CN104658253A - Highway traffic state identification method - Google Patents

Abstract

The invention provides a method capable of estimating traffic states of a highway, and aims to solve the problems that conventional highway traffic loads are increased day by day and traffic capacity and service level of a road are remarkably reduced. According to a basic idea, flow and speed data are detected by a detector, three parameters of traffic around the detector are estimated firstly, weighting estimation of different zones are obtained according to the position relation of a display unit and a detector, weighting is performed according to the traffic parameters of the zones, traffic states of the display unit are obtained, and finally, the traffic states are graded through a threshold. The efficient highway traffic state identification method is provided, and limited road network space-time resources are exerted to the greatest extent.

Description

A kind of traffic status of express way method of discrimination

Technical field

The present invention relates to a kind of traffic status of express way method of discrimination, for freeway traffic regulation and control, belong to intelligent transportation research field.

Background technology

Traffic status of express way differentiation refers to the traffic flow parameter utilizing highway main line section detector and gateway turning branch detection device to obtain, set up the computing method characterizing freeway traffic flow running status leading indicator, and then classification differentiation is carried out to freeway traffic flow running status.Traffic status of express way be rationally determined as urban traffic control person, the decision-making of traveler provides support technology and reference information, for traffic administration person understands the information of road net traffic state change in real time, take some countermeasures in time relieve traffic congestion, alleviate crowded, promote traffic safety, the efficiency playing road network time-space distribution maximum lays the foundation.

Existing traffic state estimation method has following shortcoming: the space cell three that the space section that (1) detecting device detects, the space cell of state estimation, traffic behavior are issued is inconsistent, thus carries out state estimation and issue all not making full use of sensor information; (2) the randomness impact of traffic behavior is not considered, thus the information stability issued has much room for improvement; (3) correlation technique is too complicated, is difficult to apply in practice.

Based on this, in order to judge freeway traffic operational efficiency and the degree of crowding more exactly, need to set up a kind of traffic status of express way method of discrimination efficiently.

Summary of the invention

The object of the present invention is to provide a kind of traffic status of express way method of discrimination.The basic thought of the method is the detector acquisition traffic flow parameter laid by road network, after transport data processing, and then sets up road network traffic flow running status index calculating method.For achieving the above object, the traffic status of express way method of discrimination that the present invention proposes comprises: the estimation of detector data process, traffic behavior, state classification and differentiation.

Basic step of the present invention is as follows:

C1, detector data process;

The estimation of c2, traffic status of express way;

C3, state classification and differentiation flow process.

Step c1 is specifically:

The acquisition and processing of detector data.

Each main line section detector needs the average velocity and the flow that at least record traffic flow in each sampling interval in each track.Need to carry out corresponding pre-service to data simultaneously, ensure specification, the standard of mode input data.Suggestion, when highway data acquisition, adopts 1 minute as basic data sampling interval.Meanwhile, can carry out merging and the adjustment of data according to actual needs, the issue interval of traffic behavior can be the multiple of 1 minute.Data prediction mainly comprises flow and speed data.

Master data mainly comprises two classes: in the n-th basic display unit, i-th track, be every lane detector flow q (n, i, the k of kth 1 sampling interval under sampling interval with 1 minute ₁) in (unit: veh/1 min) and the n-th basic display unit, i-th track, be car speed mean value v (n, i, the k of kth 1 sampling interval under sampling interval with 1 minute ₁) (unit: km/h).

The process of step c2 comprises:

C21, traffic status of express way variable and computing method thereof

According to the basic condition of China's highway Loop detector layout, determine that the elementary cell that traffic status of express way shows is 2km.The macroscopic description of traffic flow is needed to the variable of the average operation action of description traffic flow be defined under a certain special time period and ad-hoc location.Therefore, need to carry out discretize to Time and place.The discrete time interval can represent with T.The display unit of highway can be divided into the elementary cell of n 2 km length.Each elementary cell t=kT within each time interval can represent with corresponding Macro-traffic Flow basic parameter.

1. traffic flow density: ρ _n(k) (unit: veh/km/lane).Represent in the kT moment, in the n-th elementary cell, vehicle number is divided by the length Δ of elementary cell _mand number of track-lines λ _m.

2. average velocity: v _n(k) (unit: km/h).Represent in the kT moment, the average velocity of all vehicles in the n-th elementary cell.

3. the magnitude of traffic flow: q _n(k) (unit: veh/h).Represent within the time period [kT, (k+1) T], leave the vehicle number of the n-th elementary cell divided by time interval T.

Data acquisition intervals due to actual traffic stream is 1 minute, as required data is summed up the traffic flow data obtaining any sampling interval.Wherein, flow directly can be added summation, and speed needs to be weighted on average to obtain.

If q _d(n, i, k) represents in the n-th basic display unit, i-th track, under sampling interval duration is T minute situation, and the detecting device flow (unit: veh/T min) of a kth sampling interval.V _d(n, i, k) represents in the n-th basic display unit, i-th track, the average velocity (unit: km/h) of a kth sampling interval.

Then the computing formula of this section traffic flow hour flow (unit veh/h), section speed (unit km/h) and section periphery density (unit: veh/km/lane) is as follows respectively:

$q_d(n,k)=\frac{60\underset{i=1}{\overset{{\mathrm{\λ}}_n}{\mathrm{\Σ}}}{q}_d(n,i,k)}{T}$

$v_d(n,k)=\frac{\underset{i=1}{\overset{{\mathrm{\λ}}_n}{\mathrm{\Σ}}}{q}_d(n,i,k)v_d(n,i,k)}{\underset{i=1}{\overset{{\mathrm{\λ}}_n}{\mathrm{\Σ}}}{q}_d(n,i,k)}$

${\mathrm{\ρ}}_d(n,k)=\frac{q_d(n,k)}{v_d(n,k){\mathrm{\λ}}_n}$

In formula, λ _nfor number of track-lines.

Consider the wave characteristic of data, section detector three parameter needs the estimated value of carrying out corresponding the disposal of gentle filter and then being applied to traffic behavior.The smothing filtering model of three parameters is as follows:

${\hat{q}}_d(n,k)=\frac{{\mathrm{\β}}_1{q}_d(n,k-2)+{\mathrm{\β}}_2{q}_d(n,k-1)+{\mathrm{\β}}_3{q}_d(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

${\hat{v}}_d(n,k)=\frac{{\mathrm{\β}}_1{v}_d(n,k-2)+{\mathrm{\β}}_2{v}_d(n,k-1)+{\mathrm{\β}}_3{v}_d(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

${\widehat{\mathrm{\ρ}}}_d(n,k)=\frac{{\mathrm{\β}}_1{\mathrm{\ρ}}_d(n,k-2)+{\mathrm{\β}}_2{\mathrm{\ρ}}_d(n,k-1)+{\mathrm{\β}}_3{\mathrm{\ρ}}_d(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

Wherein, β ₁, β ₂, β ₃be respectively three level and smooth parameters, and β ₁+ β ₂+ β ₃=1.Three parameters can be arranged according to actual needs, consider β here ₁=0.2; β ₂=0.3; β ₃=0.5.

C22, method for estimating state

According to the laying situation of practical detector, the Loop detector layout situation of each basic display unit is mainly divided into following three kinds of situations: the first, does not lay detecting device; The second, lay one group of detecting device; 3rd, lay and organize detecting device more.For the section not laying detecting device, its state needs to be estimated by the state of former and later two sections, if it is too much not lay detecting device section, then cannot carry out the estimation of traffic behavior.For the interval of laying many group detecting devices, get one group as far as possible and consider that the detecting device in section centre position calculates, other detecting devices can as subsequent use or carry out the checking of state estimation.

Carried out the estimation of section traffic behavior by detector data, be mainly divided into two kinds of situations.The first: when basic display unit be first section or last section time, the traffic behavior parameter of its section just equals the traffic behavior parameter that this section internal detector calculates.The second: when all there is corresponding state before and after section, its estimation process is just comparatively complicated, as shown in Figure 2.

Wherein, basic display unit n first half speed and basic display unit n latter half speed are expressed as:

$v_1(n,k)=\frac{{\hat{q}}_d(n,k)}{{\mathrm{\λ}}_n[{\widehat{\mathrm{\ρ}}}_d(n-1,k)L_1+{\widehat{\mathrm{\ρ}}}_d(n,k)L_2]/(L_1+L_2)}$

$v_2(n,k)=\frac{{\hat{q}}_d(n,k)}{{\mathrm{\λ}}_n[{\widehat{\mathrm{\ρ}}}_d(n,k)L_3+{\widehat{\mathrm{\ρ}}}_d(n+1,k)L_4]/(L_3+L_4)}$

Then the travelling speed of basic display unit is:

$v(n,k)=\frac{L_2+L_3}{L_2/v_1(n,k)+L_3/v_2(n,k)}$

In like manner, need the smoothing process of travelling speed:

$\hat{v}(n,k)=\frac{{\mathrm{\β}}_1\hat{v}(n,k-2)+{\mathrm{\β}}_2\hat{v}(n,k-1)+{\mathrm{\β}}_3\hat{v}(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

Correlation parameter value can with reference to recommended value above.

Step c3 is specifically:

To state classification and the differentiation of traffic status of express way.According to the characteristic that freeway traffic flow runs, speed index generally can be adopted as the key variables weighing traffic circulation state.According to actual conditions, three grades or level Four traffic behavior grade scale can be set, see table 1.Concrete threshold speed needs to arrange in conjunction with the demand of real data and traffic administration person.

Table 1 traffic status of express way hierarchical table

Beneficial effect of the present invention: the invention has the beneficial effects as follows and make full use of detector data, the difference between the interval and Loop detector layout of balance display unit, state estimation.

Accompanying drawing explanation

Fig. 1 is the basic procedure that traffic status of express way differentiates;

Fig. 2 traffic status of express way differentiates section partition table;

Embodiment

Below in conjunction with accompanying drawing 1, invention specific implementation method is further illustrated.

The raw data that detecting device detects is according to mainly comprising two classes: the every track flow (q (n, i, the k that were sampling interval with 1 minute ₁)) and speed (v (n, i, k ₁)).Q _d1(n, i, k ₁) represent in the n-th basic display unit, i-th track, 1 minute to be the detecting device flow of kth 1 sampling interval under sampling interval.(unit: veh/1min) v _d1(n, i, k ₁) represent in the n-th basic display unit, i-th track, 1 minute to be the car speed mean value of kth 1 sampling interval under sampling interval.(unit: km/h)

(1) section traffic three parameter estimation

After obtaining raw data, first will estimate the Macro-traffic Flow parameter of section, be also traffic three parameter: (i) traffic flow density: ρ _n(k) (unit: veh/km/lane).Represent in the kT moment, in the n-th elementary cell, vehicle number is divided by the length Δ of elementary cell _mand number of track-lines λ _m; (ii) average velocity: v _n(k) (unit: km/h).Represent in the kT moment, the average velocity of all vehicles in the n-th elementary cell; (iii) magnitude of traffic flow: q _n(k) (unit: veh/h).Represent within the time period [kT, (k+1) T], leave the vehicle number of the n-th elementary cell divided by time interval T.

If q _d(n, i, k) represents in the n-th basic display unit, i-th track, under sampling interval duration is T minute situation, and the detecting device flow (unit: veh/T min) of a kth sampling interval.V _d(n, i, k) represents in the n-th basic display unit, i-th track, the average velocity (unit: km/h) of a kth sampling interval.

Then this section traffic flow hour flow (unit veh/h) is:

$q_d(n,k)=\frac{60\underset{i=1}{\overset{{\mathrm{\λ}}_n}{\mathrm{\Σ}}}{q}_d(n,i,k)}{T}$

Wherein, λ _nfor number of track-lines.

Section speed (unit km/h) is:

$v_d(n,k)=\frac{\underset{i=1}{\overset{{\mathrm{\λ}}_n}{\mathrm{\Σ}}}{q}_d(n,i,k)v_d(n,i,k)}{\underset{i=1}{\overset{{\mathrm{\λ}}_n}{\mathrm{\Σ}}}{q}_d(n,i,k)}$

Section periphery density (unit: veh/km/lane) is:

${\mathrm{\ρ}}_d(n,k)=\frac{q_d(n,k)}{v_d(n,k){\mathrm{\λ}}_n}$

(2) traffic three parameter is level and smooth

Consider the wave characteristic of data, section detector three parameter needs the estimation carried out corresponding the disposal of gentle filter and then be applied to traffic behavior.The smothing filtering model of three parameters is as follows:

${\hat{q}}_d(n,k)=\frac{{\mathrm{\β}}_1{q}_d(n,k-2)+{\mathrm{\β}}_2{q}_d(n,k-1)+{\mathrm{\β}}_3{q}_d(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

${\hat{v}}_d(n,k)=\frac{{\mathrm{\β}}_1{v}_d(n,k-2)+{\mathrm{\β}}_2{v}_d(n,k-1)+{\mathrm{\β}}_3{v}_d(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

${\widehat{\mathrm{\ρ}}}_d(n,k)=\frac{{\mathrm{\β}}_1{\mathrm{\ρ}}_d(n,k-2)+{\mathrm{\β}}_2{\mathrm{\ρ}}_d(n,k-1)+{\mathrm{\β}}_3{\mathrm{\ρ}}_d(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

Wherein, β ₁, β ₂, β ₃be respectively three level and smooth parameters, and β ₁+ β ₂+ β ₃=1.Three parameters can be arranged according to actual needs, consider β here ₁=0.2; β ₂=0.3; β ₃=0.5.

(3) estimation of the traffic parameter of basic display unit

According to the laying situation of practical detector, the Loop detector layout situation of each basic display unit is mainly divided into following three kinds of situations: the first, does not lay detecting device; The second, lay one group of detecting device; 3rd, lay and organize detecting device more.For the section not laying detecting device, its state needs to be estimated by the state of former and later two sections, if it is too much not lay detecting device section, then cannot carry out the estimation of traffic behavior.For the interval of laying many group detecting devices, get one group as far as possible and consider that the detecting device in section centre position calculates, other detecting devices can as subsequent use or carry out the checking of state estimation.

Carried out the estimation of section traffic behavior by detector data, be mainly divided into two kinds of situations.The first: when basic display unit be first section or last section time, the traffic behavior parameter of its section just equals the traffic behavior parameter that this section internal detector calculates.The second: when all there is corresponding state before and after section, its estimation process is just comparatively complicated, as shown in Figure 2.

Wherein, basic display unit n first half speed can be expressed as:

$v_1(n,k)=\frac{{\hat{q}}_d(n,k)}{{\mathrm{\λ}}_n[{\widehat{\mathrm{\ρ}}}_d(n-1,k)L_1+{\widehat{\mathrm{\ρ}}}_d(n,k)L_2]/(L_1+L_2)}$

Wherein, basic display unit n latter half speed can be expressed as:

$v_2(n,k)=\frac{{\hat{q}}_d(n,k)}{{\mathrm{\λ}}_n[{\widehat{\mathrm{\ρ}}}_d(n,k)L_3+{\widehat{\mathrm{\ρ}}}_d(n+1,k)L_4]/(L_3+L_4)}$

Then the travelling speed of basic display unit is:

$v(n,k)=\frac{L_2+L_3}{L_2/v_1(n,k)+L_3/v_2(n,k)}$

In like manner, need the smoothing process of travelling speed:

$\hat{v}(n,k)=\frac{{\mathrm{\β}}_1\hat{v}(n,k-2)+{\mathrm{\β}}_2\hat{v}(n,k-1)+{\mathrm{\β}}_3\hat{v}(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

Correlation parameter value can with reference to recommended value above.

(4) state classification and differentiation

According to the characteristic that freeway traffic flow runs, speed index generally can be adopted as the key variables weighing traffic circulation state.According to actual conditions, three grades or level Four traffic behavior grade scale can be set, see table 1.Concrete threshold speed needs to arrange in conjunction with the demand of real data and traffic administration person.

Table 1 traffic status of express way hierarchical table

Claims (1)

1. a traffic status of express way method of discrimination, is characterized in that the method comprises the following steps:

C1, detector data process; Specifically: raw data that detecting device detects is according to mainly comprising two classes: with to be sampling interval every in 1 minute track flow q (n, i, k ₁) and speed v (n, i, k ₁); q _d1(n, i, k ₁) represent in the n-th basic display unit, i-th track, 1 minute to be the detecting device flow of kth 1 sampling interval under sampling interval; v _d1(n, i, k ₁) represent in the n-th basic display unit, i-th track, 1 minute to be the car speed mean value of kth 1 sampling interval under sampling interval;

The estimation of c2, traffic status of express way; Specifically:

The estimation of c21, section traffic three parameter;

If q _d(n, i, k) represents in the n-th basic display unit, i-th track, under sampling interval duration is T minute situation, and the detecting device flow of a kth sampling interval; v _d(n, i, k) represents in the n-th basic display unit, i-th track, the average velocity of a kth sampling interval;

Then this section traffic flow hour flow is:

$q_d(n,k)=\frac{60\underset{i=1}{\overset{{\mathrm{\λ}}_n}{\mathrm{\Σ}}}{q}_d(n,i,k)}{T}$

Wherein, λ _nfor number of track-lines;

Section speed is:

$v_d(n,k)=\frac{\underset{i=1}{\overset{{\mathrm{\λ}}_n}{\mathrm{\Σ}}}{q}_d(n,i,k)v_d(n,i,k)}{\underset{i=1}{\overset{{\mathrm{\λ}}_n}{\mathrm{\Σ}}}{q}_d(n,i,k)}$

Section periphery density is:

${\mathrm{\ρ}}_d(n,k)=\frac{q_d(n,k)}{v_d(n,k){\mathrm{\λ}}_n}$

C22, traffic three parameter level and smooth;

Consider the wave characteristic of data, section detector three parameter needs the estimation carried out corresponding the disposal of gentle filter and then be applied to traffic behavior; The smothing filtering model of three parameters is as follows:

${\hat{q}}_d(n,k)=\frac{{\mathrm{\β}}_1{q}_d(n,k-2)+{\mathrm{\β}}_2{q}_d(n,k-1)+{\mathrm{\β}}_3{q}_d(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

${\hat{v}}_d(n,k)=\frac{{\mathrm{\β}}_1{v}_d(n,k-2)+{\mathrm{\β}}_2{v}_d(n,k-1)+{\mathrm{\β}}_3{v}_d(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

${\widehat{\mathrm{\ρ}}}_d(n,k)=\frac{{\mathrm{\β}}_1{\mathrm{\ρ}}_d(n,k-2)+{\mathrm{\β}}_2{\mathrm{\ρ}}_d(n,k-1)+{\mathrm{\β}}_3{\mathrm{\ρ}}_d(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

Wherein, β ₁, β ₂, β ₃be respectively three level and smooth parameters, and β ₁+ β ₂+ β ₃=1; Three parameters can be arranged according to actual needs, consider β here ₁=0.2; β ₂=0.3; β ₃=0.5; ;

C23, state estimation;

Carried out the estimation of section traffic behavior by detector data, be mainly divided into two kinds of situations; The first: when basic display unit be first section or last section time, the traffic behavior parameter of its section just equals the traffic behavior parameter that this section internal detector calculates; The second: when all there is corresponding state before and after section, carry out as follows:

Basic display unit n first half speed and basic display unit n latter half speed are expressed as:

$v_1(n,k)=\frac{{\hat{q}}_d(n,k)}{{\mathrm{\λ}}_n[{\widehat{\mathrm{\ρ}}}_d(n-1,k)L_1+{\widehat{\mathrm{\ρ}}}_d(n,k)L_2]/(L_1+L_2)}$

$v_2(n,k)=\frac{{\hat{q}}_d(n,k)}{{\mathrm{\λ}}_n[{\widehat{\mathrm{\ρ}}}_d(n,k)L_3+{\widehat{\mathrm{\ρ}}}_d(n+1,k)L_4]/(L_3+L_4)}$

Then the travelling speed of basic display unit is:

$v(n,k)=\frac{L_2+L_3}{L_2/v_1(n,k)+L_3/v_2(n,k)}$

In like manner, need the smoothing process of travelling speed:

$\hat{v}(n,k)=\frac{{\mathrm{\β}}_1\hat{v}(n,k-2)+{\mathrm{\β}}_2\hat{v}(n,k-1)+{\mathrm{\β}}_3\hat{v}(n,k)}{{\mathrm{\β}}_1+{\mathrm{\β}}_2+{\mathrm{\β}}_3}$

C3, state classification and differentiation flow process; Specifically: the characteristic run according to freeway traffic flow, adopt speed index as the key variables weighing traffic circulation state; According to actual conditions, three grades or level Four traffic behavior grade scale are set; Concrete threshold speed is arranged in conjunction with the demand of real data and traffic administration person;