CN106530698A - Estimation method for city road travel time taking exit turning into consideration - Google Patents

Abstract

The invention discloses an estimation method for city road travel time taking exit turning into consideration. The method comprises the steps that 1) an accumulated histogram correction model is established; 2) the accumulated upstream vehicle amounts in different exit turnings of vehicles are estimated; and 3) and travel time of a specific exit turning of a road is estimated by utilizing the estimated accumulated upstream vehicle amounts in different exit turnings of vehicles and accumulated downstream amount in different exit turnings of vehicles. According to the invention, data collected by a fixed-point detector and a detection vehicle is combined with a traditional accumulated histogram model, the accumulated histogram model taking travel time of different exit turnings is reestablished, existence of relative deviation in city roads is considered, the estimation precision for travel time of city roads is improved, and an error of the estimated value relative to an actually measured value of the detection vehicle is lower than 10%.

Description

A kind of method of estimation of the Urban Travel Time for considering outlet turning

Technical field

The invention belongs to intelligent transportation field, and in particular to a kind of Urban Travel Time of consideration outlet turning is estimated Meter method.

Background technology

Journey time is one of major parameter of intelligent transportation system, is to weigh road grid traffic congestion, go out line efficiency Important indicator.In city road network, journey time refer to from section point upstream to point downstream the time required to summation.Due to city The control of intersection signal, the non-conservation of traffic flow itself, cause journey time difference between different outlet turnings very big, Accurately estimate that city road network journey time still suffers from challenging.The method that the prediction of journey time is generally adopted includes time sequence Row method, artificial nerve network model, Kalman filtering etc., although these methods all have certain accuracy, existence foundation How computationally intensive data are solves the problem being difficult.

The content of the invention

For solving the above problems, the present invention provides a kind of estimation side of the Urban Travel Time for considering outlet turning Method, the method include：

Step one, set up accumulative histogram correction model；

Step 2, the accumulative vehicle number in upstream to vehicle difference outlet turning are estimated；

Step 3, the upstream accumulative vehicle number outlet turning different with vehicle for utilizing the vehicle difference outlet turning for estimating The accumulative vehicle number in downstream estimate the average travel time of the concrete outlet turning in section.

Further, the accumulative histogram correction model is：

In formula：U_AT () is the accumulative vehicle number in upstream section；U_mT () is the accumulative vehicle in upstream of the different outlet turning of vehicle Number；ε is error parameter；The different outlet turning of vehicle is represented with alphabetical m, m represents T, L, R, and wherein T represents straight trip, and L represents left Turn, R represents right-hand rotation.

Further, the accumulative vehicle number evaluation method in upstream to vehicle difference outlet turning is：

(1) calculate U_A(t) and D_m(t)；

The U_A(t) and D_mT () is obtained by being laid in the detector data at stop line, wherein, D_mT () represents downstream Section adds up vehicle number；

(2) vertically scale U_AT (), estimates U_m(t)；

Defined variable V_m,pFor effective scale factor, in representing one day in p time intervals m outlet turnings scale factor； t_s,pAnd t_e,pStart time and end of time (this paper V of p time interval vehicle travels is represented respectively_m,pValue can be according to road The flow proportional of each outlet turning of section arranges the corresponding scale factor value of each outlet turning, if V_m,p=1, then U_m(t)=U_A (t))；

In the initial time t of p time intervals_s,pInitial build vehicle number U to upstream m outlet turnings_m(t_s,p) estimated Meter, then using formula (2) is to each time interval U_AT () carries out vertically scale with the corresponding proportion factor, draw U_mThe estimation of (t) Value；

In formula：Y (t) is t_s,pTo the vehicle number that t time intervals m are turned to；U_A(t)-U_A(t_s,p) it is t_s,pTo in t time intervals Accumulation vehicle number；U_mT () carries out the accumulative vehicle number of m steerings for t；

(3) using detection car data, calculate U_m′(t)；

It is in order that the data definition that probe vehicles are collected is by which in crossing offer time labelling to introduce probe vehicles D′_mT (), now definition have n probe vehicles, unloading lot detection is respectively t in the time point of upstream and downstream_uAnd t_d, to U_m(t) curve The point for being passed through is defined, and step is as follows：

A, by t_dData list be ranked up according to the ascending order (sequencing for reaching) of data value；

B, by t_uData list be ranked up according to the ascending order (sequencing for setting out) of data value；

C, then by U_mT point that () passes through is defined as (t_uj,D′(t_dj)), wherein t_ujAnd t_djIt is t respectively_u、t_dIn data list The data of jth probe vehicles；

On road vehicle flowrate it is less in freestream conditionses when, accumulation histogram data are carried out into initialization process now, Define initial reference point P₀(U_m(t₀)=D '_m(t₀)=0), choose road on traffic flow from freely flow gradually to walk or drive slowly state mistake The moment crossed is used as t_r, define P_r(t_r,U_m(t_r)) for a reference point in accumulation histogram, for other points in rectangular histogram P_j, with former point P_j-1Recalculated as reference, then recalculated the U for obtaining_mPoint (the t passed through by ' (t)_p,Y_p) meter Calculate formula as follows：

U_m' (t)=U_m(t)+C (3)

Wherein,

In formula：U_m' (t) is to recalculate U_mThe upstream accumulation vehicle number of t m outlet turnings that () obtains；C is amendment variable； S is zoom factor.

Further, the scale factor V_m,p, it is by the U of the previous day each Fixed Time Interval_m' (t) and U_AT () carries out whole Close, for being predicted to future time instance road net traffic state, V_m,pComputing formula it is as follows：

Wherein：

In formula：t_e,p、t_s,pRespectively p time intervals beginning and end moment；Y_A,p、Y_m,pRespectively U_A(t) and U_m' (t) is in p Between interval in stored counts.

Further, the accumulative vehicle number in upstream using the vehicle for estimating difference outlet turning it is different with vehicle go out The accumulative vehicle number in downstream that mouth is turned to estimates that the average travel time method of the concrete outlet turning in section is：By U_m' (t) and detection Device data D_mT () is substituted in formula (8), calculate the average travel time of the concrete outlet turning in section

Wherein：

N=U '_m(t₂)-U′_m(t₁)=D_m(t₄)-D_m(t₃) (9)

In formula：U_m' (t) is the accumulative vehicle number of the upstream m outlet turnings for calculating again；S is whole journey times； N is time t₁-t₂(t₃-t₄) the interior accumulation vehicle flowrate for reaching point upstream (leaving point downstream)；

Beneficial effects of the present invention：By data and the tradition accumulation histogram mould of detector and probe vehicles collection will be pinpointed Type combines, and re-establishes the journey time accumulation histogram model of the different outlet turnings of consideration and carries out instance analysis, newly-built The estimated value of accumulation histogram model and the error of probe vehicles measured value are 10% or so, and just only need to a small amount of traffic flow data The travel time data of different outlet turnings is can be derived that, can simply and effectively judge crossing all directions whether in traffic Saturation, formulation to subsequent control strategy and prevents the aggravation of traffic congestion with higher using value.

Description of the drawings

Fig. 1 is vertically scale U of the present invention_A(t) schematic diagram；

Fig. 2 is accumulation histogram correction model Establishing process figure of the present invention；

Fig. 3 calculates U for the present invention using detection car data_m' (t) schematic diagrams；

Fig. 4 is that average travel time of the present invention calculates schematic diagram；

Fig. 5 is embodiment of the present invention survey region figure；

Left-hand rotation travel-time maps of the Fig. 6 for the embodiment of the present invention；

Left-hand rotation Travel Time Error analysis charts of the Fig. 7 for the embodiment of the present invention；

Straight trip travel-time maps of the Fig. 8 for the embodiment of the present invention；

Straight trip Travel Time Error analysis charts of the Fig. 9 for the embodiment of the present invention；

Right-hand rotation travel-time maps of the Figure 10 for the embodiment of the present invention；

Right-hand rotation Travel Time Error analysis charts of the Figure 11 for the embodiment of the present invention；

Figure 12 is method accuracy rate comparison diagram in the case of the identical traffic flow of the embodiment of the present invention；

Specific embodiment

The case that is preferable to carry out of the present invention is illustrated below in conjunction with accompanying drawing, it will be appreciated that described herein preferred Case study on implementation is only used for explaining and illustrating the present invention program, is not limited to the present invention.

Specific embodiment：Refer to Fig. 1,2,3, a kind of estimation side of the Urban Travel Time for considering outlet turning Method, the method include：

Step one, set up accumulative histogram correction model：

In formula：U_AT () is the accumulative vehicle number in upstream section；U_mT () is the accumulative vehicle in upstream of the different outlet turning of vehicle Number；ε is error parameter；The different outlet turning of vehicle is represented with alphabetical m, m represents T, L, R, and wherein T represents straight trip, and L represents left Turn, R represents right-hand rotation.

Step 2, the accumulative vehicle number in upstream to vehicle difference outlet turning estimate that the evaluation method is：

(1) calculate U_A(t) and D_m(t)；

The U_A(t) and D_mT () is obtained by being laid in the detector data at stop line, wherein, D_mT () represents downstream Section adds up vehicle number；

(2) vertically scale U_AT (), estimates U_mT (), refers to Fig. 1；

Defined variable V_m,pFor effective scale factor, in representing one day in p time intervals m outlet turnings scale factor； t_s,pAnd t_e,pStart time and end of time (this paper V of p time interval vehicle travels is represented respectively_m,pValue can be according to road The flow proportional of each outlet turning of section arranges the corresponding scale factor value of each outlet turning, if V_m,p=1, then U_m(t)=U_A (t))；

In the initial time t of p time intervals_s,pInitial build vehicle number U to upstream m outlet turnings_m(t_s,p) estimated Meter, then using formula (2) is to each time interval U_AT () carries out vertically scale with the corresponding proportion factor, draw U_mThe estimation of (t) Value；

In formula：Y (t) is t_s,pTo the vehicle number that t time intervals m are turned to；U_A(t)-U_A(t_s,p) it is t_s,pTo in t time intervals Accumulation vehicle number；U_mT () carries out the accumulative vehicle number of m steerings for t；

(3) using detection car data, calculate U_m' (t), refers to Fig. 2, Fig. 3；

It is in order that the data definition that probe vehicles are collected is by which in crossing offer time labelling to introduce probe vehicles D′_mT (), now definition have n probe vehicles, unloading lot detection is respectively t in the time point of upstream and downstream_uAnd t_d, to U_m(t) curve The point for being passed through is defined, and step is as follows：

A, by t_dData list be ranked up according to the ascending order (sequencing for reaching) of data value；

B, by t_uData list be ranked up according to the ascending order (sequencing for setting out) of data value；

C, then by U_mT point that () passes through is defined as (t_uj,D′(t_dj)), wherein t_ujAnd t_djIt is t respectively_u、t_dIn data list The data of jth probe vehicles；

On road vehicle flowrate it is less in freestream conditionses when, accumulation histogram data are carried out into initialization process now, Define initial reference point P₀(U_m(t₀)=D '_m(t₀)=0), choose road on traffic flow from freely flow gradually to walk or drive slowly state mistake The moment crossed is used as t_r, define P_r(t_r,U_m(t_r)) for a reference point in accumulation histogram, for other points in rectangular histogram P_j, with former point P_j-1Recalculated as reference, then recalculated the U for obtaining_mPoint (the t passed through by ' (t)_p,Y_p) meter Calculate formula as follows：

U_m' (t)=U_m(t)+C (3)

Wherein,

In formula：U_m' (t) is to recalculate U_mThe upstream accumulation vehicle number of t m outlet turnings that () obtains；C is amendment variable； S is zoom factor.

As shown in figure 3, in time t_rOriginal rectangular histogram is not zoomed in and out before, in t_r-t_pTime interval is entered to rectangular histogram Row vertically scale, more than t_pAfterwards due to variable C being corrected for definite value, vertical transition is carried out to which, so that U_m' (t) is continuous , and in moment t_pParallel to original U_m(t).U is utilized finally_m' (t) and D_mT () calculates the city road row for considering outlet turning Journey time average journey time.

Further, the scale factor V_m,p, it is by the U of the previous day each Fixed Time Interval_m' (t) and U_AT () carries out whole Close, for being predicted to future time instance road net traffic state, V_m,pComputing formula it is as follows：

Wherein：

In formula：t_e,p、t_s,pRespectively p time intervals beginning and end moment；Y_A,p、Y_m,pRespectively U_A(t) and U_m' (t) exists Stored counts in p time intervals.

Fig. 4 is referred to, the accumulative vehicle number in upstream and vehicle of the vehicle difference outlet turning that step 3, utilization are estimated are not Add up the average travel time that vehicle number estimates the concrete outlet turning in section with the downstream of outlet turning：By U_m' (t) and detector Data D_mT () is substituted in formula (8), calculate the average travel time of the concrete outlet turning in section

Wherein：

N=U '_m(t₂)-U′_m(t₁)=D_m(t₄)-D_m(t₃) (9)

In formula：U_m' (t) is the accumulative vehicle number of the upstream m outlet turnings for calculating again；S is whole journey times； N is time t₁-t₂(t₃-t₄) the interior accumulation vehicle flowrate for reaching point upstream (leaving point downstream)；

It should be noted that for verifying feasibility of the present invention, applicant is with Qingdao HongKong road from rivers and mountains South Road to Wuyi Mountain One section of road is Experimental Area, refers to Fig. 5.This is one section of major urban arterial highway from west to east, section total length 2.1km, two-way four Track, traffic detector are laid near stop line, from the number that on July 20th, 2015 is collected in above-mentioned interval by probe vehicles Experimental simulation is carried out according to neural network model and accumulation histogram correction model is substituted into, is simulated from S point to L respectively, 3 points of T, R enters The vehicle that row turns left, keeps straight on, turns right.The time data that probe vehicles are gathered is defined as into traveled distance time t_a, using above-mentioned two Outlet turning journey time obtained by planting model emulation is defined as predicted travel time t_e, evaluation index is as follows：

A_M=(1-MAPE) (12)

In formula：APE is journey time absolute percentage error；MAEP is journey time mean percentage error；A_MFor row Journey time Estimate effectiveness percentage rate.

Section is carried out turning left, kept straight on above two model using above-mentioned formula, being turned right three directional run times Error analyses.From Fig. 6, Fig. 8, Figure 10 Forecasting of Travel Time result can be seen that accumulation histogram curve generally with actual row The compatible degree of journey time graph is higher, and journey time can be estimated；Knowable to the error analyses of Fig. 7, Fig. 9, Figure 11, fortune When being predicted with accumulation histogram method, error is generally controlled within 10%, and relative neural network model For accuracy it is slightly higher, accurate journey time can be predicted.

In addition, in order to prove the method compared with neural network model with simple advantage, applicant's utilization public affairs Formula (10)-(12) are respectively to accumulation histogram model and neural network model in Different Traffic Flows amount in unit interval Percentage error is analyzed, and refers to Figure 12, can be found that from figure：

(1), when carrying out road travel time estimation using accumulation histogram model, estimated accuracy can reach 90% or so, And as the increase estimated accuracy of traffic flow data can increase.

(2) method presented here only need to detect on a small quantity car data can to make relatively accurate prediction, but it is refreshing Only when traffic flow is larger, accuracy of detection just can be relatively accurate for Jing network modeies.So using accumulation histogram model prediction Journey time is more simple.

Finally it should be noted that：The better embodiment of the present invention is the foregoing is only, is not intended to limit the present invention, it is all In simple any modification, equivalent, modification made according to the technical spirit of the present invention etc., all should be included in the present invention's Within protection domain.

Claims (5)

1. it is a kind of consider outlet turning Urban Travel Time method of estimation, it is characterised in that the method includes：

Step one, set up accumulative histogram correction model；

Step 2, the accumulative vehicle number in upstream to vehicle difference outlet turning are estimated；

Under step 3, the upstream accumulative vehicle number outlet turning different with vehicle using the vehicle difference outlet turning for estimating The accumulative vehicle number of trip estimates the average travel time of the concrete outlet turning in section.

2. it is according to claim 1 consider outlet turning Urban Travel Time method of estimation, it is characterised in that The accumulative histogram correction model is：

$U_A\left(t\right)=\underset{\∀m}{\Σ}{U}_m\left(t\right)+\mathrm{\ϵ}---\left(1\right)$

In formula：U_AT () is the accumulative vehicle number in upstream section；U_mT () is the accumulative vehicle number in upstream of the different outlet turning of vehicle；ε For error parameter；The different outlet turning of vehicle is represented with alphabetical m, m represents T, L, R, and wherein T represents straight trip, and L represents left-hand rotation, R Represent and turn right.

3. it is according to claim 1 consider outlet turning Urban Travel Time method of estimation, it is characterised in that The upstream to vehicle difference outlet turning adds up vehicle number evaluation method and is：

(1) calculate U_A(t) and D_m(t)；

The U_A(t) and D_mT () is obtained by being laid in the detector data at stop line, wherein, D_mT () represents downstream road section Accumulative vehicle number；

(2) vertically scale U_AT (), estimates U_m(t)；

Defined variable V_m,pFor effective scale factor, in representing one day in p time intervals m outlet turnings scale factor；t_s,pWith t_e,pStart time and end of time (this paper V of p time interval vehicle travels is represented respectively_m,pValue can be each according to section The flow proportional of outlet turning arranges the corresponding scale factor value of each outlet turning, if V_m,p=1, then U_m(t)=U_A(t))；

In the initial time t of p time intervals_s,pInitial build vehicle number U to upstream m outlet turnings_m(t_s,p) estimated, so Using formula (2) is to each time interval U afterwards_AT () carries out vertically scale with the corresponding proportion factor, draw U_mThe estimated value of (t)；

In formula：Y (t) is t_s,pTo the vehicle number that t time intervals m are turned to；U_A(t)-U_A(t_s,p) it is t_s,pAccumulate in t time intervals Vehicle number；U_mT () carries out the accumulative vehicle number of m steerings for t；

(3) using detection car data, calculate U_m′(t)；

Introduce probe vehicles be in order that its crossing provide time labelling, the data definition that probe vehicles are collected be D '_m(t), Now definition has n probe vehicles, and unloading lot detection is respectively t in the time point of upstream and downstream_uAnd t_d, to U_mWhat t () curve was passed through Point is defined, and step is as follows：

A, by t_dData list be ranked up according to the ascending order (sequencing for reaching) of data value；

B, by t_uData list be ranked up according to the ascending order (sequencing for setting out) of data value；

C, then by U_mT point that () passes through is defined as (t_uj,D′(t_dj)), wherein t_ujAnd t_djIt is t respectively_u、t_dJth in data list The data of probe vehicles；

On road vehicle flowrate it is less in freestream conditionses when, accumulation histogram data are carried out into initialization process now, define Initial reference point P₀(U_m(t₀)=D '_m(t₀)=0), chooses on road traffic flow from freely flowing gradually to status transition of walking or drive slowly Moment is used as t_r, define P_r(t_r,U_m(t_r)) for a reference point in accumulation histogram, for other points P in rectangular histogram_j, with Former point P_j-1Recalculated as reference, then recalculated the U for obtaining_mPoint (the t passed through by ' (t)_p,Y_p) calculating it is public Formula is as follows：

U_m' (t)=U_m(t)+C (3)

Wherein,

$s=\left\{\begin{array}{cc}\frac{Y_p-U_m\left(t_r\right)}{U_m\left(t_p\right)-U_m\left(t_r\right)}& U_m\left(t_p\right)\≠U_m\left(t_r\right)\\ 1& U_m\left(t_p\right)=U_m\left(t_r\right)\end{array}\right.---\left(5\right)$

In formula：U_m' (t) is to recalculate U_mThe upstream accumulation vehicle number of t m outlet turnings that () obtains；C is amendment variable；S is Zoom factor.

4. it is according to claim 3 consider outlet turning Urban Travel Time method of estimation, it is characterised in that The scale factor V_m,p, it is by the U of the previous day each Fixed Time Interval_m' (t) and U_AT () is integrated, for future time instance Road net traffic state is predicted, V_m,pComputing formula it is as follows：

$V_{m,p}=\frac{Y_{A,p}-Y_{m,p}}{Y_{A,p}}---\left(6\right)$

Wherein：

In formula：t_e,p、t_s,pRespectively p time intervals beginning and end moment；Y_A,p、Y_m,pRespectively U_A(t) and U_m' (t) is in p Between interval in stored counts.

5. it is according to claim 1 consider outlet turning Urban Travel Time method of estimation, it is characterised in that The downstream of the upstream accumulative vehicle number outlet turning different with vehicle using the vehicle for estimating difference outlet turning adds up Vehicle number estimates that the average travel time method of the concrete outlet turning in section is：By U_m' (t) and detector data D_mT () substitutes into public In formula (8), the average travel time of the concrete outlet turning in section is calculated

$\stackrel{\‾}{T_m}=\frac{S}{N}=\frac{\underset{i=1}{\overset{N}{\Σ}}\[{D_m}^{-1}\left(i\right)-{\left(U_m^{\′}\right)}^{-1}\left(i\right)\]}{N}=\frac{\underset{i=1}{\overset{N}{\Σ}}{D_m}^{-1}\left(i\right)-\underset{i=1}{\overset{N}{\Σ}}{\left(U_m^{\′}\right)}^{-1}\left(i\right)}{N}---\left(8\right)$

Wherein：

N=U '_m(t₂)-U′_m(t₁)=D_m(t₄)-D_m(t₃) (9)

In formula：U_m' (t) is the accumulative vehicle number of the upstream m outlet turnings for calculating again；S is whole journey times；When N is Between t₁-t₂(t₃-t₄) the interior accumulation vehicle flowrate for reaching point upstream (leaving point downstream).