JP3585226B2 - Traffic condition prediction method, apparatus, program, and recording medium recording the program - Google Patents

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a traffic situation prediction method and apparatus for predicting a future traffic situation based on a current traffic situation.
[0002]
[Prior art]
For example, as shown in FIG. 7, the traffic congestion point information is measured at measurement points placed every several tens to several hundred meters on a certain road, and there is no traffic congestion (0), congestion (1), and traffic congestion (2). ) And several levels. In this case, the traffic congestion state on a certain road can be represented by a series of numbers 0, 1, and 2, for example, "001112200". Conventionally, there have been the following two methods for predicting such a traffic jam situation.
1. A method of predicting sequentially for each measurement point, that is, for all measurement points.
2. A method of predicting the length of a traffic jam from a certain fixed point, that is, a method of determining a fixed point where traffic jams frequently occur, and expressing traffic congestion state data by how much traffic is congested from the fixed point.
[0003]
[Problems to be solved by the invention]
At present, traffic information centers such as ATIS (Advanced Traffic Information Systems) and VICS (Vehicle Information and Communication Service) provide traffic congestion point information as shown in FIG. 7 through car navigation and the like, which has several levels of congestion levels. are doing.
[0004]
However, the conventional traffic congestion point information prediction method described above has the following problems.
[0005]
1. Although the method of (1) does not impair the information amount of data, the prediction processing takes a very long time because the prediction processing is performed for all of a large number of measurement points.
[0006]
2. In the method of (1), it is possible to reduce the target variable as the prediction target quantity to one. However, the occurrence of traffic congestion is not limited to a specific location, but may occur at an arbitrary location on a road. When the traffic congestion state has several levels, it is impossible to sufficiently express the traffic congestion level only by the length from the fixed point. As described above, when data is expressed only by the length from the fixed point, there is a possibility that the information amount of the data is lost.
[0007]
An object of the present invention is to provide a traffic condition prediction method and device capable of quickly and accurately predicting a traffic congestion condition.
[0008]
[Means for Solving the Problems]
First, the traffic congestion state of a section, which is expressed by quantifying the congestion degree at a certain time at a plurality of measurement points in a predetermined section of a road at a plurality of levels, is chronologically represented. Next, a traffic jam status number, which is an identification number, is assigned to each type of time-series traffic jam status, and a traffic jam status table including the traffic jam status numbers and the traffic jam status is created. Next, the time-series traffic congestion status is converted into a traffic congestion status number with reference to the traffic congestion status table. Next, a traffic jam number is predicted using the converted traffic jam status number, and a predicted traffic jam status number is calculated. Finally, the predicted traffic congestion status number is converted into a traffic congestion status using the traffic congestion status table.
[0009]
By converting traffic congestion point data (traffic congestion status) into traffic congestion status numbers, it is possible to reduce the amount of data and easily handle traffic congestion status data, and accurately and quickly predict traffic congestion status. Can be realized.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
Next, embodiments of the present invention will be described with reference to the drawings.
[0011]
Referring to FIG. 1, a traffic situation prediction apparatus according to an embodiment of the present invention includes a traffic congestion point data time-series unit 101, a traffic congestion situation table creation unit 102, a traffic congestion situation number data creation unit 103, a predicted traffic congestion situation. It comprises a number calculation unit 104, a predicted traffic congestion status conversion unit 105, and a storage unit 106.
[0012]
The traffic congestion point data time-series unit 101 converts the traffic congestion point data (FIG. 7) stored in the storage unit 106 into time-series traffic congestion point data for each link, and stores the data in the storage unit 106. The traffic congestion status table creation unit 602 creates a traffic congestion status table including a traffic congestion status number, a traffic congestion status, and the number of times of observation based on the time-series traffic congestion location data in link units stored in the storage unit 106. Is stored in the storage unit 106. The traffic congestion situation number data creating unit 103 creates traffic congestion situation number data based on the traffic congestion situation table stored in the storage unit 106 and the time-series traffic congestion location data in link units, and stores the data in the storage unit 106. The predicted traffic congestion status number calculation unit 104 calculates a predicted traffic congestion status number based on the traffic congestion status number data stored in the storage unit 106, and stores it in the storage unit 106. The predicted traffic congestion status conversion unit 105 converts the predicted traffic congestion status number stored in the storage unit 106 into a traffic congestion status using a traffic congestion status table, and outputs the result as the predicted traffic congestion status.
[0013]
Next, the operation of the present embodiment will be described with reference to FIGS. As a premise of the present invention, it is assumed that traffic congestion point data on a service target road can be collected through a traffic information center such as ATIS or VICS.
[0014]
First, in step 201, the traffic congestion point data 301, 302,... Collected from the traffic information center and the like are time-series on a link basis, and the time-series traffic congestion point data 311 for link 1 and the time-series traffic congestion point data for link 2 312,... Are created. A link is obtained by dividing a road by an intersection or the like, and refers to, for example, a road portion from one intersection to the next intersection. The data collected from the traffic information center is divided at each measurement time as indicated by 301, 302,... In FIG. 3, and one data includes data of many links. Data is extracted from the data at each measurement time in units of links, as shown by 311, 312,... In FIG.
[0015]
Next, in step 202, all the types of observed data are extracted from the traffic congestion point data time-series for each link, summarized in a table as shown in FIG. 4, and assigned an identification number. The number of observations is also summarized. Hereinafter, the traffic congestion point data summarized in the extraction table will be referred to as traffic congestion status 403, and the assigned identification number will be referred to as traffic congestion status number 402. The table will be referred to as a traffic congestion state table 401. As shown in FIG. 4, the traffic congestion status table 401 sequentially stores the assigned traffic congestion status number 402, the actual traffic congestion status 403, and the number of observations 404.
[0016]
Next, in step 203, as shown in FIG. 5, the traffic congestion state 502 in the traffic congestion point data 500 time-series for each link is converted into a traffic congestion state number, and the time 521 and the traffic congestion state number 522 are used. Then, the time-series traffic congestion status number data 520 is created. Specifically, the traffic congestion state 502 of the traffic congestion point data 500 in the time series is compared with the traffic congestion state 512 of the traffic congestion state table 510 compiled in step 202, and the corresponding traffic congestion state Is converted into the traffic congestion status number 511 assigned to. By expressing the traffic congestion status at each time by the traffic congestion status number in this manner, it becomes possible to handle data on a link basis while reducing the data amount.
[0017]
Next, in step 204, a traffic congestion status number is predicted using the converted traffic congestion status number data 520, and a predicted traffic congestion status number is calculated. As a prediction method, for example, a method using the decision tree of Japanese Patent Application No. 2001-20433 can be considered. In this step, the traffic congestion status is not directly predicted but the prediction is performed with the traffic congestion status number as the prediction target. This makes it possible to improve the prediction processing speed and reduce the memory.
[0018]
FIG. 6 shows an example in which a traffic congestion number is predicted from the traffic congestion status number data 520 using the decision tree of Japanese Patent Application No. 2001-20433. Here, as shown in FIG. 6, a decision tree is a tree in which one node corresponds to a test of one attribute, child nodes of the node correspond to possible values of the attribute, and leaves are predicted values. It is a structure. The decision tree is created for each of the prediction target roads for each link and for each prediction time of how many minutes to predict. At the time of prediction, a predicted value is determined by sequentially following such a decision tree from a root node. In the example of FIG. 6, for example, a case where the traffic congestion situation number prediction value at 14:45 on Monday is calculated is described below. First, the attribute value of the day of the week is tested at the root node 601, and since it is Monday, the process proceeds to the next node 602. At the node 602, the attribute value of the time is tested, and the process proceeds to the leaf 603, where the predicted value “1” of the traffic congestion status number is determined. At 14:50, the process proceeds to leaf 604, and the predicted value “1” of the traffic congestion status number is determined.
[0019]
Finally, in step 205, the calculated predicted traffic congestion status number is converted into a predicted traffic congestion status using the traffic congestion status table. Specifically, the operation reverse to that of step 203 is performed.
[0020]
By the way, in the present embodiment, a method using a decision tree has been described as a prediction method. However, a method using a search of a past case of Japanese Patent Application No. 10-233189 may be considered, and the prediction method is limited to the present embodiment. Not done.
[0021]
In this embodiment, a method using a link unit as a unit of time series has been described, but a method using an intersection unit and a method using a mesh unit are also conceivable. Is not limited to this embodiment.
[0022]
Note that the processing shown in FIG. 2 can be recorded on a recording medium such as a floppy disk, CD-ROM, or magneto-optical disk as a traffic condition prediction program and executed on a computer such as a personal computer.
[0023]
【The invention's effect】
As described above, according to the present invention, an accurate and short processing time can be realized by tabulating the observed traffic congestion status, assigning the traffic congestion status number, and predicting the traffic congestion status number.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a traffic situation prediction device according to an embodiment of the present invention.
FIG. 2 is a flowchart showing the operation of the traffic situation prediction device of FIG.
FIG. 3 is a diagram illustrating a procedure for time-series traffic congestion point data in link units.
FIG. 4 is a diagram showing an example of a traffic congestion state table.
FIG. 5 is a diagram showing a method of converting traffic jam status number data.
FIG. 6 is a diagram illustrating an example of predicting a traffic congestion status number using a decision tree disclosed in JP-A-2001-20433.
FIG. 7 is a diagram showing an example of traffic congestion point data targeted by the present invention.
[Explanation of symbols]
101 Traffic congestion point data time series unit 102 Traffic congestion situation table creation unit 103 Traffic congestion situation number data creation unit 104 Predicted traffic congestion situation number calculation unit 105 Predicted traffic congestion situation conversion unit 106 Storage units 201 to 205 Steps 301 and 302 Traffic Traffic congestion location data 311, 312 Time-series traffic congestion location data 401 Traffic congestion status table 402 Traffic congestion status number 403 Traffic congestion status 404 Number of times observed 500 Traffic congestion location data 501 for link n Time 502 Traffic congestion status 510 Traffic congestion status table 511 Traffic congestion status number 512 Traffic congestion status 520 The traffic congestion status number data 521 at the converted link n Time 522 Traffic congestion status number 601 Root node 602 Nodes 603, 604 Leaf

Claims (5)

A traffic condition prediction method for predicting a future traffic condition based on a current traffic condition,
Time-series traffic congestion status of the section represented by quantifying the congestion degree at a certain time of a plurality of measurement points in a predetermined section of the road at a plurality of levels,
A step of assigning a traffic congestion situation number that is an identification number to each type of the traffic congestion situation in the time series, and creating a traffic congestion situation table including the traffic congestion situation number and the traffic congestion situation,
Referring to the traffic jam table, converting the time-series traffic jam situation to a traffic jam situation number,
Predicting a traffic jam number using the converted traffic jam status number, and calculating a predicted traffic jam status number;
A traffic situation prediction method comprising a step of converting the predicted traffic jam situation number into a traffic jam situation using the traffic jam situation table. The method of claim 1, wherein the traffic jam table includes a number of times the traffic jam has been observed. A traffic condition prediction device that predicts a future traffic condition based on a current traffic condition,
A storage device,
A traffic congestion state of a section expressed by numerically expressing congestion degrees at a plurality of levels at a plurality of measurement points in a predetermined section of a road in a time series, and storing the traffic congestion in the storage device. Location data time series means,
The time-series traffic jam situation is read out from the storage device, a traffic jam situation number as an identification number is assigned to each type, and a traffic jam situation table including the traffic jam situation number and the traffic jam situation is created. Means for creating a traffic jam table stored in the storage device;
Reading the traffic congestion status table from the storage device, referring to the table, converting the time-series traffic congestion status into a traffic congestion status number, and storing the traffic congestion status number data in the storage device; ,
The converted traffic congestion status number is read from the storage device, a traffic congestion number is predicted using the number, a predicted traffic congestion status number is calculated, and the predicted traffic congestion status number calculation means is stored in the storage device. When,
A traffic condition predicting device comprising a predicted traffic congestion condition converting means for reading the predicted traffic congestion condition number from the storage device and converting the number into a traffic congestion condition using the traffic congestion condition table. A traffic condition prediction program for causing a computer to execute the method according to claim 1. A recording medium on which the traffic situation prediction program according to claim 4 is recorded.

Images