CN105185103A - Road travel time management and control method - Google Patents

Abstract

The invention discloses a road travel time management and control method comprising the following steps of: S01, acquiring and processing traffic data by using a SCOOT-type traffic control system; S02, acquiring and processing traffic data by using a satellite positioning system; S03, fusing the traffic data transferred by the SCOOT-type traffic control system with the traffic data transferred by the satellite positioning system and generating a fusion output result by using a fusion server; and S04; sending the fusion output result to a traffic information release and intelligent dynamic traffic guidance system by using the fusion server. The road travel time management and control method prevents extra investment, greatly reduces the acquisition cost of dynamic traffic information, and solves problems that because of inconsistent data sampling intervals of the SCOOT-type traffic control system, the acquisition of the road travel time lacks good data basis and the satellite positioning system does not have a good matching effect in the vicinity of an intersection.

Description

A kind of management control method of Link Travel Time

Technical field

The present invention relates to a kind of technical field of intelligent traffic, particularly a kind of management control method of the Link Travel Time based on SCOOT class traffic control system and global position system.

Background technology

Link Travel Time is reflection road traffic state effective traffic parameter the most directly perceived, is the important foundation of carrying out congested in traffic management and dynamic path guidance.At present, the acquisition methods of Link Travel Time can be divided into directly collection on the whole and indirectly obtain two kinds.Wherein, direct acquisition method needs the terminus in each section to lay Car license recognition facility, and required cost is too high, is difficult to realize in Practical Project.Therefore, the indirect acquisition methods based on other traffic data is one of important subject of international traffic engineering field always.

SCOOT (Split, Cycle, OffsetOptimizationTechnique) system, i.e. " apart from optimisation technique when split, signal period and green light are started to walk ", it is one of control system of advanced person the most in the world, many big or middle cities such as Beijing, Guangzhou, Shenzhen of China all adopt this Systematical control urban transportation, have accumulated and by traffic datas a large amount of for persistent accumulation.In view of the control effects that SCOOT is excellent, domestic and international correlative study person develops again the similar traffic signal control system of some and SCOOT in succession, these systems all adopt the wagon detector layout scheme similar to SCOOT and timing designing algorithm, and itself and SCOOT system are referred to as SCOOT class traffic signal control system by the present invention.Along with GPS (Globalpositionsystem, GPS), Beidou satellite navigation system (BeiDouNavigationSatelliteSystem, etc. BDS) global position system is in the widespread adoption of field of traffic, mobile unit has been widely used in the aspects such as the vehicle schedulings such as taxi, logistic car, bus, monitoring and navigation, have accumulated a large amount of data equally.These two kinds of data sources have become of paramount importance traffic data collection means in urban road, utilize it to calculate journey time, improve traffic control, the harmony of traffic guidance and point duty has important practical significance for low cost.

But because SCOOT class traffic control system is the time interval obtain traffic data with long green light time, and the parameter that long green light time becomes when being, thus reduce the comparability of each time interval traffic data, Link Travel Time is obtained and lacks good data basis.Achievement in research relevant is at present also less, and all suppose that SCOOT class traffic control system wagon detector can provide traffic data according to certain fixing sampling interval, even there is no premised on some data that method provides at present, these hypothesis all do not meet the actual conditions of SCOOT class traffic control system, or are difficult in a short time be realized by engineering means.

For global position system, Floating Car sample size determines that Link Travel Time obtains the key factor of effect, but due to system operation cost, to determine Floating Car vehicle often more single, the vehicle of traffic flow forms but more complicated, dissimilar vehicle has different operation characteristics, therefore, the operation conditions of Floating Car is difficult to the integral status representing traffic flow completely.In addition, due to blocking satellite-signal such as trees, pile and tunnels, GPS check frequency can be caused in some localities, the calculating effect of Link Travel Time can be reduced equally.In addition, there are some researches show, due to the impact that signal controls, vehicle loiters near crossing, causes map match poor effect, and then causes Link Travel Time to obtain poor effect.

Summary of the invention

The object of the invention is: the management control method that a kind of Link Travel Time is provided, it obtains for the Link Travel Time caused due to reasons such as SCOOT class traffic control system data sampling interval are inconsistent problems such as lacking good data basis, proposes a kind of effectively based on the Link Travel Time acquisition methods of SCOOT class traffic control system data; For the problem that the Link Travel Time output effect caused due to reasons such as sample size and subregion map match difficulties is not good, propose a kind of effectively based on the Link Travel Time acquisition methods of global position system data; Based on the Link Travel Time Output rusults of above-mentioned two kinds of data sources, propose a kind of Urban road journey time and merge acquisition methods, improve the output effect of Link Travel Time further.

The technical scheme achieved the above object is: a kind of management control method of Link Travel Time, and it comprises the following steps,

S01: utilize SCOOT class traffic control system to gather traffic data and to go forward side by side row relax, and the traffic data after process is sent to converged services device;

S02: utilize global position system to gather traffic data and to go forward side by side row relax, and the traffic data after process is sent to converged services device;

S03: utilize described converged services device to be merged by the traffic data that described SCOOT class traffic control system and described global position system are passed to, and generate fusion Output rusults;

S04: utilize described converged services device that fusion Output rusults is sent to Traffic information demonstration, Intelligent Dynamic system for traffic guiding.

Described step S01 comprises the following steps,

S011: the first space scale utilizing SCOOT class traffic control system to obtain Link Travel Time and very first time yardstick are determined to extract;

S012: in periodic intervals, gather the traffic data at crossing, and utilize communication system it to be uploaded to the information server of the described SCOOT class traffic control system of traffic information center, described traffic data comprises traffic parameter data and traffic signalization data, and described traffic parameter data comprise the magnitude of traffic flow, average speed, occupation rate; Described traffic signalization data comprise cycle duration, long green light time;

S013: utilize the information server of described SCOOT class traffic control system to carry out virtual traffic data sequence structure to actual traffic data;

S014: utilize BP neural network to design Link Travel Time first and obtain model, generate the first Output rusults of Link Travel Time, and the first Output rusults is sent to described converged services device.

The first space scale in described step S011 is the road between continuous two stop line reverse extending lines.

Described step S013 comprises the following steps,

S0131: traffic parameter data are converted;

Described step S0131 comprises the following steps,

S01311: the conversion of the magnitude of traffic flow:

If vehicle evenly arrives, shown in the following formula of the vehicle number passed through in the unit interval:

${\stackrel{\‾}{q}}_i\left(s\right)=\frac{q_i\left(s\right)}{C_i\left(s\right)}$

In formula, C _i(s) and q _is () is respectively cycle and the magnitude of traffic flow in i-th time interval of actual traffic data sequence; for the vehicle number passed through in the unit time;

Then the magnitude of traffic flow from the mapping relations that the conversion of actual traffic data sequence is virtual traffic data sequence is:

$q_j\left(x\right)=\underset{N=i}{\overset{i+n}{\Σ}}{\stackrel{\‾}{q}}_i\left(s\right)\×t_i$

Wherein, q _jx () is the total flow in a virtual traffic data sequence jth time interval; t _ifor a virtual traffic data sequence jth time interval is positioned at the duration in i-th time interval of actual traffic data sequence; N is a number virtual traffic data sequence jth time interval taking actual traffic data sequence;

S01312: average speed is converted;

Average speed refers to the mean value of vehicle spot speed, and its mapping relations being virtual traffic data sequence from the conversion of actual traffic data sequence are:

$v_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}\left(v_i\right(s)\×t_i\×{\stackrel{\‾}{q}}_i(s\left)\right)/{\mathrm{\Σ}}_{N=i}^{i+n}\left({\stackrel{\‾}{q}}_i\right(s)\×t_i)$

Wherein, v _jx () is the average speed in a virtual traffic data sequence jth time interval; v _is () is the average speed in i-th time interval of actual traffic data sequence, for the vehicle number passed through in the unit time; t _ifor a virtual traffic supplemental characteristic sequence jth time interval position and the duration in i-th time interval of actual traffic data sequence;

S01313: occupation rate is converted;

Occupation rate refers to the ratio of the time accumulated value that vehicle takies and minute, and its mapping relations being virtual traffic data sequence from the conversion of actual traffic data sequence are:

$o_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}\left(o_i\right(s)\×t_i)/{\mathrm{\Σ}}_{N=i}^{i+n}\left(t_i\right)$

Wherein, o _jx () is the occupation rate in a virtual traffic data sequence jth time interval; o _is () is the occupation rate in i-th time interval of actual traffic data sequence, t _ifor a virtual traffic supplemental characteristic sequence jth time interval position and the duration in i-th time interval of actual traffic data sequence;

S0132: traffic signalization data are converted,

Described step S0132 comprises the following steps,

S01321: cycle duration and long green light time conversion,

The mode of continuous micro adjustment that what the optimization of SCOOT class traffic control system timing parameter adopted is, the long green light time in adjacent signals cycle and cycle duration change less, and cycle duration and long green light time from the mapping relations that the conversion of actual traffic data sequence is virtual traffic data sequence are:

$g_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}{g}_i\left(s\right)/n$

$C_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}{C}_i\left(s\right)/n$

Wherein, g _j(x), C _j(x) be respectively a virtual traffic data sequence jth time interval average long green light time, average period duration; g _is () is the long green light time in i-th time interval of actual traffic data sequence, C _is () is actual traffic data sequence i-th periodic intervals.

Described BP neural network in described step S014 is the three layers of BP neural network comprising hidden layer, and its training function adopts Sigmiod function; Training error adopts root-mean-square error RMSE.

Described step S02 comprises the following steps,

S021: the second space yardstick utilizing global position system to obtain Link Travel Time and the second time scale are determined to extract, and its second space yardstick is consistent with the first space scale, and its very first time yardstick is consistent with very first time yardstick;

S022: in certain sampling time interval, utilizes global position system to gather the various information datas of mobile unit, and by communication apparatus, by the various uploading information datas of mobile unit to the information server of the global position system of traffic information center;

S023: the information server and the generalized information system that utilize global position system, to running time and the intersection delay time design second acquisition model of basic road stroke, generate the second Output rusults of Link Travel Time, and the second Output rusults is sent to described converged services device.

Described step S023 comprises the following steps,

S0231: the model of the time of vehicle operation of design basic road;

Described step S0231 comprises the following steps,

S02311: bicycle basic road running time calculates,

Suppose that vehicle remains a constant speed traveling between adjacent positioned point, then border, the section moment extracts formula and is

$t^{\′\′}\left(t\right)=t^{\′}(t+1)=\frac{L^{\′}\left(t\right)\×t(t-n(t\left)\right)+L^{\′}(t-n(t\left)\right)\×t\left(t\right)}{L^{\′}\left(t\right)+L^{\′}(t-n(t\left)\right)}$

Wherein, " (t), t ' (t+1) represent current road segment terminal border moment and downstream road section starting point border moment to t respectively; T (t), t (t-n (t)) represent the location moment of current matching point and last match point data respectively; L ' (t), L ' (t-n (t)) represent the distance on current matching point and last match point data and current road segment terminal border respectively;

Then bicycle basic road running time computing formula is

T′＝t″-t′

Wherein: T ' is single sample car basic road running time; T " and t ' is respectively the moment of sample car through start and end border, section;

S02312: sample car basic road running time calculates,

Sample car basic road running time refers to the average level through all bicycle basic road running times of specific road section in special time yardstick, directly can get the average of bicycle basic road running time, namely

$\hat{T}={\mathrm{\Σ}}_{j=1}^n{T}_j^{\′}$

Wherein, for sample car basic road running time;

S02313: traffic flow basic road running time calculates,

Traffic flow basic road running time refers to the average level through all vehicle basic road running time times of specific road section in special time yardstick, and employing statistical study computing formula is

$T\left(l_1\right)=f\left(\hat{T}\right)$

Wherein: T (l ₁) be traffic flow basic road running time; The regression function of f for utilizing regretional analysis to set up;

S0232: design intersection delay time model;

Described step S0232 comprises the following steps,

S02321: intersection delay Time Calculation,

Robert Webster formula is utilized to estimate intersection delay, namely

$\stackrel{\‾}{d}=\frac{C{(1-u)}^2}{2(1-ux)}+\frac{x^2}{2q(1-x)}-0.65{\left(\frac{C}{q^2}\right)}^{\frac{1}{3}}\·x^{(2+5u)}$

Wherein, for the intersection delay time, C is cycle duration; G is long green light time; Q is the magnitude of traffic flow; U is split, i.e. g/C; X is saturation degree, i.e. (q × C)/(s × g); S is saturation volume rate.

S0233: Link Travel Time calculates; Link Travel Time is basic road running time and intersection delay time sum, namely

$T=T\left(l_1\right)+\stackrel{\‾}{d}$

The advantage of technique scheme is: the management control method of Link Travel Time of the present invention, without the need to extra input, significantly reduces the procurement cost of dynamic information; Overcome the Link Travel Time caused due to reasons such as SCOOT class traffic control system data sampling interval are inconsistent and obtain difficult problems such as lacking good data basis, improve Link Travel Time output effect; Overcome because the Link Travel Time that cause of reason such as sample size and subregion map match difficulty obtain the difficult problem of poor effect, improve Link Travel Time output effect; Fusion based on two kinds of data sources obtains the output effect that model further increases Link Travel Time, and then traffic signalization can be significantly improved, transport information guides and the coordination degree of point duty, for improving the congested in traffic decision support of dredging effect and providing stronger to greatest extent.

Accompanying drawing explanation

Below in conjunction with drawings and Examples, the invention will be further described.

The overview flow chart of the Link Travel Time management control method of Fig. 1 embodiment of the present invention.

The Link Travel Time based on SCOOT class traffic control system of Fig. 2 embodiment of the present invention obtains process flow diagram.

The Link Travel Time based on global position system of Fig. 3 embodiment of the present invention obtains process flow diagram.

The space scale determination schematic diagram of the SCOOT class traffic control system of Fig. 4 embodiment of the present invention.

The SCOOT class traffic control system cycle of Fig. 5 embodiment of the present invention and long green light time schematic diagram.

The BP Artificial Neural Network Structures figure based on SCOOT class traffic control system of Fig. 6 embodiment of the present invention.

Border, the section moment extracting method schematic diagram based on global position system of Fig. 7 embodiment of the present invention.

Fig. 8 is that the Link Travel Time of the embodiment of the present invention merges acquisition process flow diagram.

Embodiment

Embodiment: as shown in Figure 1, a kind of management control method of Link Travel Time, traffic information collection subsystem wherein in the City ITS that adopts of the present embodiment, primarily of part compositions such as airfield equipment, communication system and traffic information center associated servers, comprises 3 modules below concrete operations flow process.

1, based on SCOOT class traffic control system: the space scale obtain Link Travel Time and time scale are determined; Utilizing communication system, take cycle as the time interval, the traffic parameter data at crossing and signal lamp arrangement information is uploaded to traffic information center SCOOT class traffic control system information server; SCOOT class traffic control system information server carries out virtual traffic data sequence structure to actual traffic data sequence; SCOOT class traffic control system information server utilizes Link Travel Time to obtain model, generates Link Travel Time Output rusults, and is sent to converged services device.

2, based on global position system: the space scale obtain Link Travel Time and time scale are determined; The fields such as car number, car plate, precision, latitude, instantaneous velocity, deflection and positioning time at a certain time interval, by Wireless Telecom Equipment, are uploaded to traffic information center global position system information server by mobile unit; Global position system information server is in conjunction with GIS (GeographicInformationSystem, GIS) system, the basic road running time utilizing the present invention to propose and intersection delay time obtain model, generate Link Travel Time Output rusults, and be sent to converged services device.

3, Link Travel Time merges acquisition module: wherein converged services device receives the Link Travel Time Output rusults based on two kinds of data sources, utilize Link Travel Time to merge and obtain model, generate and merge Output rusults, and be sent to the follow-up subsystem of the intelligent transportation system such as Traffic information demonstration, dynamic traffic guidance, to provide Data support for it.

The management control method of the Link Travel Time of the present embodiment, it comprises the following steps.

S01: utilize SCOOT class traffic control system to gather traffic data and to go forward side by side row relax, and the traffic data after process is sent to converged services device.

As shown in Figure 2, described step S01 comprises the following steps.

As shown in Figure 4, S011: utilize SCOOT class traffic control system to determine to extract to the first space scale of link travel and very first time yardstick; The first space scale in described step S011 is the road between continuous two stop line reverse extending lines.

Concrete: the first space scale defining method that Link Travel Time obtains refers to, using the border of upstream, track stop line reverse extending line position as pavement section, the road between continuous two stop line reverse extending lines is referred to as section; The very first time yardstick defining method that Link Travel Time obtains refers to, considers described virtual traffic data sequence and builds the minimum mental habit with traveler of the road network total delay caused, advise being advisable with 150s or 300s.

S012: in periodic intervals, gather the traffic data at crossing, and utilize communication system it to be uploaded to the information server of the described SCOOT class traffic control system of traffic information center, described traffic data comprises traffic parameter data and traffic signalization data, and described traffic parameter data comprise the magnitude of traffic flow, average speed, occupation rate; Described traffic signalization data comprise cycle duration, long green light time.

S013: utilize the information server of described SCOOT class traffic control system to carry out virtual traffic data sequence structure to actual traffic data.SCOOT class traffic control system virtual traffic data sequence builds and refers to, by the transformational relation of correspondence, by the virtual traffic data sequence that the conversion of the actual traffic data sequence of dynamic change is set time yardstick, specifically comprise the conversion of the magnitude of traffic flow, average speed, occupation rate and traffic control parameter etc.

Wherein, 1) the space scale defining method of Link Travel Time acquisition: SCOOT class traffic control system is smallest record unit with crossing, and the traffic data obtained comprises traffic signalization data and traffic parameter data two type.Wherein, the former mainly comprises cycle, green time etc.; The latter mainly comprises the magnitude of traffic flow, average speed and occupation rate etc.The wagon detector of SCOOT class traffic control system is laid near the stop line reverse extending line position of upstream, track, and the sampling interval of traffic parameter data is long green light time.In order to the basic data of SCOOT class traffic control system can be made full use of, using the border of upstream, track stop line reverse extending line position as pavement section, the road between continuous two stop line reverse extending lines is referred to as section.

2) the time scale defining method of Link Travel Time acquisition: the inconsistent meeting at virtual traffic data sequence and actual traffic data sequence time interval causes corresponding data to extract some difference to some extent, thus regular hour delay can be produced.When determining the time scale of Link Travel Time Estimation, the mental habit of the minimum and traveler of road network total delay should be considered, advise being advisable with 150s or 300s.

Concrete step is as follows.

Described step S013 comprises the following steps.

S0131: traffic parameter data are converted.Described step S0131 comprises the following steps.

S01311: the conversion of the magnitude of traffic flow.

If vehicle evenly arrives, the vehicle number passed through in the unit interval is as shown in formula (1):

${\stackrel{\‾}{q}}_i\left(s\right)=\frac{q_i\left(s\right)}{C_i\left(s\right)}---\left(1\right)$

Wherein, C in formula _i(s) and q _is () is respectively actual traffic data sequence i-th periodic intervals, the magnitude of traffic flow; for the vehicle number passed through in the unit time.

Then the magnitude of traffic flow from the mapping relations that the conversion of actual traffic data sequence is virtual traffic data sequence is:

$q_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}{\stackrel{\‾}{q}}_i\left(s\right)\×t_i---\left(2\right)$

Wherein, q _jx () is the total flow in a virtual traffic data sequence jth time interval; t _ifor a virtual traffic supplemental characteristic sequence jth time interval position and the duration in i-th time interval of actual traffic data sequence; N is a number virtual traffic data sequence jth time interval taking actual traffic data sequence.

S01312: average speed is converted.

Average speed refers to the mean value of vehicle spot speed, and its mapping relations being virtual traffic data sequence from the conversion of actual traffic data sequence are:

$v_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}\left(v_i\right(s)\×t_i\×{\stackrel{\‾}{q}}_i(s\left)\right)/{\mathrm{\Σ}}_{N=i}^{i+n}\left({\stackrel{\‾}{q}}_i\right(s)\×t_i)---\left(1\right)$

Wherein, v _jx () is the average speed in a virtual traffic data sequence jth time interval; v _is () is the average speed in i-th time interval of actual traffic data sequence, for the vehicle number passed through in the unit time; t _ifor a virtual traffic supplemental characteristic sequence jth time interval position and the duration in i-th time interval of actual traffic data sequence.

S01313: occupation rate is converted.

Occupation rate refers to the ratio of the time accumulated value that vehicle takies and minute, and its mapping relations being virtual traffic data sequence from the conversion of actual traffic data sequence are:

$o_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}\left(o_i\right(s)\×t_i)/{\mathrm{\Σ}}_{N=i}^{i+n}\left(t_i\right)---\left(4\right)$

Wherein, o _jx () is the occupation rate in a virtual traffic data sequence jth time interval; o _is () is the occupation rate in i-th time interval of actual traffic data sequence, t _ifor a virtual traffic supplemental characteristic sequence jth time interval position and the duration in i-th time interval of actual traffic data sequence;

As shown in Figure 5, S0132: traffic signalization data are converted, and described step S0132 comprises the following steps.

S01321: cycle duration and long green light time conversion.

The mode of continuous micro adjustment that what the optimization of SCOOT class traffic control system timing parameter adopted is, the long green light time in adjacent signals cycle and cycle duration change less, and cycle duration and long green light time from the mapping relations that the conversion of actual traffic data sequence is virtual traffic data sequence are:

$g_j\left(x\right)={\Σ}_{N=i}^{i=n}{g}_i\left(s\right)/n---\left(5\right)$

$C_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}{C}_i\left(s\right)/n---\left(6\right)$

Wherein, g _j(x), C _j(x) be respectively a virtual traffic data sequence jth time interval average long green light time, average period duration; g _is () is the long green light time in i-th time interval of actual traffic data sequence, C _is () is actual traffic data sequence i-th periodic intervals.

The inconsistent meeting at virtual traffic data sequence and actual traffic data sequence time interval causes corresponding data to extract some difference to some extent, thus can produce regular hour delay.When determining the time scale that Link Travel Time is simulated, the mental habit of the minimum and traveler of road network total delay should be considered, advise being advisable with 150s or 300s.

Described BP neural network in described step S014 is the three layers of BP neural network comprising hidden layer, and its training function adopts Sigmiod function; Training error adopts root-mean-square error RMSE.

As shown in Figure 6, concrete: the Link Travel Time based on SCOOT class traffic control system obtains model and refers to, build three layers of BP neural network containing a hidden layer, using the virtual traffic data sequence of SCOOT class traffic signal control system as input, comprise the magnitude of traffic flow, average speed, occupation rate, long green light time and cycle duration totally 5 parameters; Using the Link Travel Time of same space and time scale as output; Hidden nodes is set to 9 according to the principle of (2 × input neuron number-1); Training function adopts Sigmiod function; Training error adopts root-mean-square error RMSE.

S014: utilize BP neural network to design Link Travel Time first and obtain model, generate the first Output rusults of Link Travel Time, and the first Output rusults is sent to described converged services device.

BP neural network is one of neural network be most widely used at present, its characteristic that there is study and store a large amount of input-output mode map relation, and without the need to disclosing the mathematical model describing this mapping relations in advance, the present invention will set up Link Travel Time analogy model based on it.

Because three layers of BP neural network containing a hidden layer can approach any non-linear continuous function, therefore hidden layer number is set to 1.Using the virtual traffic data sequence of the SCOOT class traffic signal control system of aforementioned structure as input, comprise the magnitude of traffic flow, average speed, occupation rate, long green light time and cycle duration totally 5 parameters; Using the Link Travel Time of same space and time scale as output; Hidden nodes is set to 9 according to the principle of (2 × input neuron number-1); Training function adopts Sigmiod function; Training error adopts root-mean-square error RMSE.

S02: utilize global position system to gather traffic data and to go forward side by side row relax, and the traffic data after process is sent to converged services device.

Merge to make Link Travel Time and obtain model there is good data basis, global position system module volume yardstick and time scale consistent with SCOOT class traffic control system module.

In order to solve the global position system problem that effect is not good near crossing, Link Travel Time is divided into nature link travel time and intersection delay time two parts, and proposes to calculate this two-part journey time respectively, l ₁for natural road section length, l ₂for crossing segment length, the two, is generally as the criterion with real road investigation for separation with canalization starting point, and for the road not having canalization section, trunk roads, secondary distributor road, branch road distance stop line gets 70 ~ 90m, 50 ~ 70m, 30 ~ 40m respectively.

As shown in Figure 3, described step S02 comprises the following steps.

S021: the space scale that the Link Travel Time based on global position system obtains and time scale consistent with SCOOT class traffic control system.Wherein, the second space yardstick utilizing global position system to obtain Link Travel Time and the second time scale are determined to extract, and its second space yardstick is consistent with the first space scale, and its very first time yardstick is consistent with very first time yardstick.

S022: in certain sampling time interval, utilizes global position system to gather the various information datas of mobile unit, and by communication apparatus, by the various uploading information datas of mobile unit to the information server of the global position system of traffic information center.

S023: the information server and the generalized information system that utilize global position system, second analogy model is designed to the running time of basic road and intersection delay time, generate the second Output rusults of Link Travel Time, and the second Output rusults is sent to described converged services device.Link Travel Time based on global position system data obtains model and refers to, Link Travel Time is divided into nature link travel time and intersection delay time two parts, and calculates this two-part journey time respectively.

1, the acquisition of basic road running time comprises the acquisition of bicycle basic road running time, sample car basic road running time obtains and traffic flow basic road running time obtains three parts;

2, Robert Webster formulae discovery intersection delay is utilized.Concrete step is as follows.

Described step S023 comprises the following steps.

S0231: design basic road running time obtains model.

Described step S0231 comprises the following steps.

S02311: bicycle basic road running time calculates.

As shown in Figure 7, global position system data generally can provide the fields such as car number, car plate, precision, latitude, instantaneous velocity, deflection and positioning time, specifically as shown in table 1.

Table 1 global position system data instance.

Suppose that vehicle remains a constant speed traveling between adjacent positioned point, then border, the section moment extracts formula and is

$t^{\′\′}\left(t\right)=t^{\′}(t+1)=\frac{L^{\′}\left(t\right)\×t(t-n(t\left)\right)+L^{\′}(t-n(t\left)\right)\×t\left(t\right)}{L^{\′}\left(t\right)+L^{\′}(t-n(t\left)\right)}---\left(7\right)$

Wherein, " (t), t ' (t+1) represent current road segment terminal border moment and downstream road section starting point border moment to t respectively; T (t), t (t-n (t)) represent the location moment of current matching point and last match point data respectively; L ' (t), L ' (t-n (t)) represent the distance on current matching point and last match point data and current road segment terminal border respectively;

Then bicycle basic road running time computing formula is

T′＝t″-t′(8)

Wherein: T ' is single sample car basic road running time; T " and t ' is respectively the moment of sample car through start and end border, section.

S02312: sample car basic road running time calculates.

Sample car basic road running time refers to the average level through all bicycle basic road running times of specific road section in special time yardstick, directly can get the average of bicycle basic road running time, namely

$\hat{T}={\mathrm{\Σ}}_{j=1}^n{T}_j^{\′}---\left(9\right)$

Wherein, for sample car basic road running time.

S02313: traffic flow basic road running time calculates.

Traffic flow basic road running time refers to the average level through all vehicle basic road running time times of specific road section in special time yardstick, and employing statistical study computing formula is

$T\left(l_1\right)=f\left(\hat{T}\right)---\left(10\right)$

Wherein: T (l ₁) be traffic flow basic road running time; The regression function of f for utilizing regretional analysis to set up, those skilled in the art should understand, according to sample car basic road running time, also other statistical analysis technique can be adopted to calculate traffic flow basic road running time, and fundamental purpose is the average level that can represent all vehicle basic road running time time;

S0232: design intersection delay time model; Described step S0232 comprises the following steps.

S02321: intersection delay Time Calculation.

Robert Webster formula is utilized to estimate intersection delay, namely

$\stackrel{\‾}{d}=\frac{C{(1-u)}^2}{2(1-ux)}+\frac{x^2}{2q(1-x)}-0.65{\left(\frac{C}{q^2}\right)}^{\frac{1}{3}}\·x^{(2+5u)}---\left(11\right)$

Wherein, for the intersection delay time, C is cycle duration; G is long green light time; Q is the magnitude of traffic flow; U is split, i.e. g/C; X is saturation degree, i.e. (q × C)/(s × g); S is saturation volume rate.

S0233: Link Travel Time calculates; Link Travel Time is basic road running time and intersection delay time sum, namely

$T=T\left(l_1\right)+\stackrel{\‾}{d}.---\left(12\right)$

S03: utilize described converged services device to be merged by the traffic data that described SCOOT class traffic control system and described global position system are passed to, and generate fusion Output rusults.

As shown in Figure 8, described step S03 comprises the following steps.

S031: build three layers of BP neural network containing a hidden layer.

S032: using described first Output rusults of SCOOT class traffic signal control system, global position system second Output rusults and sample car quantity as input.

S033: using the Link Travel Time of same space and time scale as output.

Concrete: Link Travel Time merges acquisition model and refers to, builds three layers of BP neural network containing a hidden layer, using SCOOT class traffic signal control system module result of calculation, global position system module result of calculation and sample car quantity as input; Using the Link Travel Time of same space and time scale as output; Hidden nodes is set to 5 according to the principle of (2 × input neuron number-1); Training function adopts Sigmiod function; Training error adopts root-mean-square error RMSE.

S04: utilize described converged services device that fusion Output rusults is sent to Traffic information demonstration, Intelligent Dynamic system for traffic guiding.

It should be pointed out that for the present invention through absolutely proving, also can have the embodiment of multiple conversion and remodeling, be not limited to the specific embodiment of above-mentioned embodiment.Above-described embodiment is as just explanation of the present invention, instead of limitation of the present invention.In a word, protection scope of the present invention should comprise those for.

Claims (9)

1. a management control method for Link Travel Time, is characterized in that: it comprises the following steps,

S01: utilize SCOOT class traffic control system to gather traffic data and to go forward side by side row relax, and the traffic data after process is sent to converged services device;

S02: utilize global position system to gather traffic data and to go forward side by side row relax, and the traffic data after process is sent to converged services device;

S03: utilize described converged services device to be merged by the traffic data that described SCOOT class traffic control system and described global position system are passed to, and generate fusion Output rusults;

S04: utilize described converged services device that fusion Output rusults is sent to Traffic information demonstration, Intelligent Dynamic system for traffic guiding.

2. the management control method of Link Travel Time according to claim 1, is characterized in that: described step S01 comprises the following steps,

S011: the first space scale utilizing SCOOT class traffic control system to obtain Link Travel Time and very first time yardstick are determined to extract;

S012: in periodic intervals, gather the traffic data at crossing, and utilize communication system it to be uploaded to the information server of the described SCOOT class traffic control system of traffic information center, described traffic data comprises traffic parameter data and traffic signalization data, and described traffic parameter data comprise the magnitude of traffic flow, average speed, occupation rate; Described traffic signalization data comprise cycle duration, long green light time;

S013: utilize the information server of described SCOOT class traffic control system to carry out virtual traffic data sequence structure to actual traffic data;

S014: utilize BP neural network to design Link Travel Time first and obtain model, generate the first Output rusults of Link Travel Time, and the first Output rusults is sent to described converged services device.

3. the management control method of Link Travel Time according to claim 2, is characterized in that: the first space scale in described step S011 is the road between continuous two stop line reverse extending lines.

4. the management control method of Link Travel Time according to claim 2, is characterized in that: described step S013 comprises the following steps,

S0131: traffic parameter data are converted;

Described step S0131 comprises the following steps,

S01311: the conversion of the magnitude of traffic flow:

If vehicle evenly arrives, shown in the following formula of the vehicle number passed through in the unit interval:

${\stackrel{\‾}{q}}_i\left(s\right)=\frac{q_i\left(s\right)}{C_i\left(s\right)}---\left(1\right)$

In formula, C _i(s) and q _is () is respectively cycle and the magnitude of traffic flow in i-th time interval of actual traffic data sequence; for the vehicle number passed through in the unit time;

Then the magnitude of traffic flow from the mapping relations that the conversion of actual traffic data sequence is virtual traffic data sequence is:

$q_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}{\stackrel{\‾}{q}}_i\left(s\right)\×t_i$

Wherein, q _jx () is the total flow in a virtual traffic data sequence jth time interval; t _ifor a virtual traffic data sequence jth time interval is positioned at the duration in i-th time interval of actual traffic data sequence; N is a number virtual traffic data sequence jth time interval taking actual traffic data sequence;

S01312: average speed is converted;

Average speed refers to the mean value of vehicle spot speed, and its mapping relations being virtual traffic data sequence from the conversion of actual traffic data sequence are:

$v_j\left(x\right)={\Σ}_{N=i}^{i+n}\left(v_i\right(s)\×t_i\×{\stackrel{\‾}{q}}_i(s\left)\right)/{\Σ}_{N=i}^{i+n}\left({\stackrel{\‾}{q}}_i\right(s)\×t_i)$

Wherein, v _jx () is the average speed in a virtual traffic data sequence jth time interval; v _is () is the average speed in i-th time interval of actual traffic data sequence, for the vehicle number passed through in the unit time; t _ifor a virtual traffic supplemental characteristic sequence jth time interval position and the duration in i-th time interval of actual traffic data sequence;

S01313: occupation rate is converted;

Occupation rate refers to the ratio of the time accumulated value that vehicle takies and minute, and its mapping relations being virtual traffic data sequence from the conversion of actual traffic data sequence are:

$o_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}\left(o_i\right(s)\×t_i)/{\mathrm{\Σ}}_{N=i}^{i+n}\left(t_i\right)$

Wherein, o _jx () is the occupation rate in a virtual traffic data sequence jth time interval; o _is () is the occupation rate in i-th time interval of actual traffic data sequence, t _ifor a virtual traffic supplemental characteristic sequence jth time interval position and the duration in i-th time interval of actual traffic data sequence;

S0132: traffic signalization data are converted,

Described step S0132 comprises the following steps,

S01321: cycle duration and long green light time conversion,

The mode of continuous micro adjustment that what the optimization of SCOOT class traffic control system timing parameter adopted is, the long green light time in adjacent signals cycle and cycle duration change less, and cycle duration and long green light time from the mapping relations that the conversion of actual traffic data sequence is virtual traffic data sequence are:

$g_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}{g}_i\left(s\right)/n$

$C_j\left(x\right)={\mathrm{\Σ}}_{N=i}^{i+n}{C}_i\left(s\right)/n$

Wherein, g _j(x), C _j(x) be respectively a virtual traffic data sequence jth time interval average long green light time, average period duration; g _is () is the long green light time in i-th time interval of actual traffic data sequence; C _is () is the cycle in i-th time interval of actual traffic data sequence.

5. the management control method of Link Travel Time according to claim 2, is characterized in that: the described BP neural network in described step S014 is the three layers of BP neural network comprising hidden layer, and its training function adopts Sigmiod function; Training error adopts root-mean-square error RMSE.

6. the management control method of Link Travel Time according to claim 1, is characterized in that: described step S02 comprises the following steps,

S021: the second space yardstick utilizing global position system to obtain Link Travel Time and the second time scale are determined to extract, and its second space yardstick is consistent with the first space scale, and its very first time yardstick is consistent with very first time yardstick;

S022: in certain sampling time interval, utilizes global position system to gather the various information datas of mobile unit, and by communication apparatus, by the various uploading information datas of mobile unit to the information server of the global position system of traffic information center;

S023: the information server and the generalized information system that utilize global position system, to running time and the intersection delay time design second acquisition model of basic road, generate the second Output rusults of Link Travel Time, and the second Output rusults is sent to described converged services device.

7. the management control method of Link Travel Time according to claim 6, is characterized in that: described step S023 comprises the following steps,

S0231: design basic road travels and obtains model;

Described step S0231 comprises the following steps,

S02311: bicycle basic road running time calculates,

Suppose that vehicle remains a constant speed traveling between adjacent positioned point, then border, the section moment extracts formula and is

$t^{\′\′}\left(t\right)=t^{\′}(t+1)=\frac{L^{\′}\left(t\right)\×t(t-n(t\left)\right)+L^{\′}(t-n(t\left)\right)\×t\left(t\right)}{L^{\′}\left(t\right)+L^{\′}(t-n(t\left)\right)}$

Wherein, t " (t), t ^'(t+1) current road segment terminal border moment and downstream road section starting point border moment is represented respectively; T (t), t (t-n (t)) represent the location moment of current matching point and last match point data respectively; L ' (t), L ' (t-n (t)) represent the distance on current matching point and last match point data and current road segment terminal border respectively;

Then bicycle basic road running time computing formula is

T′＝t″-t′

Wherein: T ' is single sample car basic road running time; T " and t ' is respectively the moment of sample car through start and end border, section;

S02312: sample car basic road running time calculates,

Sample car basic road running time refers to the average level through all bicycle basic road running times of specific road section in special time yardstick, directly can get the average of bicycle basic road running time, namely

$\hat{T}={\mathrm{\Σ}}_{j=1}^n{T}_j^{\′}$

Wherein, for sample car basic road running time;

S02313: traffic flow basic road running time calculates,

Traffic flow basic road running time refers to the average level through all vehicle basic road running time times of specific road section in special time yardstick, and employing statistical study computing formula is

$T\left(l_1\right)=f\left(\hat{T}\right)$

Wherein: T (l ₁) be traffic flow basic road running time; The regression function of f for utilizing regretional analysis to set up;

S0232: design intersection delay Time Calculation model;

Described step S0232 comprises the following steps,

S02321: intersection delay Time Calculation,

Robert Webster formula is utilized to estimate intersection delay, namely

$\stackrel{\‾}{d}=\frac{C{(1-u)}^2}{2(1-ux)}+\frac{x^2}{2q(1-x)}-0.65{\left(\frac{C}{q^2}\right)}^{\frac{1}{3}}\·x^{(2+5u)}$

Wherein, for the intersection delay time, C is cycle duration; G is long green light time; Q is the magnitude of traffic flow; U is split, i.e. g/C; X is saturation degree, i.e. (q × C)/(s × g); S is saturation volume rate.

S0233: Link Travel Time calculates; Link Travel Time is basic road running time and intersection delay time sum, namely

$T=T\left(l_1\right)+\stackrel{\‾}{d}.$

8. the management control method of Link Travel Time according to claim 6, is characterized in that: described step S03 comprises the following steps,

S031: build three layers of BP neural network containing a hidden layer;

S032: using described first Output rusults of SCOOT class traffic signal control system, global position system second Output rusults and sample car quantity as input;

S033: using the Link Travel Time of same space and time scale as output.

9. the management control method of Link Travel Time according to claim 8, is characterized in that:

The principle of hidden nodes foundation (2 × input neuron number-1) of described BP neural network is set to 5; Training function adopts Sigmiod function; Training error adopts root-mean-square error RMSE.