JP2000067360A - Traffic information prediction method, device therefor, recording medium having recorded traffic information prediction program thereon, link classification method and recording medium having recorded link classification program thereon - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately and efficiently predict complicatedly changing traffic information. SOLUTION: A selection part 11A inputs the traffic information 21 of a link and selects and outputs the traffic information 221-22m of the links for every link group and the traffic information 23 of a prediction object link. The traffic information 23 is stored in a storage part 14. Link group traffic information calculation parts 121-12m calculate the traffic information 241-24m of the link groups based on the traffic information 221-22m and respectively store them in the storage parts 131-13m. A prediction part 16A reads the link group traffic information 241-24m from the storage parts 131-13m and the traffic information 23 of the prediction object link from the storage part 14, predicts the traffic information of the prediction object link and outputs predicted information 26.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a traffic information prediction method and apparatus, and a link classification method.

[0002]

2. Description of the Related Art Conventional traffic information prediction methods include predicting future traffic information from past traffic information of a link to be predicted, and past traffic information of a link for predicting traffic information as shown in FIG. There has been a method of predicting future traffic information of a link for which traffic information is predicted from information and past traffic information of an adjacent link connected to the link for which traffic information is predicted. However, this method has a problem that prediction accuracy is low in predicting traffic information that changes in a complicated manner.

In order to solve this, there is a method shown in FIG. Here, the traffic information of the link to be predicted is predicted using the traffic information of the link in the area including the link to be predicted. According to this method, traffic information of a wide range of links can be used, so that traffic information that changes in a complicated manner can be accurately predicted.

[0004]

However, in this method, when the number of links included in the area increases, the capacity for storing traffic information increases, and the number of traffic information used for prediction becomes enormous. When the prediction unit is realized by the device, there is a problem that the configuration is complicated, and when the prediction unit is realized by the program processing, there is a problem that the prediction process takes time.

In this method, the link to be predicted is
If it's not always at the center of the area but around the area,
There was a problem that the traffic information of the surrounding links could not be reflected and the accuracy of the prediction did not increase.

An object of the present invention is to provide a traffic information prediction method for accurately and efficiently predicting traffic information that changes in a complicated manner.

Another object of the present invention is to provide a specific method for dividing links into groups.

[0008]

According to the present invention, a link is divided into groups by an adjacent link connected to a link to be predicted, traffic information of each of the divided link groups is obtained, and traffic information of the obtained link group is obtained. The traffic information of the link to be predicted is predicted from the traffic information of the link to be predicted.

In the present invention, when the traffic information of the link to be predicted is predicted, the traffic information of each link group is
The traffic information of the adjacent link is used in addition to the traffic information of the link to be predicted.

In addition to the traffic information of the link to be predicted (and the traffic information of the adjacent link), the links are grouped by adjacent links connected to the link to be predicted, traffic information of the link group is obtained, and traffic of the link group thus obtained is obtained. It differs from the prior art in that traffic information of a prediction target link is predicted using information.

Claims 3 to 5 provide a method of dividing a link into groups by a link adjacent to a link to be predicted.

[0013] In the third aspect, a shortest path to each node is obtained starting from one node at the end of the link to be predicted, and a first path which does not pass through the link to be predicted from among the obtained optimum paths.
, And the shortest path to each node is obtained, starting from the other node at the end of the link to be predicted.
A second route that does not pass through the prediction target link is extracted from the obtained optimum route, and the link is divided into groups for each link included in the route that passes through each adjacent link of the prediction target link from the extracted route. It is characterized by the following.

[0013] Claim 4 corresponds to the fact that a link may be overlapped between the first route and the second route extracted in claim 3, and overlaps with the first route and the second route. If there is a duplicate link, remove one of the duplicate links, and replace each of the links included in the route that passes through each adjacent link of the prediction target link from the remaining route. It is characterized by being divided into groups.

A fifth aspect of the present invention corresponds to the third and fourth aspects, wherein the shortest route obtained by the optimal route search includes a link that does not pass through among links to be divided into groups. By extracting links that are not included in the extracted route among the links to be divided into groups, and incorporating the extracted links into a group of links to which one of the nodes at both ends of the extracted links belongs, the links are extracted. It is characterized by being divided into groups.

[0015]

Next, embodiments of the present invention will be described with reference to the drawings.

Referring to FIG. 1, first the embodiment of the traffic information prediction apparatus selection section 11A and the link group traffic information calculating unit 12 ₁ to 12 _m and the storage unit 13 ₁ to 13 _m of the present invention
And a storage unit 14 and a prediction unit 16A. Here, the number of link groups is m.

The selecting unit 11A divides the links into groups based on the connection relations of the links adjacent to the link to be predicted, and stores the link groups. When the link traffic information 21 is input, the link traffic information 22 _{1 for} each link group is entered.
M22 _m , and outputs the traffic information 23 of the traffic link.

The link group traffic information calculators 12 ₁ to 1
2 _m is the link traffic information 22 ₁ to 2 from the selection unit 11A.
If 2 _m is input, calculates the traffic information 24 ₁ to 24 _m link group on the basis of these, the storage unit 13 ₁ to 1
Send to 3 _m . Link group traffic information 24 _1-2
4 _m, specifically, for example, when the link traffic information 21 is represented by the average speed of the link, can take an average for all links in the average speed of the link group of links, the links When the traffic information is represented by the travel time of the link, the total travel time for all the links in the link group can be calculated. In this way, an appropriate amount may be determined as the link group traffic information 24 _{1 to} 24 _m according to the amount represented by the link traffic information 21.

The storage units 13 _{1 to} 13 _m store N values from the present to the past of the traffic information 24 _{1 to} 24 _m of the link group. Link group traffic information calculators 12 ₁ -12
from _m every time the traffic information 24 ₁ to 24 _m of the link group is input, and discard the traffic information of the oldest N-th link group, and updates the traffic information link group.

The storage unit 14 stores the past N values of the traffic information 23 of the link to be predicted. Every time the traffic information 23 of the prediction target link is input from the selecting unit 11A, the traffic information of the Nth prediction target link is discarded, and the traffic information of the prediction target link is updated.

The prediction unit 16A to the storage unit 13 ₁ to 13 _m and a storage unit 14, the traffic information 24 of each link group
_{1 to} 24 _m , each time the traffic information 23 of the link to be predicted is input, the traffic information 24 _{1 to} 24 of N link groups
Read out the traffic information 23 of _m and N prediction target links,
The traffic information of the link to be predicted is predicted, and prediction information 26 is output. As a specific method of the prediction, an autoregressive method or a neural network can be provided.

In this way, when the link traffic information 21 is input, the link traffic information 2 in the link group is entered.
2 ₁ than ~22 _m of the link group traffic information 24 _1-2
4 _m is calculated and stored, and the traffic information 23 of the link to be predicted is stored, so that prediction can be performed based on the traffic information.

Referring to FIG. 2, the second the embodiment of the traffic information prediction apparatus selecting section 11B and the link group traffic information calculating unit 12 ₁ to 12 _m and the storage unit 13 of the present invention _1-13
m , 14, 15 ₁ to 15 _m and a prediction unit 16B.

The traffic information predicting apparatus according to the present embodiment has
The selection unit 11A and the prediction unit 16A of the communication information prediction device are respectively
Storage unit instead of the selection unit 11B and the prediction unit 16B
Fifteen ₁ ~ 15_m Is added.

The selecting unit 11B stores adjacent links and link groups for each link to be predicted. When the link of the traffic information 21 is input, and outputs to select the traffic information 22 ₁ through 22 _m of links each link group, the traffic information 23 in the prediction target link traffic information 25 ₁ to 25 _m of the adjacent link Output.

The storage unit 15 ₁ to 15 _m stores the current past N values from the traffic information 25 ₁ to 25 _m of the adjacent link. Then, the traffic information 2 of the adjacent link from the selection unit 11B
Every time 5 _{1 to} 25 _m is input, the traffic information of the oldest N-th adjacent link is discarded, and the traffic information of the surrounding link is updated.

The prediction unit 16B stores the traffic information 24 _{1 to} 24 _{m of} the link group, the traffic information 23 of the link to be predicted, and the surroundings in the storage units 13 _{1 to} 13 _m , the storage unit 14 and the storage units 15 ₁ to 15 _m , respectively. Each time the link traffic information 25 _{1 to} 25 _m is input, N traffic information 24 ₁ of the link group is input.
２４24 _m , N traffic information 23 of the link to be predicted and N traffic information 25 ₁ 125 _m of each adjacent link are read out,
The traffic information of the link to be predicted is predicted, and the prediction information 26
Is output.

As described above, when the link traffic information 21 is input, the link traffic information 2 in the link group is input.
2 ₁ than ~22 _m of the link group traffic information 24 _1-2
4 _m is calculated and stored, and the traffic information 23 of the link to be predicted and the traffic information 25 _{1 to} 25 _{m of the} adjacent link are stored, and the prediction can be performed based on the traffic information.

Referring to FIG. 3, a traffic information prediction device according to a third embodiment of the present invention includes an input device 31 and a storage device 32 _1.
32 _m , 33, an output device 35, a recording medium 36A, and a data processing device 37.

The input device 31 is an input device such as a modem for inputting the traffic information 21 of the link. Storage device 32 ₁ to
32 _m corresponds respectively to the storage unit 13 ₁ to 13 _m in FIG. The storage device 33 corresponds to the storage unit 14 in FIG.
The output device 35 is an output device, such as a display or a printer, to which the prediction information 26 is output. The recording medium 36A is
Selector 11A in FIG. 1, the link traffic information calculator 12 ₁ -
12 _m , a floppy disk or a CD on which a traffic information prediction program including the processing of the prediction unit 16A is recorded.
-Recording media such as ROMs, magneto-optical disks, and semiconductor memories. The data processing device 37 reads the traffic information prediction program from the recording medium 36A and executes the CP for executing the program.
U.

Referring to FIG. 4, a traffic information prediction device according to a fourth embodiment of the present invention includes an input device 31 and a storage device 32 _1.
32 _m , 33, 34 _{1 to} 34 _m , an output device 35, a recording medium 36B, and a data processing device 37.

In this embodiment, storage devices 34 _{1 to} 34 _m are added to the traffic information prediction device of FIG.
6B replaces the recording medium 36A.

[0033] corresponds to a storage device 34 ₁ to 34C _m storage unit 15 in the 2 each view ₁ to 15 _m. The recording medium 36B, the selection section 11B in FIG. 2, link group traffic information calculating unit 12 ₁ to 12 _m, the traffic information prediction program of the processing of the prediction unit 16B is recorded.

FIG. 5 is a flowchart showing a link dividing method according to an embodiment of the present invention.

Step 41: First, the shortest route to each node is obtained by an optimal route search method, starting from one node at the end of the link to be predicted (for example, node 1 in FIG. 7). Among them, the route passing through the link to be predicted is excluded. The shortest path to each node of the remaining paths is represented by a tree structure. The Dijkstra method can be applied as the optimum route search method.

Step 42: Next, the shortest path to each node is determined by the optimal path search method, starting from the other node (node 2) at the end of the link to be predicted. Among them, the route passing through the link to be predicted is excluded. The shortest path to each node of the remaining paths is represented by a tree structure.

When the above processing is performed on the link network of FIG. 6, two tree structures as shown in FIG. 8 are obtained, for example.

Step 43 (Claim 4): At this time,
When two routes that go through the same link are calculated, such as a route that goes through the link of No. 7, the highest link of the overlapping link (the link of No. 7 in the example of FIG. 7).
The highest link of this overlapping link is incorporated into the shortest link to the link (in this case, 6), and the lower link than the highest link of this overlapping link is separated.

As a result, as shown in FIG. 9, a tree of links that do not include the same link can be calculated with the link adjacent to the link to be predicted being at the top.

Step 44 (Claim 5): In the optimal route search method, some links do not pass. In the example of FIG. 6, the links with numbers 10, 13, 18, 21, 29, 31, 33, and 39 correspond to this. This is incorporated into a link group. As a specific method, the shortest path to the nodes at both ends of the link that does not pass can be incorporated into the shorter node.

Step 45: In this manner, as shown in FIG. 10, for example, in the optimal route search method, the same link is not included, with the link adjacent to the prediction target link being the highest rank, including the link that does not pass. A tree of links can be derived, and the link can be divided into links based on links adjacent to the link to be predicted.

In the example shown in FIG. 7, an example in which the tree structure has three stages has been described.
Links can be split.

With respect to the divided links, a link group can be determined based on the distance from the link to be predicted and the number of steps in the tree, excluding links far from the link to be predicted. In this way, the links in FIG. 6 can be divided into link groups as in FIG.

Note that the link classification method shown in FIG. 5 uses a floppy disk, CD-ROM, magneto-optical disk,
The program can be stored as a link classification program in a recording medium such as a semiconductor memory, and can be read and executed by a CPU of a computer.

[0045]

As described above, the present invention uses not only the link to be predicted and the adjacent link connected thereto but also the traffic information of a wide range of links around the link to be predicted. The fact that the traffic information input to the prediction means can be reduced and prediction can be performed can also accurately and efficiently predict traffic information that changes in a complicated manner.

[Brief description of the drawings]

FIG. 1 is a block diagram of a traffic information prediction device according to a first embodiment of the present invention.

FIG. 2 is a block diagram of a traffic information prediction device according to a second embodiment of the present invention.

FIG. 3 is a block diagram of a traffic information prediction device according to a third embodiment of the present invention.

FIG. 4 is a block diagram of a traffic information prediction device according to a fourth embodiment of the present invention.

FIG. 5 is a flowchart of a link classification method according to an embodiment of the present invention.

FIG. 6 is a diagram illustrating an example of a link.

FIG. 7 is a diagram in which the links in FIG. 6 are divided into link groups.

FIG. 8 is a diagram showing link division.

FIG. 9 is a diagram showing link division.

FIG. 10 is a diagram showing link division.

FIG. 11 is an explanatory diagram of a conventional traffic information prediction method.

FIG. 12 is a diagram for explaining regions and links in a conventional traffic information prediction method.

[Explanation of symbols]

11A, 11B selectors 12 _{1 to} 12 _m link group traffic information calculator 13 _{1 to} 13 _m storage unit (link group traffic information) 14 storage unit (traffic information of link to be predicted) 15 ₁ to 15 _m storage unit (adjacent link) 16A, 16B Prediction unit 21 Traffic information of links 22 _{1 to} 22 _m Traffic information of links in link group 23 Traffic information of link to be predicted 24 _{1 to} 24 _m Traffic information of link group 25 _{1 to} 25 _m adjacent Link traffic information 26 Prediction information 31 Input device 32 _{1 to} 32 _m , 33, 34 _{1 to} 34 _m Storage device 35 Output device 36A, 36B Recording medium 37 Data processing device 41 to 45 steps

Continuing on the front page (72) Inventor Hitoshi Mohri Nippon Telegraph and Telephone Corporation, 3-19-2 Nishi-Shinjuku, Shinjuku-ku, Tokyo (72) Inventor Tsutomu Horikoshi 3-2-1, Nishi-Shinjuku, Shinjuku-ku, Tokyo Nippon Telegraph and Telephone Telephone Co., Ltd. (72) Inventor Satoshi Suzuki 3-19-2 Nishi-Shinjuku, Shinjuku-ku, Tokyo F-term in Japan Telegraph and Telephone Co., Ltd. 5H180 AA01 BB13 BB19 EE02

Claims (12)

[Claims] 1. A link is divided into groups based on a connection relationship between adjacent links connected to a link to be predicted, traffic information of each of the divided link groups is obtained, and traffic information of the obtained link group and a link to be predicted are calculated. From traffic information of
A traffic information prediction method for predicting traffic information of a link to be predicted. 2. A link is divided into groups based on a connection relationship between adjacent links connected to a link to be predicted, traffic information of each of the divided link groups is obtained, and traffic information of the obtained link group and a link to be predicted are calculated. A traffic information prediction method for predicting traffic information of a link to be predicted from traffic information of an adjacent link and traffic information of an adjacent link. 3. A shortest route to each node is determined starting from one node at the end of the link to be predicted, and a first route that does not pass through the link to be predicted is extracted from the determined optimal route. Starting from the other node at the end of the target link, a shortest path to each node is determined, a second path that does not pass through the prediction target link is extracted from the determined optimum paths, and a prediction target link is extracted from the extracted path. A link dividing method for dividing a link into groups for each link included in a route passing through each adjacent link. 4. A shortest route to each node is determined starting from one node at the end of the link to be predicted, and a first route that does not pass through the link to be predicted is extracted from the determined optimal route. Starting from the other node at the end of the target link, a shortest path to each node is obtained, and a second path that does not pass through the prediction target link is extracted from the obtained optimum paths, and the first path and the second path are extracted. If there is a common link in the route, extract the route excluding one of the common links, and group the links from the extracted route into groups for each link included in the route passing through each adjacent link of the prediction target link. Link splitting method to split. 5. Extracting links that are not included in an extracted route from links to be divided into groups, and incorporating the extracted links into a group of links to which one of the nodes at both ends of the extracted links belongs. 5. The link dividing method according to claim 3, wherein the link is divided into groups. 6. Selection means for inputting traffic information of a link, selecting and outputting traffic information of the link for each link group, and outputting traffic information of a link to be predicted, provided for each link group, Link group traffic information calculation means for inputting traffic information of links in the link group output from the selection means and calculating traffic information of the link group based on the traffic information, provided for each link group; A first storage unit that stores a plurality of pieces of traffic information of the link group calculated by the link group traffic information calculation unit of the link group from the present to the past;
Storage means for storing a plurality of past traffic information of the link to be predicted output from the selection means from the present time to the past, and link groups stored in the first storage means and the second storage means, respectively. Each time the traffic information of the prediction target link and the traffic information of the link to be predicted are input, a plurality of pieces of traffic information of each link group from the first storage means and the second storage means,
A traffic information prediction device having prediction means for reading a plurality of pieces of traffic information of a prediction target link and predicting traffic information of the prediction target link based on all or a part of the traffic information. 7. A means for inputting traffic information of a link, selecting and outputting traffic information of the link for each link group, and outputting traffic information of a link to be predicted and an adjacent link, A link group traffic information calculating means for inputting traffic information of a link in the link group output from the selecting means and calculating traffic information of the link group based on the traffic information; A first group for storing a plurality of pieces of traffic information of the link group calculated by the link group traffic information calculation means of the link group from the present to the past;
Storage means for storing a plurality of pieces of traffic information of the prediction target link output from the selection means in the past from the present, and provided for each adjacent link and output from the selection means. A third storage unit that stores a plurality of pieces of traffic information of adjacent links from the present to the past, and the first storage unit, the second storage unit, and the third storage unit respectively include link group traffic information and a prediction target. Each time the traffic information of the link and the traffic information of the adjacent link are input, a plurality of traffic information of each link group from the first storage unit, the second storage unit, and the third storage unit are calculated. Prediction means for reading a plurality of traffic information of a target link and a plurality of traffic information of each adjacent link, and predicting the traffic information of the prediction target link based on all or a part of the traffic information Traffic information prediction apparatus having. 8. Traffic information of a link is input, traffic information of the link is selected and output for each link group, traffic information of a link to be predicted is output, and traffic information of the link to be predicted is predicted. A selection process of storing the traffic information of the target link in a second storage unit that stores a plurality of traffic information from the present to the past; and providing traffic information of a link in the link group output from the selection unit, provided for each link group. The traffic information of the link group is calculated based on the traffic information, and the traffic information is provided for each link group. Link group traffic information calculation processing to be stored; traffic information of a link group and traffic of a link to be predicted in the first storage means and the second recording means, respectively. Each time a report is input, a plurality of traffic information of each link group and a plurality of traffic information of the link to be predicted are read from the first storage means and the second storage means, respectively, and all or A recording medium storing a traffic information prediction program for causing a computer to execute a prediction process of predicting traffic information of the prediction target link based on a part thereof. 9. Traffic information of a link is input, traffic information of the link is selected and output for each link group, and traffic information of a link to be predicted and an adjacent link is output.
The traffic information of the link to be predicted is stored in second storage means for storing a plurality of pieces of traffic information of the link to be predicted from the present to the past, and the traffic information of the adjacent link is provided for each adjacent link, and A selection process of storing the traffic information of the present time in a third storage unit that stores a plurality of past traffic information from the present time; and inputting the traffic information of the links in the link group, which is provided for each link group and output from the selecting unit, The traffic information of the link group is calculated based on the traffic information, and the traffic information is stored in a first storage unit that is provided for each link group and stores a plurality of pieces of traffic information of the link group from the present to the past. Link group traffic information calculation processing, and link group traffic information and prediction stored in the first storage unit, the second recording unit, and the third storage unit, respectively. Each time the traffic information of the target link and the traffic information of the adjacent link are input, a plurality of pieces of traffic information of each link group from the first storage unit, the second storage unit, and the third storage unit, The computer performs a prediction process of reading a plurality of traffic information of the prediction target link, a plurality of traffic information of each adjacent link, and predicting the traffic information of the prediction target link based on all or a part of the traffic information. A recording medium on which a traffic information prediction program to be executed is recorded. 10. A process of obtaining a shortest route to each node starting from one node at the end of the link to be predicted, and extracting a first route not passing through the link to be predicted from the obtained shortest routes; Starting from the other node at the end of the link to be predicted as a start point, a shortest path to each node is determined, and a second path that does not pass through the link to be predicted is extracted from the determined optimal path. A recording medium recording a link division program for causing a computer to execute a process of dividing a link into groups for each link included in a path passing through each adjacent link of a prediction target link. 11. A process of obtaining a shortest route to each node starting from one node at the end of the link to be predicted, and extracting a first route not passing through the link to be predicted from the obtained shortest route; Starting from the other node at the end of the link to be predicted as a start point, a shortest path to each node is determined, and a second path that does not pass through the link to be predicted is extracted from the determined optimal path; When there is a link common to the second route, a route excluding one of the common links is extracted, and a link is extracted from the extracted route for each link included in a route passing through each adjacent link of the prediction target link. Recording medium for storing a link division program for causing a computer to execute a process of dividing a program into groups. 12. Extracting links that are not included in an extracted path from links to be divided into groups, and incorporating the extracted links into a group of links to which one of the nodes at both ends of the extracted links belongs. The recording medium according to claim 10, wherein the link is divided into groups by: