CN108269411B

CN108269411B - Expressway ETC traffic flow prediction method

Info

Publication number: CN108269411B
Application number: CN201611266475.0A
Authority: CN
Inventors: 张重阳; 陆建峰; 杨静宇
Original assignee: Nanjing University of Science and Technology
Current assignee: Nanjing University of Science and Technology
Priority date: 2016-12-31
Filing date: 2016-12-31
Publication date: 2020-12-25
Anticipated expiration: 2036-12-31
Also published as: CN108269411A

Abstract

The invention provides a method for predicting the ETC traffic flow on a highway. Detecting a city round-trip annular route in a traffic record according to historical data of vehicle traveling in an ETC system, selecting a prediction reference sample according to the characteristic that the traffic has correlation in the same time period, and predicting the traffic in a specified time interval according to the real-time traffic and the prediction reference sample; when the ETC vehicle flow is predicted, the flow of a similar reference sample is considered, the difference between the reference sample and the predicted flow is also considered, and the predicted difference is corrected through the proportional coefficient of the flow in the recent time range and the flow in the corresponding time range of the reference sample. The method of the invention can be used as a supplement to the existing traffic flow prediction, thereby improving the accuracy of the flow prediction.

Description

Expressway ETC traffic flow prediction method

Technical Field

The invention belongs to the technical field of intelligent traffic, and particularly relates to a highway ETC traffic flow prediction method.

Background

With the rapid development of social economy and the gradual popularization of household automobiles, the rhythm of people between cities is more and more frequent, and more vehicles run on expressways. The method has the advantages that the whole change trend of the traffic state of the highway is effectively grasped, and is a necessary premise for effectively managing the operation of the highway and preventing traffic jam; meanwhile, the overall grasp of the flowing trend of the vehicles among cities is beneficial to improving the management of urban traffic and improving the comprehensive reception capacity and service level of the cities.

The prediction of traffic flow is always a research hotspot in the field of intelligent transportation, and scholars at home and abroad put forward a large number of prediction methods which can be roughly divided into: the prediction method based on the linear system, the prediction method based on the nonlinear system, the intelligent prediction method based on the knowledge discovery and the prediction method based on the combination mode. These prediction methods basically predict based on historical data and real-time data of traffic flow without considering the influence of the start and end points of vehicles on traffic.

An Electronic Toll Collection (ETC) system is a kind of Electronic automatic Toll Collection system commonly applied to highways, bridges and tunnels, and can automatically calculate rates and deduct Toll fees under the condition of no parking. With the popularization of ETC, more and more vehicles are equipped with ETC. The ETC system records a large amount of vehicle traffic information including driving routes, driving time and the like, the information has close correlation with the changes of highway and urban traffic flow, and currently, research on a prediction method of the highway flow of the ETC is less.

The highway flow prediction is one of the core problems with great difficulty in the field of intelligent transportation at present, but the accuracy of the conventional prediction method is generally insufficient. The method for predicting the ETC vehicle flow on the highway by utilizing the vehicle travel information of the ETC system can be used as a supplement of the conventional traffic flow prediction method, thereby improving the accuracy of flow prediction and having important significance for improving the highway traffic management level.

Disclosure of Invention

The invention aims to provide a highway ETC traffic flow prediction method, which can predict the highway ETC traffic flow by utilizing the vehicle travel information of an ETC system, can be used as a supplement of the existing traffic flow prediction, and thus improves the accuracy of the traffic flow prediction.

In order to solve the technical problem, the invention provides a highway ETC traffic flow prediction method, which comprises the following steps:

step one, extracting ETC vehicle passing historical data, and combining a pair of entrance and exit data into a high-speed entrance and exit record;

step two, sorting the high-speed access records according to time, and screening the round-trip routes in the records;

step three, segmenting the time, and counting the distribution data of vehicles entering and leaving the city in each time period, wherein the method specifically comprises the following steps:

first, in time interval s of date D, from city CIn_iTo city COut_jNumber of vehicles Num (D, s, CIn)_i，COut_j) The calculation method is as follows,

Num(D，s，CIn_i，COut_j)＝NumW(D，s，CIn_i，COut_j)+NumO(D，s，CIn_i，COut_j)

wherein, NumW (D, s, CIn)_i，COut_j) The number of vehicles constituting the round trip route;

NumO(D，s，CIn_i，COut_j) Do not constitute a round tripThe number of vehicles IN the route is that i is more than or equal to 1 and less than or equal to IN, j is more than or equal to 1 and less than or equal to JN, IN is the number of entrance cities, JN is the number of exit cities, s is more than or equal to 1 and less than or equal to SN, and SN is the number of time intervals dividing one day;

then, in time interval s of date D, from city CIn_iNumber of vehicles coming out

NumCOut(D，s，CIn_i) COut to the city_jNumber of vehicles NumCIn (D, s, COut)_j) From city CIn_iNumber of outgoing shuttle vehicles NumCouWs (D, s, CIn)_i) COut to the city_jNumber of vehicles NumClinW (D, s, COut) to and from_j) The calculation methods are as follows, respectively,

selecting the flow of the similar time interval as a prediction reference sample according to the characteristics of the time interval to be predicted;

estimating the flow in the time interval to be predicted according to the prediction reference sample and the flow in the latest time range; the flow in the time interval to be predicted is from the city CIn in the time interval s of the date Dpre to be predicted_iNumber of vehicles driving out NumCout (Dpre, s, CIn)_i) COut to the city_jNumber of vehicles (Dpre, s, COut)_j) From city CIn_iNumber of outgoing shuttle vehicles NumCOutW (Dpre, s, CIn)_i) And reach city COut_jNumber of vehicles (Dpre, s, COut) to and from NumCinW_j) The calculation methods are shown below, respectively,

NumCOut(Dpre，s，CIn_i)＝α×NumCOut(Dref，s，CIn_i)

wherein,

NumCIn(Dpre，s，COut_j)＝β×NumCIn(Dref，s，COut_j)

wherein,

NumCOutW(Dpre，s，CIn_i)＝γ×NumCOutW(Dref，s，CIn_i)

wherein,

NumCInW(Dpre，s，COut_j)＝λ×NumCInW(Dref，s，COut_j)

wherein,

the date to be predicted is Dpre, the date of the prediction reference sample is Dref, the current date is Dnow, the date of the reference sample corresponding to the current date Dnow is Drnow-Dpre + Dhow, the nearest time range is DN days, and d is the number of days from the current date.

Preferably, the historical data of ETC vehicle passage refers to vehicle running information automatically recorded by an ETC system when a vehicle enters or exits the expressway through an ETC special channel;

the pair of entrance and exit data refers to vehicle driving information of a toll station for driving in and out of the same vehicle during one-time highway passing, and specifically comprises the following information:

driving in high speed through entrance toll station PortIn at time TIn, wherein the city where the toll station PortIn is located is CIn;

outputting high speed through an exit toll station PortOut at the time TOut, wherein the city where the toll station PortOut is located is COut, the running time is TOut-TIn, and the running distance is Dist;

the one high speed access record comprises: the vehicle identification CarID, the entrance city CIn, the entrance time TIn, the exit city COut, the exit time TOut, the travel time TOut-TIn, and the travel distance Dist. .

Compared with the prior art, the method has the obvious advantages that the method predicts the ETC travel vehicles on the highway based on the historical information of ETC traffic, analyzes the driving trend of the vehicles among cities, the travel condition of urban vehicles, the driving condition of the urban vehicles and the urban conditions of driving short trips in each vacation period, is favorable for improving the service level of a highway management department and a city management department, and can be used as a supplement for the conventional traffic flow prediction.

Drawings

Fig. 1 is a schematic flow chart of the highway ETC vehicle flow prediction method of the invention.

Detailed Description

It is easily understood that according to the technical solution of the present invention, those skilled in the art can imagine various embodiments of the highway ETC vehicle flow prediction method of the present invention without changing the essential spirit of the present invention. Therefore, the following detailed description and the accompanying drawings are merely illustrative of the technical aspects of the present invention, and should not be construed as all of the present invention or as limitations or limitations on the technical aspects of the present invention.

With the combination of the attached drawings, the urban round-trip circular route in the traffic record is detected based on the historical information of vehicle travel in the ETC system, the prediction reference sample is selected according to the characteristic that the traffic in the same time period has correlation, and the traffic in the specified time interval is predicted according to the real-time traffic and the prediction reference sample. When the ETC vehicle flow is predicted, the flow of a similar reference sample is considered, and the difference between the reference sample and the predicted flow is also considered. And correcting the difference of the prediction through the proportional coefficient of the flow in the recent time range and the flow in the corresponding time range of the reference sample.

As shown in fig. 1, a flowchart of the method for predicting the vehicle flow on the highway ETC according to the embodiment includes the following basic steps:

step 101, extracting ETC vehicle passing history data, and combining a pair of entrance and exit data into a high-speed entrance and exit record. In accordance with the present embodiment of the invention,

the ETC vehicle passing historical data refers to vehicle running information automatically recorded by an ETC system when a vehicle passes through an ETC special channel to enter and exit the expressway;

the pair of entrance and exit data refers to driving data of the same vehicle entering and exiting the toll station during one-time highway passing, namely vehicle driving information corresponding to the starting point (entrance) and the ending point (exit) of the highway for one-time toll collection. For example, the history data of one pass of the vehicle Car1 includes a pair of entrance/exit data:

an inlet: driving in high speed through entrance toll station PortIn at time TIn, wherein the city where the toll station PortIn is located is CIn;

and (4) outlet: outputting high speed through an exit toll station PortOut at the time TOut, wherein the city where the toll station PortOut is located is COut, the running time is TOut-TIn, and the running distance is Dist;

then the merged one high speed entry record is: the vehicle identification CarID, the entrance city CIn, the entrance time TIn, the exit city COut, the exit time TOut, the travel time TOut-TIn, and the travel distance Dist. The vehicle identification is used for uniquely identifying one vehicle, the time unit adopts minutes, and the distance unit adopts kilometers.

And 102, sorting the high-speed access records according to time, and screening the round-trip routes in the records.

The holiday self-driving travel is generally a short-distance travel, and one characteristic of the route is that the cities of the route form a round-trip annular route, namely, the city which starts from a certain city and finally returns to the starting city.

According to the present embodiment, assuming that the high speed access records of the Car1 are sorted by time, the high speed entrance city of the route of the continuous KN high speed access records is CIn_kHigh speed export city COut_kAnd heightFast driving time TIn_kWherein k is more than or equal to 1 and less than or equal to KN, and if the following conditions are met, the KN high-speed access records form a round-trip route:

(1) the number of days of travel is less than a specified number of days, i.e. (TIn)_KN-TIn₁) (24 × 60) < DayNum +1, where DayNum takes the number of days of vacation for vacation and non-vacation takes a constant of 3 days;

(2)CIn_k＝COut_k-1wherein k is more than or equal to 2 and less than or equal to KN;

(3)CIn₁＝COut_KN。

and 103, segmenting the time, and counting the distribution data of vehicles entering and exiting the city in each time period.

According to the specific embodiment, the time is segmented into 30-minute intervals, and the SN is 48 time intervals in 24 hours a day. Suppose that from city CIn, in the time interval with date D interval s (s is more than or equal to 1 and less than or equal to SN)_iTo COut_jNumber of vehicles (D, s, CIn) is recorded as Num (D, s, CIn)_i，COut_i)：

NumO(D，s，CIn_i，COut_j) I is more than or equal to 1 and less than or equal to IN, j is more than or equal to 1 and less than or equal to JN, wherein IN is the number of entrance cities, and JN is the number of exit cities. Then there is a change in the number of,

(1) in the time interval with the date D interval being s, from city CIn_iThe number of outgoing vehicles is recorded as:

(2) in the time interval with the date D interval of s, the city COut is reached_jThe number of vehicles is noted as:

(3) in the time interval with the date D interval being s, from city CIn_iThe number of outgoing shuttle vehicles is recorded as:

(4) in the time interval with the date D interval of s, the city COut is reached_jThe number of round trip vehicles is recorded as:

and 104, selecting the flow of the similar time interval as a prediction reference sample according to the characteristics of the time interval to be predicted.

According to the specific embodiment, the dates are divided into three types, namely holidays, weekend holidays and working days, and the similar time intervals respectively selected by the prediction reference samples are as follows: the holiday data is the same as the holiday data of the previous year, the holiday data of the last week is the holiday data of the last week, and the data of the same working day of the last week is the working day. The date of the prediction reference sample that records the date to be predicted Dpre is Dref.

And 105, estimating the flow in the time interval to be predicted according to the reference sample and the flow at the latest time.

The prediction takes into account the flow rate of the similar reference sample, as well as the difference between the reference sample and the predicted flow rate. According to the present embodiment, the difference in the prediction is corrected by the scaling factor of the flow in the recent time range to the flow in the corresponding time range of the reference sample. Assuming that the current date Dnow corresponds to a reference sample date Drnow ═ Dref-Dpre + Dnow, and the latest time range selects DN days, then

(1) From city CIn in time interval with date Dpre interval s_iThe number of driven-out (high-speed entrance) vehicles is predicted as:

NumCOut(Dpre，s，CIn_i)＝α×NumCOut(Dref，s，CIn_i)

wherein the proportionality coefficient

d is the number of days from the current date

(2) At a time interval with a date Dpre of s, the city COut is reached_jThe number of (high exit) vehicles is noted:

NumCIn(Dpre，s，COut_j)＝β×NumCIn(Dref，s，COut_j)

wherein the proportionality coefficient

(3) From city CIn in time interval with date Dpre interval s_iThe number of outgoing round-trip vehicles (vacation trips) is recorded as:

NumCOutW(Dpre，s，CIn_i)＝γ×NumCOutW(Dref，s，CIn_i) Wherein the proportionality coefficient

(4) At a time interval with a date Dpre of s, the city COut is reached_jThe number of round-trip vehicles (vacation receptions) is recorded as:

NumCInW(Dpre，s，COut_j)＝λ×NumCInW(Dref，s，COut_j) Wherein the proportionality coefficient

Claims

1. A highway ETC vehicle flow prediction method is characterized by comprising the following steps:

first, in time interval s of date D, from city CIn_iTo city COut_jNumber of vehicles Num (D, s, CIn)_i,COut_j) The calculation method is as follows,

Num(D,s,CIn_i,COut_j)＝NumW(D,s,CIn_i,COut_j)+NumO(D,s,CIn_i,COut_j)

wherein, NumW (D, s, CIn)_i,COut_j) The number of vehicles constituting the round trip route; NumO (D, s, CIn)_i,COut_j) For the number of vehicles which do not form a round-trip route, i is more than or equal to 1 and less than or equal to IN, j is more than or equal to 1 and less than or equal to JN, IN is the number of entrance cities, JN is the number of exit cities, s is more than or equal to 1 and less than or equal to SN, and SN is the number of time intervals dividing one day;

then, in time interval s of date D, from city CIn_iNumber of vehicles exiting NumCout (D, s, CIn)_i) COut to the city_jNumber of vehicles NumCIn (D, s, COut)_j) From city CIn_iNumber of outgoing shuttle vehicles NumCouWs (D, s, CIn)_i) COut to the city_jNumber of vehicles NumClinW (D, s, COut) to and from_j) The calculation methods are as follows, respectively,

selecting the flow of the similar time interval as a prediction reference sample according to the characteristics of the time interval to be predicted; the similar time intervals are: the holiday takes the same holiday data of the previous year, the last holiday takes the last holiday data of the previous week, and the working day takes the same working day data of the previous week;

NumCOut(Dpre,s,CIn_i)＝α×NumCOut(Dref,s,CIn_i)

wherein,

NumCIn(Dpre,s,COut_j)＝β×NumCIn(Dref,s,COut_j)

wherein,

NumCOutW(Dpre,s,CIn_i)＝γ×NumCOutW(Dref,s,CIn_i)

wherein,

NumCInW(Dpre,s,COut_j)＝λ×NumCInW(Dref,s,COut_j)

wherein,

the date to be predicted is Dpre, the date of the prediction reference sample is Dref, the current date is Dnow, the date of the reference sample corresponding to the current date Dnow is Drnow-Dpre + Dnow, the nearest time range is DN days, and d is the number of days from the current date.

2. The method for predicting the ETC vehicle flow on the expressway according to claim 1, wherein the historical data of ETC vehicle traffic refers to vehicle driving information automatically recorded by an ETC system when a vehicle enters and exits the expressway through an ETC special channel;

the one high speed access record comprises: the vehicle identification CarID, the entrance city CIn, the entrance time TIn, the exit city COut, the exit time TOut, the travel time TOut-TIn, and the travel distance Dist.