And control the intersection group traffic signal control strategy after intelligent robot combustion adjustment，Realize crossing control signal timing designing scheme steady-state operation and intelligent robot interlinked command，Solving urban road oversaturated intersection single-point operation optimization problem；Display module 3 is for showing the first video camera module 1 and the video data of the second video camera module 2 shooting.

In embodiments of the present invention, first video camera module 1 is the 360 ° of panoramic high-definition video cameras that highly can stretch, it is located at the above-head of intelligent robot, second video camera module 2 is HD video video camera, it is located at the eye of intelligent robot, display module 3, for touching display screen, is positioned at the body part of intelligent robot, it is simple to artificial touch operation.

See also Fig. 8, be the structural representation of the data processor module of the embodiment of the present invention.The data processor module of the embodiment of the present invention includes model and sets up unit, traffic characteristics analytic unit, traffic circulation state recognition unit, policy optimization unit and scheme running unit；Specifically:

Model sets up unit for receiving the first video camera module and the video data of the second video camera module transfer, and carry out video data sorting out after screening, image recognition and feature extraction etc. process and generate crossing real-time dynamic information environment, set up that picture is clear, the crossing moving model of broad view；

Traffic characteristics analytic unit for carrying out operation situation monitoring to crossing moving model, and analyzes intersection group traffic characteristics according to crossing moving model；Wherein, the method that crossing moving model carries out operation situation monitoring includes: analysis intersection group blocks up and forms and evacuate collection and the process of mechanism and traffic circulation parameter；

Analyze intersection group to block up and form and evacuate mechanism and specifically include: analyze the risk factor blocked up of intersection group, determine that crossing overflow, green light sky are put, are detained the harmful effects such as the queuing impact for intersection group traffic congestion, it is determined that the process that hypersaturated state is formed；Judging traffic flow running rate when traffic flow bottleneck dissipates, application transportation network load balancing theoretical description congestion status evacuates the traffic flow character of process, for analyzing the traffic behavior based theoretical of hypersaturated state intersection group.

The collection of traffic circulation parameter and process specifically include: determine the traffic circulation parameter analyzed required for the logical running status of urban road intersection flock-mate, the pluses and minuses of the various traffic circulation parameter collecting method of comparative analysis and the adaptability to hypersaturated state traffic signalization, it is preferable that the Data Source needed for intersection group traffic status identification and traffic control；Set up traffic circulation parameter cleaning treatment method, it is determined that the algorithm of Data-parallel language, the differentiation of traffic flow wrong data, correction and traffic flow Reduction for Redundant Data is lost in traffic flow, lays the foundation for traffic state analysis.The traffic circulation parameter acquisition of the embodiment of the present invention and processing method specifically include video encoder server and the modeling of traffic relatedness index；Wherein,

The concrete mode of video encoder server is:

1) moving target candidate region is extracted, it is determined that vehicle region that may be present；

2) goal verification, the candidate region that the upper stage was produced confirms, it is judged that be vehicle or background；

3) Target Segmentation, by identifying the pixel meeting vehicle characteristics in image, separates target to be identified from background；

4) target following, according to the vehicle in frame before and after the characteristic matching extracted, thus calculating traffic circulation parameter；

5) target classification, classifies to different types of vehicle according to geometric shape, textural characteristics etc.；

6) later stage processes, and calculates traffic circulation parameter according to detection demand, such as vehicle flowrate, speed etc..

Traffic relatedness index includes discreteness coupling index and retardancy coupling index；

Discreteness coupling index is: by the impact of the discrete factor of wagon flow, downstream intersection all passes through crossing to the first bus and last bus ensureing fleet in same green time, then need to design the green wave band that broadens of a kind of diffusion type.But this design can make the green time of the crossing of most downstream look cannot be accepted, and is a kind of control mode that discreteness is thrown the reins to, often inadvisable in practical engineering application.The wide green ripple of the many employings of control method to discrete constraint, but the method can make the Some vehicles being positioned at wagon flow stem or afterbody can have certain delay at each crossing.Set discreteness relatedness index I1 rise as path in the signal control cycle, the settled isometric green time of the some ratio by vehicle, it may be assumed that

In formula (1): q0 (i) represents the wagon flow of a certain paths initial upstream intersection parking i period of line by number；The wagon flow of qd (i+T) delegated path terminal chiasma mouth the i-th+T period arrives number；T represents the running time from path starting point to the end；Tg represents the green wave duration in the signal period.Q0 (i) and qd (i+T) can adopt field observation value, it is also possible to is calculated by Robertson motorcade dispersion formula, it may be assumed that

In formula (2): the wagon flow of qd (j) delegated path terminal chiasma mouth jth period arrives number, t=β T=β (j-i), coefficient of dispersionα, β represent undetermined parameter, and Robertson advises value respectively 0.35 and 0.8.

Retardancy coupling index is: intersection group is formed to any section m on certain road, if there is N number of different flow direction in the crossing inlet road along this path direction of advance, calculates the functional areas length value of each flow direction Queue lengthField observation statistical value can be adopted, it is possible to use queue length computing formula is estimated, the present invention adopts the queue length computational methods of Synchro7, deceleration distanceWith perception-reaction distanceComputational methods, willIt is defined as section m along the crossing inlet road of path direction of advance, flows to the ratio of functional areas length maximum and path L, it may be assumed that

If this path is made up of M section, then its retardancy index I2 is:

The mode analyzing intersection group traffic characteristics is: understand the traffic characteristics of intersection group respectively from aspects such as intersection group geometry topological property, path space characteristic, traffic stream characteristics, traffic signalization characteristics, find the variation characteristic of traffic flow in intersection group, provide foundation for application supersaturation traffic control strategy.Wherein, intersection group is classified by intersection group geometry topological property according to the road path number feature between in intersection group two crossings；Traffic flow can be run the impact produced by the design of path space specificity analysis road traffic facility；Traffic stream characteristics gives suitable in the descriptive model of cutout between hypersaturated state urban road, according to intersection group traffic stream characteristics, chooses suitable traffic circulation data acquisition means, sets up data cleansing and processing method.Traffic signalization specificity analysis basic control principle and control structure, lay a good foundation for setting up traffic control method.

Traffic circulation state recognition unit, for carrying out crossing assessment index and in-circuit emulation analysis according to traffic characteristics, identifies intersection group traffic circulation state；Wherein, the method for traffic circulation state recognition unit identification intersection group traffic circulation state includes: intersection group scope defines, intersection group hypersaturated state identification and assessment, the detection of intersection group critical path and classification and traffic parameter short-term prediction modeling and simulation；

The principle that intersection group scope defines is as follows:

1) crossing having relatively High relevancy should be divided into an intersection group, and the crossing that relatedness is not strong should be divided in different intersection groups；

2) in urban road network, the crossing number in each intersection group should be roughly equal, and meets the hsrdware requirements of traffic controlling machine；

3) time complexity of algorithm is low, and committed memory to lack；

4) the result reply traffic flow that scope defines runs the impact having front.

In embodiments of the present invention, the method defining intersection group scope specifically includes: judging on the basis of intersection group spatial character and internal association mechanism, analyze the traffic relatedness between crossing in intersection group, set up the intersection group scope confining method of feature based matrix and based on the intersection group scope confining method of self organizing neural network.Application vehicle queue length describes intersection group linked character with the ratio of line crossing space length and the Quality degree of effective utilization of green time respectively, the former is in conjunction with flow factor and distance factor, the latter takes into account flow factor and timing factor, the various characteristic analysis method of integrated application, defines the scope of intersection group.

The mode of intersection group hypersaturated state identification and assessment is particularly as follows: based on the method analyzing intersection group degree of super saturation, propose the ratio applying invalid green time and the total green time caused by negative effect to define degree of super saturation index, and weigh intersection group degree of super saturation with this.Based on hypersaturated state intersection group characteristic of produced negative effect on Spatial Dimension and time dimension, calculate the degree of super saturation index of intersection group respectively in room and time dimension.By shock wave model and space-time diagram on Spatial Dimension, what produce when the shock wave produced when dissipating by queuing up and green ripple start calculates crossing maximum queue length from sailing shock wave, the parking shock wave produced when the shock wave produced when dissipating by queuing up and lower cycle red light start calculates the tributary queue length of crossing, calculates the degree of super saturation coefficient of Spatial Dimension with this.At time dimension, the upstream detector mainly through being produced by crossing queuing overflow occupies phenomenon for a long time to calculate the degree of super saturation coefficient of crossing.The degree of super saturation coefficient of general space dimension and time dimension, identifies the hypersaturated state of intersection group.

Hypersaturated state directly can not be measured by traffic parameter or calculate identification, indirectly obtains only by negative effects such as overflows produced by hypersaturated state.For the hypersaturated state of quantitative judge intersection group, the definition of intersection group hypersaturated state is extended, calculate supersaturation coefficient by the negative effect caused by hypersaturated state, so that it is determined that the hypersaturated state of intersection group.Hypersaturated state refers to the situation when a means of transportation generation transport need by traffic signalization is more than its traffic capacity state (green time maximum by number), its negative effect that by the delay queuing in certain cycle, negative effect or the upstream means of transportation in next cycle can be produced within a cycle because overflow goes out defines, and the ratio (supersaturation coefficient) applying invalid green time and total green time weighs degree of super saturation.

In embodiments of the present invention, adopting the hypersaturated state of induction coil Vehicle Detection data assessment intersection group, induction coil typical case lays mode and includes stop line detector and senior detector (laying in stop line upstream) two kinds.Under hypersaturated state, intersection group is queued up longer, all can not accurately detect, regardless of stop line detector or senior detector, the traffic organization identifying hypersaturated state crossing, it is necessary to identify the hypersaturated state of intersection group with method for parameter estimation.The negative effect that under application hypersaturated state, traffic control produces in space-time unique replaces traditional method of estimation to evaluate the state of means of transportation.Viscous flow queue length when the identified negative effect of algorithm mainly has the signal period to terminate and the overflow phenomena of crossing, upstream, two kinds of negative effects all can cause the effective green time of signalized intersections to reduce.Queue length is detained in the method estimation crossing adopting shock wave (Shockwave), according to the detector high occupation rate (QueueOverDetector that queuing vehicle extended stationary periods face on the detector is caused, QOD) overflow phenomena in phenomenon identification intersection group, identifies the hypersaturated state of intersection group further.

The calculating of shock wave velocity of wave, if velocity of wave (u2, u3, u4) is also used to calculate the maximum queue length in the cycle, because traffic arrives flow rate variance relatively greatly, queuing shock wave (u1) is not suitable for estimation queue length.Selecting from making shock wave (u2) and deviating from shock wave (u3) and estimate queue length, computing formula is:

In formula (4): flow rate when qm and km represents maximum flow respectively and density, kj represents jamming density,WithRepresent traffic arrival rate and corresponding density.WithRefer to the traffic flow modes through detector after time Tc, when solving u2 it is assumed herein that qm, km and kj are fixed value, compression shock ripple u4 and have identical velocity of wave from sailing shock wave u2.

High-resolution traffic data is used to estimation and includesQm, km at interior various traffic variations, wherein traffic flow rate data, asWith qm can directly by detector acquisition, butKm isodensity data must be estimated.The independent holding time can be provided, it is assumed that effectively vehicle commander is it is known that space mean speed can be obtained based on the traffic data of event；Now available AFR estimates density data divided by space mean speed.Estimating individual speed ui, space mean speed us, the method for flow rate q and density k is:

In formula (5) to formula (8): t0, i and tg, i represents detector holding time and the interval of vehicle i, ui and hi represents speed and the space headway of vehicle i, q, us and k represent AFR, space mean speed and density respectively, Le represents effective vehicle commander, and n represents the vehicle number of a fleet in same traffic behavior.It is detained queue length and degree of super saturation Index for Calculation, the maximum queue length in the n-th cycleWith the moment reaching maximum queue lengthFor:

In formula (9) and formula (10): Ld represents stop line to the distance between detector.

The detection of intersection group critical path with the mode of classification particularly as follows: based on the strong feature of fleet's traffic relatedness in intersection group, adopt the intersection group critical path identification way based on wavelet transformation and spectrum analysis, analyze and extract intersection group traffic flow variation characteristic in short-term, utilize the critical path of the method detection intersection group of data mining analysis, for intersection group path ranking.In conjunction with the characteristic that intersection group critical path upstream and downstream wagon flow dispersion degree is little, traffic signal are pressed different frequency and are decomposed by application wavelet transformation technique, retain the high-frequency signal of reflection traffic flow variation characteristic in short-term and the low frequency signal of reflection traffic flow basis variation characteristic, filtered traffic signal are reconstructed the new traffic signal highlighting traffic flow variation characteristic in short-term, as the input data of critical path identification and classification.Each import of intersection group of calculating wavelet transformation reconstruct flows to the cross-spectral density between the power spectral density of traffic signal and the flow direction.The degree of association of two traffic signal is determined by calculating the consistency coefficient of each cross spectrum, obtain correspondence and specify the criticality index in all paths of import, again through the phase place calculated between two signals, the travel time checking being aided with at 2 calculates effectiveness, the comprehensive significance level analyzing all import critical paths.

In intersection group, the power of intersection traffic relatedness is mainly manifested between crossing wagon flow dispersion degree size, the i.e. similarity arriving stream characteristics and upstream stream characteristics of downstream intersection.The performance in critical path of this similarity becomes apparent from, once related grouped intersections middle and upper reaches crossing causes the traffic flow parameter such as flow, speed to change because of traffic signalization or traffic congestion, High relevancy according to relatedness Adjacent Intersections, the variation characteristic in short-term of traffic flow parameter can keep to downstream intersection.Under hypersaturated state, because wagon flow passes through crossing with saturation volume rate always, when in section, the dispersion degree of traffic flow running parameter is than stable state less, crossing is respectively flowed to the variation characteristic in short-term of traffic flow parameter as foundation, the critical path of Model Identification hypersaturated state intersection group can be set up.Model is it needs to be determined that suitable traffic parameter is to describe wagon flow feature, and chooses the variation characteristic in short-term of appropriate data digging method extraction wagon flow.

The feature little in order to highlight intersection group critical path upstream and downstream wagon flow dispersion degree, use small wave converting method that traffic signal are pressed different frequency to decompose, retain the high-frequency signal of reflection traffic flow variation characteristic in short-term and the low frequency signal of reflection traffic flow basis variation characteristic, filtered traffic signal are reconstructed the new traffic signal highlighting traffic variation characteristic in short-term, as the input data of critical path identification and classification.Wavelet transformation (WaveletTransformation) is the localization analysis of time (space) frequency, signal (function) is progressively carried out multi-scale refinement by flexible shift operations by it, it is finally reached high frequency treatment time subdivision, the frequency segmentation of low frequency place, can automatically adapt to the requirement that time frequency signal is analyzed, thus any details of signal can be focused on, solve the difficult problem of Fourier transformation.Wavelet transformation is that a kind of window size is fixed and its shape is variable, the time frequency resolution that time window and frequency window can change, and HFS has higher temporal resolution and relatively low frequency resolution.

Wavelet transformation is inherited and has developed the thought of Short Time Fourier Transform localization, can overcome again window size simultaneously and can not provide a T/F window with frequency shift with shortcomings such as frequency changes, time carry out the ideal tools of signal time frequency analysis and process.It is mainly characterized by by convert can the feature of abundant some aspect of outstanding problem, be obtained for successful application in a lot of fields.

Wavelet transformation is and is analysed to signal and is launched into the weighted sum of family's small echo machine, and its implication is morther wavelet (MotherWavelet) functionAfter making displacement τ, then make inner product with signal f (t) to be analyzed under different scale α:

In formula (11): α represents scale factor, α > 0；τ represents displacement, and its value can just can be born；Represent wavelet function and displacement thereof and yardstick stretches.

For the degree of association of quantitative Analysis intersection group each path upstream and downstream traffic flow, the method adopting spectrum analysis, using traffic flow change as input signal, analyze its spectral change feature at different frequencies.By calculating the cross-spectral density of each crossing inlet traffic signal, analyze the consistency coefficient of its signal, to determine the degree of association of two traffic signal, and apply the phase contrast of two signals, with the effectiveness of evaluation algorithm.

Frequency spectrum refers to the signal of time domain representation under frequency domain, it is possible to carries out Fourier transformation for signal and obtains, and the conclusion of gained is respectively with amplitude or phase place for the longitudinal axis, and frequency is transverse axis.Representing, with amplitude frequency spectrum, the situation that amplitude changes with frequency, phase frequency spectrum represents the situation that phase place changes with frequency.Frequency spectrum can represent that a signal is made up of the string ripple of which frequency, it is also possible to the information such as size and phase place of finding out each frequency string ripple.Spectrum analysis is a kind of technology that sophisticated signal is decomposed into relatively simple signal, finds out the way of signal information (such as amplitude, power, intensity, phase place etc.) at different frequencies and ascends the throne spectrum analysis.

Power spectrum is the sign of sequence power distribution properties on a different frequency digit time, if time series auto-covariance function γ_kSatisfy conditionThen there is following corresponding relation between power spectral density f (μ) and γ k:In formula: f (μ) defines on [-π, π], is real-valued nonnegative function.

The concrete mode of traffic parameter short-term prediction modeling and simulation is: the variation characteristic of the EXSMOOTH of application enhancements, state space neural network, EKF method, data fusion method prediction intersection group traffic parameter in short-term.By utilizing the traffic data of present period and historical period, the traffic data of subsequent period being predicted, model is not by the restriction of hypersaturated state.Traffic parameter short-term prediction has important effect in dynamic traffic control algorithm designs, it was predicted that precision have appreciable impact for the effectiveness of traffic control algorithm.The difference of the basic mode according to prediction, short-term traffic flow forecasting model is divided into data-driven and based on model two types.The method of the method mathematical statistics of data-driven or artificial intelligence processes, and such as historical traffic data such as traffic flow, traffic speed, hourages, and predicts the change of future time period traffic flow；Mainly apply traffic flow propagation model based on the method for model the traffic flow modes on Xue Ding path is estimated and predicts, according to the careful degree that traffic flow is described by model, model can be divided into macromodel, mesoscopic model, micromodel three kinds.The method form being applied to traffic parameter short-term prediction is various, and effect is different, adopts the short-sighted forecasting traffic flow model based on state space neural network (StateSpaceNeuralNetwork, SSNN) and EKF in this patent.With tradition, neutral net is different, state space neural network by the state layer of neuron state before adding a storage as impermanent memory layer, so that neutral net can predict output valve according to the Determines of the state of current time and previous moment, the spatio-temporal state of study complexity that can be more efficient.By the mathematical description of state space neural network it can be seen that the vectorial s (t) of hidden layer is input vector and deviation weighted sum, it can pass through to transmit functional expression and be calculated by input layer vector x (t):

In formula (12): sm represents the value of m-th hidden layer neuron,Represent the weight connecting i-th input layer and m-th hidden layer neuron,Represent and connect e hidden layer neuron and the neuronic weight of m-th state layer,Representing the deviation value weight with m-th hidden layer neuron, bm represents the deviation value of m-th hidden layer neuron, and its value is fixed as 1, h () and represents transmission function.

Policy optimization unit, for the critical path of hypersaturated state intersection group is carried out oversaturated intersection signal timing dial design approach and induction, adjusts hypersaturated state intersection group traffic signal control strategy；Wherein, the oversaturated intersection critical path of the embodiment of the present invention and control strategy optimization method include: intersection signal timing control prioritization scheme static optimization, dynamic cooperation traffic signalization intersection group, hierarchical screening hypersaturated state intersection group traffic control strategy, coordinate timing scheme, traffic parameter real time dynamic optimization algorithm based on non-dominated sorted genetic algorithm optimization；

Intersection signal timing controls prioritization scheme static optimization；Under hypersaturated state, stable state traffic control so that traffic flow to run optimization aim smoothly no longer applicable.The current optimization aim such as car number is maximum, queue length is minimum in analysis of key path are in the suitability of hypersaturated state traffic control, and determine traffic control optimization aim, and the optimization for traffic control parameter lays the foundation.Need in conjunction with hypersaturated state intersection group to optimize the control target dredging bottleneck road traffic flow, select the traffic control structure of hierarchical when traffic control, and be divided into intersection group layer, critical path layer, single-point intersection layer.Intersection group layer is mainly through the method such as current limliting, Self Adaptive Control, and by intersection group internal transportation stream rapid evacuation, suitably restriction external traffic flows into simultaneously；Critical path layer pays close attention to the coordination signal time distributing conception in the most prominent path of intersection group traffic problems；Single-point intersection layer then optimizes timing parameter by the semaphore of crossing according to the Coordinated Control Scheme of real-time traffic parameter and critical path layer, finally determines that intersection signal timing controls prioritization scheme.

Dynamic cooperation traffic signalization intersection group；

The traffic control strategy of hierarchical screening hypersaturated state intersection group；Three layers hierarchical optimal Controlling model according to intersection group, in existing control strategy, screening is applicable to the traffic control strategy of hypersaturated state.Wherein the traffic control strategy of single-point intersection layer have green light time delay, in advance termination phase, phase place to service again, dynamically turn left, left turn phase in advance/move after, short line crossing adopt identical timing scheme etc.；Critical path layer includes the phase contrast design etc. reversely coordinated control, synchronous transport control, blue flash and prevent overflow, green light sky from putting；The main current limiting of control strategy of intersection group layer, Self Adaptive Control etc..

Timing scheme is coordinated based on non-dominated sorted genetic algorithm optimization；The benchmark timing scheme of dynamic optimization is controlled as signal, based on the off-line data that intersection group runs, traffic control objective according to hypersaturated state, choose that the current vehicle number of weighting that critical path passes through is maximum and critical path is on average queued up minimum for optimization aim, with the green time of each crossing for input variable, timing scheme is coordinated in application second filial generation multiple target non-dominated sorted genetic algorithm optimization, controls the benchmark timing scheme of dynamic optimization as signal.

Traffic parameter real time dynamic optimization algorithm；Based on traffic state information, Forecasting Short-term Traffic, key control parameter span, on the basis of benchmark control program, dynamically adjust the value of traffic control parameter consumption analysis when each step is carried out according to real time traffic data.For reaching to prevent hypersaturated state intersection group from producing the target of negative effect by traffic control, adjustment period length can be passed through, it is to avoid before the joint of discrete shock wave and queue clearance shock wave is positioned at crossing, upstream, thus reaching to avoid being detained the purpose queued up；By adjusting the phase contrast of two crossings, overflow and green light sky can be avoided too to put the generation of phenomenon.Adopting said method obtains the span of each traffic parameter, it is possible to as the span of traffic parameter dynamic optimization.

Scheme running unit is for the intersection group traffic signal control strategy after combustion adjustment, it is achieved crossing control signal timing designing scheme steady-state operation and intelligent robot interlinked command；Specifically, the integrative design intersection prioritization scheme of the embodiment of the present invention and the method for intelligent robot interlinked command include:

(1) the quiet collaborative traffic control flow process of urban road oversaturated intersection group motion

According to the logical Controlling model structure of urban road intersection flock-mate, hypersaturated state intersection group is carried out traffic control, the hypersaturated state of intersection group in conjunction with intersection group state recognition algorithm, should be identified.When determining that intersection group is in hypersaturated state, and adjust traditional traffic signal control method when can not eliminate current congestion status, should first determine the reason that intersection group hypersaturated state is formed, if it is create overflow because of traffic design due to indivedual crossings or green light sky such as puts at the negative effect that intersection group produces hypersaturated state, corresponding traffic administration control measure should be adopted, to get rid of traffic congestion as early as possible；If the volume of traffic is excessive, then should carry out retaining or the method for current limliting at crossing bounds, evacuate the delay queuing vehicle within intersection group as early as possible, result in combination with short-term traffic flow prediction, based on static optimization scheme, traffic signal are carried out dynamic optimization by the bottleneck road for intersection group--critical path, to discongest the wagon flow in critical path as quickly as possible.When optimizing the traffic timing scheme of each crossing, it is necessary to make full use of the wagon flow storage capacity of road network, ensure that smooth flow runs, make to block up and dissipate as early as possible.If the formation of the hypersaturated state of intersection group normalization, then need in the entire scope of city, transport need to be analyzed, by improving supply and the traffic management measure of means of transportation, and in conjunction with modes such as traffic guidances, reduce the traffic flow of bottleneck road.

(2) selection in control cycle is coordinated

Intersection group critical path coordinates the mission critical that the selection in control cycle is hypersaturated state signal coordinated control, and choosing non-optimal signal period length will increase crossing queuing overflow and stop the probability occurred.Under stable traffic flow state, Cycle Length can be determined by parameters such as exterior traffic amount and road section capacities；And under hypersaturated state, the major influence factors coordinating to control Cycle Length is section storage capacity and the arrival rate of red time and green time vehicle.

The main target that hypersaturated state traffic coordinated control Cycle Length is chosen is in that queuing overflow phenomena occurs in the crucial crossing avoiding intersection group, application closure in the upperreaches strategy, avoids the generation of crossing overflow phenomena by coordinating the Cycle Length of crossing, upstream.The suggestion Cycle Length applying this strategy generating is guarantee the shock wave formed maximum cycle length of dissipation before arriving crossing, upstream of queuing up.

The computing formula calculating the peak signal control cycle preventing queuing overflow is drawn out as follows by space-time diagram:

In formula (13): L-road section length；Crossing, W-upstream width；The effective green time of Ga-downstream intersection；H-from when sailing vehicle headstock from；L-lost time；Lu-average traffic effective wagon degree；RL-shock wave dissipation place；C1-prevents the Cycle Length of overflow；Safety coefficient when SF-vehicle empties；U-is from the velocity of wave sailing shock wave；The speed of next wagon flow first car of v-；The velocity of wave of ω-parking shock wave；Δ-coordination controls phase contrast.

Coordinate traffic control Cycle Length under hypersaturated state it is also contemplated that dissociate under critical path sail rate and road section length [5], therefore, calculate Cycle Length and should be:

Signal period length on the scope that critical path coordinates the control cycle, should be scanned in conjunction with actual traffic arrival rate by the Cycle Length of each crossing of intersection group according to the traffic control optimisation strategy of single-point intersection layer and signal control constraints.When section or short line intersection traffic flow are bigger, it should avoid using the short cycle；For avoiding short line crossing to produce queuing spillover, the method adjusting phase contrast can be adopted to reduce the arrival rate of red time when the short cycle can not be used.Same extend downstream intersection green time so that the effect that produces in crossing, upstream to dam also can be avoided producing queuing flooding problems.Short line crossing is as described below to the restriction of Cycle Length when the volume of traffic is higher.

1. the minimum Capacity Constraints at each crossing:

In formula:The time span of-the i-th each crossing phase place j；The total losses time of each crossing of Li-i-th.

2. the maximum Capacity Constraints of each crossing:

3. the maximum saturation constraint of each crossing:

In formula (17):The phase place maximum saturation of each crossing of-；The flow-rate ratio sum of Yi-i-th crossing, it is calculated as follows shown in formula:

In formula (18): the phase contrast in mono-cycle of j-；yj,y’_jThe flow-rate ratio of-jth phase place and design discharge ratio；Qd-design traffic volume, unit pcu/h；Sd-designs saturation volume, unit pcu/h.

Intersection group is coordinated traffic control reference cycle length and is taken above-mentioned conditionary periodic minima:

C_ref=min (C₁,C₂,C₃,C₄,C₅)(19)

(3) offset optimization computational methods

The optimization problem that it is parameters optimization with phase contrast that offset optimization can be counted as, the value that its target is certain complicated function is maximum or minimum, during oversaturated intersection faciation potential difference on-line optimization, should optimize the phase contrast of critical path.When optimizing phase contrast, section each in intersection group is divided into some paths, and according to the significance level of critical path, it is optimized.In the path comprising n crossing, it is understood that there may be phase contrast number be (C/r) n-1, C be Cycle Length (s), r is step-size in search (s).Therefore, the computational complexity solving phase contrast is the exponential depth growth of n, need to adopt efficient optimization method [6], adopt line-axle associated methods (Link-PivotingCombinationMethod, LPCM) to carry out the phase contrast of Optimizing City intersection group's critical path.

Line-axle combined techniques is equivalent to a section by the step of a series of search, combination road network, combine every time and be equivalent to the section that extra to be converted into the section identical with section before, with the link flow that section before directly utilizing optimizes, it is relatively specific for the main line type intersection group of inner city.It optimizes the phase contrast of traffic signalization network by the form that " series connection " and " parallel connection " combines.

The span assuming j is from jo to jmax:

Step one: the start position at optimized arterial road defines crossing Jo in fact；

Step 2: combine each crossing on Trunk Road Network according to procedure below successively；

1. make { Δ }=Δ jo, Δ jo+1 ..., Δ j-1} (sets jth crossing as crucial crossing, offset optimization is the highest with jth crossing optimization level)；

2. { Δ }={ Δ } U{ Δ j} (wherein Δ j is the phase contrast previously merged) is made；

3. assume that each cycle can be divided into B period, each Period Length to be ω, if δ=1,2 ..., (B-1), increased by the phase contrast that each crossing is current and the phase contrast combined before, set up network phase difference evaluation model:

4. select suitable δ-value to obtain best evaluation effect so that { Δ } ← { Δ } δ.

Step 3: for isolated blob, { Δ j} is to particular value to specify crossing phase contrast to reach requirement to may specify the adjustment set of phase contrast.

Queue up in the crossing that the heavy turning traffic flow of the restriction and other flow direction remittance critical paths that optimize the phase contrast especially needed consideration downstream intersection traffic capacity of hypersaturated state intersection group is formed.Optimizing of hypersaturated state intersection group phase contrast needs to consider two constraints on the basis of original scheme: namely designed phase difference prevents crossing generation overflow phenomena and green light sky from putting phenomenon.

(4) signal controls real-time adaptive renewal

Optimizing and revising of split is active, the most most frequent parameter during the big parameter of traffic signal control system four (cycle, phase place phase sequence, split, phase contrast) adjusts.Single-point intersection Split Optimization real-time adaptive control key content is as follows:

1. the defining of split

After traffic control signal cycle duration is determined, the ratio of the effective green time of one of them signal phase and cycle duration is defined as the filtering ratio of signal phase, namelyWherein λ is split, and C is signal period duration, and ge is effective green time, and ge=g (green time)+A (yellow time)-L (starts lost time)；After C determines in the signal period, the optimization of split λ is optimized effective green time ge exactly, and determines that ge after determining display green time g simultaneously, optimize ge herein and just determine that optimization g.

2. Split Optimization arrange purpose and premise

After the signal period duration of traffic control system has optimized and determined, for the actual change of dynamically corresponding traffic flow, it is necessary to the green time of each phase place is carried out reallocation and adjusts by each cycle, so that the desired value that whole intersection traffic stream runs reaches optimization.Ensure that the optimum results of signal period and phase contrast is carried out simultaneously.Set up and assume:

1) signal period is reasonably determined；

2) phase place phase sequence is reasonably selected to optimize；

3) upstream and downstream of each entrance driveway line in crossing has all buried wagon detector underground；

4) mixed traffic flow impact reasonable consideration in minimax green time and copper sulfate basic etc. retrain on Split Optimization.

3. the determination of split initial value

When whistle control system brings into operation, the green time of phase place can be determined by offline optimization, or is invoked at the scheme of close temporal proximity before this, along with system operation can constantly on-line optimization adjustment, progressively met the running status of actual traffic stream by optimized algorithm.The ratio of each phase place Optimal green signal ratio in unlike signal cycle is substantially directly proportional to the ratio of phase place saturation volume ratio substantially:

In formula: gi, gj represent the Optimal green signal ratio of phase place i, j；Yi, yj represent the saturation volume ratio of phase place i, j；Qi, qj represent the flow of phase place i, j, and si, sj represent the saturation volume of phase place i, j.Therefore, when the signal period optimized determine, it is possible to according to etc. the principle of saturated distribution, carry out the determination of the split initial value under single-point real-time adaptive control according to the ratio of the saturation volume ratio of each phase place.

4. the constraints of Split Optimization

The constraints of Split Optimization is mainly signal period constraint, the constraint of minimax green time, Capacity Constraints:

In formula (22), i represents number of phases；Qi represents the flow of phase place i, C representation signal cycle；Gi represents the split of phase place i；S represents the saturation volume of phase place i；Xp represents the saturated acceptable maximum critical saturation of each phase place, generally takes Xp=0.95；Gmin represents the minimum green time of phase place, and gmax represents the maximum green light persistent period of phase place, and gmin and gmax can determine according to the traffic off-line in city, to advantageously ensure that traffic safety and to improve efficiency.

5. Split Optimization principle and algorithm

There is maximum difference with signal period optimization and be in that in Split Optimization: split is multi-C vector, and its dimension is equal to number of phases.Therefore, the complexity simplifying hyperspace optimization when ensureing to optimize precision and the memory cost taken is must take into consideration how when Split Optimization.Under the signal period determines situation, the distribution of split is generally of following methods:

The saturated timing method such as a.: based on fair principle, be used for the foundation of Split Optimization by saturation volume, has a feature simple, quick, near-optimization, but be open to traffic efficiency and service level not as total delay minimizes timing.

B. total delay minimizes timing method: the principle based on efficiency carries out split distribution, and the efficiency that is open to traffic and service level are best, but calculates time length, model needs complexity.

C. average traffic delay equal timing method: the average traffic delay making each phase place wagon flow is equal.

D. queuing rate equal timing method: the queuing rate making each phase place wagon flow is equal.

Based on this, select based on etc. the optimization that minimizes of the total delay of saturated distribution, using etc. the split of saturated distribution as the initial split of system optimizing, then the split that Step wise approximation is best.

6. Split Optimization flow process

According to above-mentioned analysis, the computing flow process of Split Optimization can be divided into three phases:

A. the initial dispensing phase of split

Upstream detector is utilized to detect the cycle traffic spirogram formula of generation in real time, principle according to equisaturation, carrying out original allocation according to the comparison signal period duration of the saturation volume ratio of each phase place, the split sum of each phase place obeys signal period constraint and minimax long green light time, the constraint of maximum critical saturation:

In formula (23), m represents the number of phases of crossing；

In formula (24), qi represents the volume of traffic of the i-th phase place, and Si represents the saturation volume of the i-th phase place.

B. the double optimization of split

If increasing the delay of phase place green time minimizing and the total revenue of stop frequency, more than the total losses suffered by the vehicle incured loss through delay by red light, green light timing just should be increased；Otherwise then should reduce green time.Based on this, optimizing from the prolongation phase place split on the main road of crossing of split, use climbing method, before green light is opened, the split performed with a upper cycle compares, search+Δ gs, 0, the change of intersection delay size in-Δ gs situation, find and incur loss through delay minimum split trimming scheme, now optimize all of non-prolongation phase place in tentative calculation crossing according to equisaturation principle, according to the ratio distribution split arriving saturation volume ratio, it follows that optimize and show that crossing extends the split of phase place and the split of other all non-prolongation phase places.If there is arbitrary non-prolongation phase place to be unsatisfactory for the constraint of minimax long green light time, the constraint of maximum critical saturation, then the saturated distribution such as newly carry out after meeting constraints above again:

In formula (25),WhereinFor the split of this cycle stretch-out phase place,For the split of a upper cycle stretch-out phase place, its optimization object function is:

C. the execution adjusting and optimizing of split

Owing to system upstream and downstream is provided with detector, therefore can controlling to save the situation of green time according to sensing, carry out saving redistributing of green time, to obtain better benefit, the delay value of accelerating system reduces further.Set up three class phase places: extend phase place, induced phase, master phase；It is primarily intended to is easy to when sensing controls reasonably to adjust the surplus and deficiency of each phase place green time, green time priority allocation unnecessary for non-prolongation phase place to the big prolongation phase place of the volume of traffic.

7. the principle that phase contrast is arranged is extended

Extend phase place and be generally arranged at that the volume of traffic is big or saturation volume is than big main road, its final green time, just can only can determine that after the split of other phase places is determined, it deducts the remaining time after other all phase places equal to cycle duration, and the total number extending phase place should be generally less than the sum arranging induced phase.

Introduce after extending phase place, it is necessary to extending phase place immediately following, after being arranged on induced phase, when induced phase is skipped or when having unnecessary green light to save, extending phase place and can obtain the whole green times having more of induced phase.Otherwise be arranged on before induced phase finishes then inadvisable extend phase place, because when induced phase not yet reaches maximum green light, then save green time and cannot adjust to extending phase place to ensure that the cycle duration optimized is carried out.

One main road direction is generally up to about arranging one and extends phase place, it is not necessary to each coordination direction all needs to be provided with prolongation phase place, particularly in two phase place situation.The phase place that master phase is intended merely to direction when regulation performs adjustment and introduces, it is not necessary to each crossing must be provided with master phase, particularly in two phase place situation.Redirected in a upper cycle if sensing controls phase place, then minimum green time when general phase place being arranged when the Split Optimization in next cycle is assigned to the induced phase split initialization of initial optimization.

8. based on the Split Optimization of double; two prolongation phase places

Single Split Optimization extended under phase condition is mainly described by foregoing, but can there is the not unique situation of prolongation phase place, and the large-scale crossing such as intersected at two major trunk roads exists typical four phase place situations.Now there are two and extend phase place, it is possible to adopt twocouese climbing method to be optimized search, obtain the Optimal green signal ratio under total delay minimum.Now to all non-prolongation phase places according to etc. saturated carry out green light distribution, to all prolongation phase places also according to etc. saturated carry out green light distribution, but be not equal to all phasetophases calculated in initial split situation and complete etc. saturated, but similar phasetophase relative etc. saturated.For typical four phase condition, the split making g1, g3 be non-prolongation phase place 1 and 3, g2, g4 are the split extending phase place 2 and 4, the Optimizing Search of split is adopted twocouese climbing method, then has:

The Split Optimization of double; two prolongation phase places adopts twocouese climbing method, and its optimization object function is:

In formula (28), d (g1), d (g3), d (g4) represent along extending each non-prolongation phase place delay value that phase place g2 direction uses climbing method to obtain；D (g11), d (g33), d (g44) represent along extending each non-prolongation phase place delay value that phase place g4 direction uses climbing method to obtain；Δ g2 represents the step-size in search extending phase place 2；Δ g4 represents the step-size in search extending phase place 4；Represent and extend the split that in phase place 2, a signal performs；Represent and extend the split that in phase place 4, a signal performs.

9. the interval of Split Optimization

Minimize to finally realize the delay that the signal period determines in situation, it is necessary to real-time matching is in the traffic of each import line being continually changing.When the adjustment interval of split is oversize, then real-time is poor, and it is excessively delayed that system tackles each phase place transport need change.When split adjustment interval is too short, then frequently adjust the instability that system will be brought to run.Owing to being spaced apart for two cycles as the optimization of signal period of strategy principal parameter, and split is as pure tactics parameter, its adjust interval should lower than the signal period, therefore split be optimized for each cycle once.The optimization time of split optimizes the split in next cycle before being typically in this cycle signal ended.Consider the time required for system optimization computing and communication transfer, therefore before first group of phase place green light of next cycle is opened the split of the necessary various phase place of optimization, its pre-set time, T was made up of following two parts: first be system optimization computing needed for time T1: depend on the performance of algorithm, calculating scale, hardware configuration situation；Second is that system schema performs required time T2: determined by communication transfer time and semaphore decoding time.

(5) set up new intersection signal timing and control collaborative linkage and commander's operational mode

The hypersaturated state intersection group signal set up controls to optimize new departure, in crossing, actual control signal machine environment runs three cycles, run with intelligent robot interlinked command with integrative design intersection optimization, 48 actions realizing artificial traffic signalization commander are harmonious with integrative design intersection, giving intelligent robot to optimize the traffic signalization function of hypersaturated state intersection group, the new function of manual coordination intersection traffic running command exercised by intelligent robot.

The present invention carries out dynamic traffic control with the key road segment trunk roads of inner city, Shenzhen road network Yu intersection group optimization and is optimized for example, specifically as shown in Fig. 9 to Figure 12, wherein, and City Road Network and associated cross mouth group static models figure centered by Fig. 9；Figure 10 is that intersection group dynamic traffic control Current Situation analyzes schematic diagram；Figure 11 is that Lianhua Road signalized crossing carries out dynamic traffic control optimization schematic diagram；Figure 12 is that Hong Li road signalized crossing carries out dynamic traffic control optimization schematic diagram.Described dynamic traffic control optimize mode particularly as follows:

1. the choosing of inner city road network critical path and intersection group

The important major trunk roads Xin Zhoulu in road network north-south, inner city, Shenzhen, is positioned at inner city, Feitian, and south from Fu Rong road, north, to Mei Hualu, is a key road segment of the external traffic undertaking the section along the line such as plum forests, Jing Tian, inner city, Xinzhou；This section is main through 4 grade separation crossings along the line: Beihuan grade separation, deep south grade separation, good fortune people's grade separation and riverfront grade separation and 2 level-crossings: Lianhua Road, Hong Li road.

2. traffic control Current Situation in Xinzhou road is analyzed

Plane hypersaturated state crossing, 2, Xinzhou road: Lianhua Road, Hong Li road gateway, section mostly be " four become three, three change two " track；Grade separation crossing, 4, Xinzhou road: Beihuan grade separation, deep south grade separation, good fortune people's grade separation and riverfront grade separation discrepancy ring road wagon flow have a strong impact on inner side main line wagon flow；Xinzhou road main line gradient is relatively big, and driver is blocked at visual angle, not easily finds gateway；Xinzhou Lu Shangyu red litchi north of a road import craspedodrome green time is long, causes that south imported vehicle accumulation is bigger: a. Cycle Length C=225s；B. green interval duration is 5s；C. in order to ensure that northing mouth vehicle queue does not spill over, the straight left phase place green time of northing mouth is additionally added.

Traffic control Current Situation in Xinzhou road is mainly manifested in:

1) traffic planning and design aspect: Xinzhou road is two-way all exists imbalance section, track, namely after shunting, number of track-lines, lower than number of track-lines before shunting, causes that vehicle interweaves and aggravates, in turn result in traffic congestion, queue up and spread backward；Grade separation discrepancy ramp distance is shorter, and traffic weave is serious, has a strong impact on the properly functioning of adjacent main line wagon flow.

2) Traffic Signal Timing aspect: crossing, Hong Li road is by north orientation south craspedodrome green time long (t=114s, C=225s), cause that vehicle accumulation is comparatively serious in the cycle from south to north, form Hong Li road crossing southing stomatodeum vehicle queue and extend to deep south grade separation；And north exit ramp is relatively low to prunus mume (sieb.) sieb.et zucc. China bus current density, link flow is unbalanced.

3) traffic operation and management aspect: surface road and grade separation road are mutually linked, the ring road that comes in and goes out is positioned at the climb and fall of joining place, causes that driver's sighting distance is bad, visual field property is poor；Road ground lacks inductivity identifier marking.

3. on the road of Xinzhou, associated cross mouth group's dynamic traffic control optimizes

Dynamic traffic control optimization design is carried out, it is achieved queue up do not spill over Lianhua Road, Lu Yuhong Li Lu crossing, Xinzhou northing mouth in Xinzhou road for Xinzhou road and Lianhua Road, Lu Yuhong litchi road, Xinzhou signalized crossing；Reduce Xinzhou road to queue up with Lianhua Road, Xinzhou Lu Yuhong Li Lu crossing southing mouth.

4. on the road of Xinzhou associated cross mouth group motion state based on the traffic control evaluation of intersection group

1) Philodendron ‘ Emerald Queen' is set up: make an on-the-spot survey finding and the timing data of associated cross mouth group on the spot according to Xinzhou road, during by peak, travel speed is decided to be 40km/h, and the system cycle is 194s；Green ripple scheme is divided into: southern by north orientation, two-way two kinds.

2) road design is improved: first, lane balance, and urban road setup of entrances and exits should keep the seriality in the basic track of main line, divides in gateway, interflow place maintains the balance of number of track-lines simultaneously；Second, reduce interwoven region vehicle coverage, in the scope of road section license, by completely isolated for main and auxiliary road, main road only collimates row, by the bypass after concentrating on broadening that interweaves, to improve interleaved order, it is ensured that through vehicles smooth and easy.

3) emulation quantitative assessment: by comparing simulation data flow and non-disguised survey profile data, error is 10.76%, meets simulation modeling flow condition.

4) optimize design and simulation relative analysis, incur loss through delay part: the delay of green ripple starting point Lianhua Road crossing is basically unchanged, and the delay of red litchi south of road import substantially reduces；Total travel time part: improved by Philodendron ‘ Emerald Queen' and road design so that morning peak two-way journey time in Xinzhou road has shortening；Queue length part: improved by green wave coordination, red litchi south of road northing stomatodeum through vehicles is queued up and is all improved.

nullThe traffic control method based on intersection group of the embodiment of the present invention and system are monitored in real time by building 360 ° of crossing panoramic videos and are modeled、Crossing assessment index and in-circuit emulation analysis、Oversaturated intersection critical path and control strategy optimization、Integrative design intersection optimizes " the four step methods " method with intelligent robot interlinked command，Set up the city traffic control intelligent robot based on intersection group，Solving urban road oversaturated intersection single-point operation optimization problem，Form intellectuality commander's urban road intersection、Oversaturated intersection、Oversaturated intersection group's traffic control and prioritization scheme，And adopt the traffic control intelligent robot based on intersection group，Realize intelligent robot and traffic signal controlling machine linkage，Set up integrative design intersection robot service mode，Promote traffic efficiency and service level that crossing single-point controls，Be conducive to scientifically and rationally the traffic flow of urban road network being carried out dynamic monitoring and optimizing tissue，Thus increasing substantially the operational efficiency of Traffic Systems，Alleviate urban traffic blocking.

The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all make within the spirit and principles in the present invention any amendment, equivalent replace and improvement etc., should be included in protection scope of the present invention it.

Claims (10)

1., based on a traffic control method for intersection group, comprise the following steps:

Step a: gather 360 °, crossing panoramic video by intelligent robot Real-time and Dynamic, sets up crossing moving model according to video data, and analyzes intersection group traffic characteristics according to crossing moving model；

Step b: carry out crossing assessment index and in-circuit emulation analysis according to traffic characteristics, identifies intersection group traffic circulation state；

Step c: the critical path of hypersaturated state intersection group is carried out oversaturated intersection signal timing dial design approach, adjusts hypersaturated state intersection group traffic signal control strategy；

Step d: the intersection group traffic signal control strategy after combustion adjustment, it is achieved crossing control signal timing designing scheme steady-state operation and intelligent robot interlinked command.

2. the traffic control method based on intersection group according to claim 1, it is characterised in that described step a also includes: crossing moving model is carried out operation situation monitoring；Described operation situation monitoring method includes: analysis intersection group blocks up and forms and evacuate collection and the process of mechanism and traffic circulation parameter；Described traffic circulation parameter acquisition and processing method specifically include: video encoder server and the modeling of traffic relatedness index.

3. the traffic control method based on intersection group according to claim 2, it is characterized in that, in described step b, described identification intersection group traffic circulation state specifically includes: intersection group scope defines, intersection group hypersaturated state identification, intersection group critical path detection and traffic parameter short-term prediction modeling and simulation.

4. the traffic control method based on intersection group according to claim 3, it is characterized in that, in described step c, the described critical path to hypersaturated state intersection group carries out oversaturated intersection signal timing dial design approach mode and specifically includes: intersection signal timing controls prioritization scheme static optimization；Dynamic cooperation traffic signalization intersection group；The traffic control strategy of hierarchical screening hypersaturated state intersection group；Coordinate timing scheme based on non-dominated sorted genetic algorithm optimization, control the benchmark timing scheme of dynamic optimization as signal；Traffic parameter real time dynamic optimization algorithm.

5. the traffic control method based on intersection group according to claim 1, it is characterized in that, in described step d, the method for described integrative design intersection prioritization scheme and intelligent robot interlinked command includes: the quiet collaborative traffic control of urban road oversaturated intersection group motion；Intersection group critical path coordinates the selection in control cycle；The phase contrast on-line optimization of oversaturated intersection group's critical path；Mixed traffic flow impact reasonable consideration in minimax green time and copper sulfate basic retrain on Split Optimization；Set up new intersection signal timing and control collaborative linkage and commander's operational mode.

6. the traffic control method based on intersection group according to claim 5, it is characterised in that described intersection group critical path is coordinated the Cycle Length computing formula of control cycle selection and is:

$C_1\≤\frac{Lh}{L_{\mathrm{\υ}}}\frac{(1-\frac{W}{L}-\frac{\left(SF\right)L_{\mathrm{\υ}}}{L})}{\frac{G_a-L}{C_1}}$

In above-mentioned formula, L is road section length；W is crossing, upstream width；Ga is the effective green time of downstream intersection；H for from when sailing vehicle headstock from；L is lost time；Lu is average traffic effective wagon degree；RL is shock wave dissipation place；C1 is the Cycle Length preventing overflow；SF is vehicle safety coefficient when emptying；U is from the velocity of wave sailing shock wave；V is the speed of next wagon flow first car；ω is the velocity of wave of parking shock wave；Δ is for coordinating to control phase contrast.

7. the traffic control system based on intersection group, it is characterised in that include intelligent robot, described intelligent robot includes the first video camera module, the second video camera module and data processor module；Described first video camera module and the second video camera module are connected with data processor module respectively；nullDescribed first video camera module and the second video camera module gather 360 °, crossing panoramic video for Real-time and Dynamic，And by the video data transmission of shooting to data processor module，Described data processor module is for setting up crossing moving model according to video data，Intersection group traffic characteristics is analyzed according to crossing moving model，Crossing assessment index and in-circuit emulation analysis is carried out according to intersection group traffic characteristics，Identify intersection group traffic circulation state，Thus the critical path of hypersaturated state intersection group is carried out oversaturated intersection signal timing dial design approach，Adjust hypersaturated state intersection group traffic signal control strategy，And control the intersection group traffic signal control strategy after intelligent robot combustion adjustment，Realize crossing control signal timing designing scheme steady-state operation and intelligent robot interlinked command.

8. the traffic control system based on intersection group according to claim 7, it is characterized in that, described first video camera module is the 360 ° of panoramic high-definition video cameras that highly can stretch, it is located at the above-head of intelligent robot, described second video camera module is HD video video camera, is located at the eye of intelligent robot.

9. the traffic control system based on intersection group according to claim 8, it is characterized in that, described data processor module includes model and sets up unit, traffic characteristics analytic unit, traffic circulation state recognition unit, policy optimization unit and scheme running unit；

Model sets up unit for receiving the first video camera module and the video data of the second video camera module transfer, and carry out video data sorting out after screening, image recognition and feature extraction etc. process and generate crossing real-time dynamic information environment, set up that picture is clear, the crossing moving model of broad view；

Traffic characteristics analytic unit for carrying out operation situation monitoring to crossing moving model, and analyzes intersection group traffic characteristics according to crossing moving model；

Traffic circulation state recognition unit, for carrying out crossing assessment index and in-circuit emulation analysis according to traffic characteristics, identifies intersection group traffic circulation state；

Policy optimization unit, for the critical path of hypersaturated state intersection group is carried out oversaturated intersection signal timing dial design approach and induction, adjusts hypersaturated state intersection group traffic signal control strategy；

Scheme running unit is for the intersection group traffic signal control strategy after combustion adjustment, it is achieved crossing control signal timing designing scheme steady-state operation and intelligent robot interlinked command.

10. the traffic control system based on intersection group according to claim 9, it is characterized in that, described intelligent robot also includes display module, described display module is for touching display screen, it is positioned at the body part of intelligent robot, described first video camera module and the second video camera module are connected with display module respectively, described first video camera module and the second video camera module are by the video data transmission of shooting to display module, and described display module is for showing the first video camera module and the video data of the second video camera module photograph.