KR20180121241A - Prediction system for public transport travel time using big data - Google Patents

Abstract

The present invention relates to a public transportation travel time prediction system using big data, capable of providing more accurate travel time information to a public transportation user by linking the big data such as historical data of a bus that is public transportation, unexpected and weather information, real-time driving information and route characteristics, and promoting the convenience of the public transportation user by providing real-time travel time and prediction information of the bus through the provided information. The public transportation travel time prediction system using big data is implemented by including a data collection unit for making the big data by collecting information required for bus travel time prediction and a data processing unit for analyzing and processing the data collected in the data collection unit and calculating a bus travel time through real-time prediction.

Description

{Prediction system for public transportation travel time using big data}

The present invention relates to a public transit time predicting system using Big Data, and particularly relates to a public transit time predicting system using Big Data, This paper proposes a system for predicting the transit time of public transit using Big Data which provides more precise traffic time information to the traffic users and provides the information about the time required for real time traffic and the prediction information to improve the convenience of the users. will be.

Currently, most municipalities have established Bus Management System (BMS) and Bus Information System (BIS) to enhance the convenience of using public transportation (especially buses). In addition, traffic information prediction technology utilizing big data such as past road speed information is being developed.

The bus information currently provided only provides real-time arrival information, location information, and base information (bus stop and route information) of the bus.

In addition, the time information provided for the existing bus users does not reflect the surrounding traffic situation because it calculates the simple average bus speed and the distance based time, As a result, the reliability of the bus users is degraded, and inaccurate arrival time information is provided when an event such as an unexpected situation or an abnormal weather occurs, so that the bus operator can not perform a flexible dispatch.

In order to solve such a problem, conventional technologies such as the following Patent Documents 1 to 2 have been proposed.

The prior art disclosed in Patent Document 1 predicts a traffic predicted time of a section (between a stop and a stop) by learning a traffic volume pattern of a nearby road, taking into account the traffic volume of the surrounding road of the section, An apparatus and method for predicting a stop arrival time of a bus by adding a travel prediction time of a remaining section are provided.

In the prior art disclosed in Patent Document 2, real-time traffic information is used in the case of short-term traffic prediction, and in the case of long-term traffic prediction, Thereby providing optimal information on the traffic situation to the user.

(Apparatus and method for predicting the arrival time of a bus by learning a traffic pattern of a nearby road) (Korean Patent No. 10-1057223) (Method and system for predicting traffic situation and providing dynamic route guidance using past and real time traffic information) Korean Patent Registration No. 10-0702852 (Apr. 04, 2007)

However, the conventional bus information providing system has a disadvantage in that it can not provide bus travel time prediction / forecast in connection with BIS / BMS information, unexpected situation, weather information,

The prior art disclosed in Patent Document 1 has a disadvantage in that it can not take into account a weather situation or an unexpected situation, and the prior art disclosed in Patent Document 2 provides traffic situation information for a driver, It is impossible to provide information to the user.

SUMMARY OF THE INVENTION Accordingly, the present invention has been made to solve the above-mentioned problems occurring in the prior art, and it is an object of the present invention to provide a system and method for linking big data such as historical data of buses, unexpected and weather information, Providing more precise traffic time information to the users, providing the bus time information and forecast information through the bus, and providing the convenience of the public transportation users with better convenience. It has its purpose.

Another object of the present invention is to construct a link system between local government real-time bus information, history information and traffic card data to provide a real-time arrival information and a public transit prediction / forecast service so that a bus operator can efficiently adjust the bus dispatch interval And to provide a public transit time predicting system using Big Data.

Another object of the present invention is to provide a method and system for estimating the time required to travel by using public transportation, Thereby providing a travel time predicting system.

Another object of the present invention is to provide a public transit time predicting system that utilizes big data for planning basic data for government information and local information bus information system information elevation and public transportation policy establishment .

According to another aspect of the present invention, there is provided a public transit time predicting system using Big Data according to the present invention, comprising: a data collecting unit collecting information necessary for predicting a bus travel time and converting it into big data; And a data processing unit for analyzing and processing the data collected by the data collecting unit and calculating the bus travel time through real-time prediction.

The public transit time predicting system using the Big Data according to the present invention may further include a service provision device capable of displaying information calculated through the data processor.

The information collected in the data collecting unit may be at least one of historical data, real-time BIS data, weather data, unexpected data, traffic communication information, BIS / BMS-based information, and route characteristic information.

The data processing unit may include a data linkage analysis and processing server for analyzing and processing the data collected through the data collection unit. A bus travel time predicting server for calculating a bus travel time through information processed through the data linkage analysis and processing server; A data storage module for storing information collected or generated through the data collection unit, the data linkage analysis, the processing server, and the bus travel time prediction server; And a WEB server for externally transmitting information stored in the data storage module.

The bus travel time predicting server may include a short-term predicting unit for predicting a short-term travel time by analyzing real-time bus traffic information using a kNN pattern matching algorithm; A prediction result determination unit that determines a travel time prediction result predicted by the short term prediction unit; And a prediction information providing unit for providing the prediction information finally determined through the prediction result determination unit.

Wherein the short-term prediction unit comprises: a cluster generating unit for generating a cluster based on real-time bus traffic information including real-time bus position information and travel time; A pattern extracting unit for extracting a plurality of patterns before a departure stop from a cluster generated by the cluster generating unit, a pattern retrieving unit for retrieving a latest travel history from a viewpoint and extracting a pattern; A pattern matching unit for extracting a pattern value extracted by the pattern extracting unit and an operation record closest to the Euclidean distance; And a final predicted traveling time calculating unit for calculating a final predicted traveling time by calculating a median value of the travel time per link for the pattern extracted by the pattern matching unit.

The final predicted traveling time calculating unit calculates the actual physical distance for each link and calculates the first forecast value by dividing the traveling time by the bus average speed according to the physical distance and calculates the final predicted traveling time according to the number of links from the target link to the starting point Calculating a second predicted value, and then adding the first and second predicted values to calculate a final predicted traveling time.

Also, it is preferable that the service providing apparatus transmits information transmitted through the WEB server to an operator, a user, a stop display device, and the like.

According to the present invention, it is possible to provide more accurate information to the public transportation users by combining historical data of buses, unexpected situations and weather information, real-time bus information, traffic communication information, BIS / BMS-based information, .

In addition, according to the present invention, it is possible to improve the convenience of the public transportation user by providing the bus real-time traveling time information and prediction information.

In addition, according to the present invention, it is possible to improve the efficiency of the existing infrastructure by establishing a linkage system such as local government real-time bus information, history information, weather information and erroneous information, and provide bus prediction service in addition to real- Therefore, the reliability of the bus information can be improved and the bus utilization can be achieved.

Brief Description of the Drawings Fig. 1 is a block diagram of a public transit time predicting system using Big Data according to an embodiment of the present invention; Fig.
FIG. 2 is an explanatory diagram illustrating an application state of a public transit time predicting system using Big Data according to an embodiment of the present invention;
FIG. 3 is a block diagram of a public transit time predicting system using Big Data according to the present invention;
FIG. 4 is a block diagram of an embodiment of the short term prediction unit of FIG. 3;
5 is a route path example of the Goyang city 1000 bus,
FIG. 6 is a result of predicting the travel time when FIG. 5 is applied to the present invention,
7 is a route path example of the bus 72 of Goyang City,
FIG. 8 is a result of the travel time prediction applied to FIG. 7,
FIG. 9 is a diagram illustrating an example of arrival time setting and travel time prediction result in the present invention,
10 is a diagram illustrating an example of a departure time setting example and a travel time prediction result in the present invention.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a configuration diagram of a public transit time predicting system using Big Data according to a preferred embodiment of the present invention, and FIG. 2 is an explanatory diagram of a utilization state of a big data-based bus transit time predicting system according to an embodiment of the present invention .

Referring to FIG. 1, a public transit time predicting system using Big Data according to an embodiment of the present invention includes a data collecting unit 110, a data processing unit 120, and a service providing apparatus 130 .

1 and 2, the data collecting unit 110 collects historical data, real-time BIS

Data, meteorological data, unexpected data, traffic communication information, BIS / BMS-based information, and route characteristic information, and data is linked with public institutions providing the corresponding information.

The data processing unit 120 analyzes data collected through the data collecting unit 110 and processes the data through the data link analysis and processing server 121 and the data link analysis and processing server 121 A bus travel time predicting server 122 for calculating a bus travel time, information collected or generated through the data collecting unit 110, the data linkage analysis and processing server 121, and the bus travel time predicting server 122 And a WEB server 124 for transmitting information stored in the data storage module 123 to the outside.

3, the bus travel time predicting server 122 includes a short-term prediction unit 122a for analyzing real-time bus traffic information using a kNN pattern matching algorithm to predict a short-term travel time, a short- And a prediction / forecast information providing unit 122d for providing prediction / forecast information finally determined through the prediction result determining unit 122c do.

4, the short-term prediction unit 122a includes a cluster generating unit 142 for generating a cluster based on real-time bus traffic information including real-time bus position information and travel time, A pattern extracting unit 143 for extracting a plurality of patterns before the departure stop from the cluster generated in the clustering unit 142, a pattern extracting unit 143 for retrieving the most recent operation history from the inquiry point and extracting a pattern, A pattern matching unit 144 for extracting the extracted pattern value and the driving record closest to the Euclidean distance, and a pattern matching unit 144 for calculating a final predicted traveling time by calculating a median value of the travel time of each link on the pattern extracted by the pattern matching unit 144 And a predicted traveling time calculating section 145. [

The service providing apparatus 130 may transmit information transmitted through the WEB server 124 to an operator, a user, a stop display device, or the like. That is, the user can confirm the information through the Web or the mobile terminal, the operator can confirm the information through the management server or the monitoring device of the operator, and can confirm the information through the display device at the stop.

The operation of the public transit time predicting system using the Big Data according to the present invention will be described in detail as follows.

First, basic data for predicting the traveling time of public transportation is collected as big data through the data collection unit 110. FIG. Here, the public transportation includes buses, taxis, trains, trains and the like. However, the present invention will be described only for buses for convenience of explanation. However, the present invention is not limited to this and can be equally applied to all public transportation. Those skilled in the art will appreciate it.

The basic data includes at least one of historical data, real-time BIS data, weather data, unexpected situation data, traffic communication information, BIS / BMS-based information, and route characteristic information.

Next, the data linkage analysis of the data processing unit 120 and the information linking unit 121a in the processing server 121 extract information for predicting the transit time in connection with the data collecting unit 110. [ The information for predicting the transit travel time thus extracted is appropriately processed through the information processing unit 121b. For example, unlike existing link-node schemes that define intersections as nodes and intersection roads as links, we define the stopping point as a node and the interstate road as a link so that the link travel time and the delay time at the node can be calculated Data is analyzed and processed.

In this case, we can apply the algorithm to predict the traffic situation through kNN (k-Nearest Neighbor) analysis method by using BMS and BIS. . Concretely, the road characteristic of the intermittent flow (city road) and the continuous flow (high speed road) together with the characteristic of the bus travel time is analyzed and processed through the history data of the bus lane.

In addition, the lag in the node (stopping point) does not exist in the general traffic flow. The variables influencing the stay time at the stopping point are the number of the stopping vehicle, the number of the passengers and the departing person, the time zone And the data linkage analysis and processing server 121. [0064] FIG.

Next, the bus travel time predicting server 122 can estimate the bus travel time by calculating the travel time of the link and the stay time at the node by calculating the travel time of the bus through the collected and processed information. At this time, it is possible to increase the accuracy of the actual bus travel time by considering the data on the weather condition, the unexpected situation, and the route characteristic.

For example, the bus travel time predicting server 122 predicts the short-term travel time by analyzing real-time bus traffic information using a k-NN pattern matching algorithm using the short-term prediction unit 122a.

More specifically, it is as follows.

As shown in FIG. 4, the cluster generating unit 142 generates a cluster based on real-time bus traffic information including real-time bus position information and traffic time. Then, the pattern extracting unit 143 selects several patterns before the departure stop from the cluster generated by the cluster generating unit 142 as a target link, and inquires the most recent travel history from the point of view to derive a pattern. That is, a pattern corresponding to the travel time pattern passed in the nearest past at the time of inquiry is derived from the past history data. Next, the pattern matching unit 144 calculates the pattern value and the Euclidean distance extracted by the pattern extracting unit 143, and sets the five most recent driving records as a matching pattern. That is, the total travel time of the five target links to be extracted in the last 6 months of the target link of the same day and the same time of day on the basis of the view point is calculated as a whole.

The Euclidean distance with the pattern for extracting the pattern to be matched uses the following Equation (1).

m ₁ = _i- (link pattern value of _{i-1), r i-} 1 = ( link record actual value of the i-1)

Then, the final predicted traveling time calculating unit 145 calculates a median value of the travel time per link for the pattern extracted by the pattern matching unit 144, and calculates the final predicted traveling time.

Here, the final predicted traveling time calculating unit 145 calculates the actual physical distance for each link, and calculates the first predicted value according to the calculated physical distance. In other words, if the calculated physical distance is less than 1 km, it is a short distance. Therefore, the travel time is calculated by dividing the bus average speed to the physical distance, and converted into the segment. If the calculated physical distance is 1 km or more, the second forecast value is calculated as follows according to the distance from the start point. That is, the second forecast value is calculated according to the number of links from the target link to the departure point. In other words, if the number of links from the target link to the departure point exceeds 5, the target value of the target link uses the reference value used in the mid-and-long term prediction. If the number of links from the target link to the start point is 5 or less, The predicted value calculated through pattern matching is used. Next, the first predicted value and the second predicted value are added to calculate a final predicted traveling time.

Next, the prediction result determination unit 122c verifies the travel time prediction result predicted by the short term prediction unit 122a. It is desirable to verify the prediction results by verifying the short-term prediction results.

In addition, the prediction information providing unit 122d provides the service providing apparatus 130 with the prediction information finally determined through the prediction result determining unit 122c.

The service providing apparatus 130 may transmit information transmitted through the WEB server 124 to an operator, a user, a stop display device, or the like. That is, the user can confirm the information through the Web or the mobile terminal, the operator can confirm the information through the management server or the monitoring device of the operator, and can confirm the information through the display device at the stop.

Here, the travel time prediction information provided by the present invention can provide information such as an optimal route and travel time / cost when using public transportation, automobiles, bicycles, and walking on a web or application basis, such as Naver maps and navigation.

In addition, the following maps can be used to provide information such as the optimal route and travel time / cost for public transportation, automobiles, bicycles, and walking based on the web and applications.

It is also possible to provide travel time information using the SKT T map public transportation application. For example, SK T Map 4.0 is a typical application for providing personal terminal information related to traffic information. SK Planet provides mobile-based traffic information service that provides fast route search, real-time public transportation information, integrated search function, etc. In the latest version of TeMap 4.0, 'Time Machine' Provides predicted travel time information of the point.

Fig. 5 shows the route of the 1000th bus route in Goyang (Ilsan-Shinchon), and Fig. 6 shows the result of predicting travel time by applying the present invention to the route.

The number of predictive links is 13 (Interpretation - Yonsei University), and the number of links (LINK) for matching is 3 (Interpretation - Interpretation).

It can be seen that the predicted time and the actual travel time error at the time of real-time prediction (short-term prediction) are almost the same except for the working time on Monday, and it is understood that the mid- and long-term prediction is 10% or less even in consideration of all days of the week and commute time .

7 is a route of Bus No. 72 in Goyang City (Ilsan-Shinchon), and Fig. 8 is a result of predicting travel time by applying the present invention to the route.

The number of predictive links is 41 (Interpretation - Yonsei University), and the number of links (LINK) for matching is 5 (Interpretation - Interpretation).

It can be seen that the predicted time and the actual travel time error at the time of real-time prediction (short-term prediction) are almost the same except for the working time on Monday, and it is understood that the mid- and long-term prediction is 10% or less even in consideration of all days of the week and commute time .

FIG. 9 is an example of the arrival time setting example and the travel time prediction result calculated by applying the present invention, FIG. 10 is an example of the departure time setting example and the travel time prediction result calculated by applying the present invention.

The bus travel time prediction result can be provided to the user in real time prediction and history based prediction. The real time prediction provides a time taken to move a specific route based on the travel log data acquired in real time. And predicts the bus travel time when a specific event or an unexpected event occurs, thereby providing the corresponding information.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, It is obvious to those who have.

The present invention is effectively applied to a technique for predicting the travel time of a public bus.

110: Data collecting unit
120:
121: Data linkage analysis and processing server
122: bus travel time prediction server
123: Data storage module
124: WEB server
130: Service providing device

Claims (8)

As a system for predicting the travel time of public transportation by utilizing big data,
A data collection unit for collecting information necessary for bus travel time prediction and converting the collected information into big data; And
And a data processing unit for analyzing and processing data collected by the data collecting unit and calculating a bus travel time through real-time prediction.
The system of claim 1, further comprising a service providing device capable of displaying information calculated through the data processor.
The information collected in the data collecting unit is at least one of history data, real-time BIS data, weather data, unexpected data, traffic communication information, BIS / BMS-based information, Data Transmission Time Estimation System.
The data processing system of claim 1, wherein the data processing unit analyzes and processes data collected through the data collecting unit; A bus travel time predicting server for calculating a bus travel time through information processed through the data linkage analysis and processing server; A data storage module for storing information collected or generated through the data collection unit, the data linkage analysis, the processing server, and the bus travel time prediction server; And a WEB server for transmitting information stored in the data storage module to the outside.
The bus travel time predicting server of claim 4, wherein the bus travel time predicting server comprises: a short-term predicting unit for predicting a short-term travel time by analyzing real-time bus traffic information using a kNN pattern matching algorithm; A prediction result determining unit that determines a travel time prediction result predicted by each of the short term prediction units; And a prediction information providing unit for providing the prediction information finally determined through the prediction result determination unit.
The system of claim 5, wherein the short-term prediction unit comprises: a cluster generator for generating a cluster based on real-time bus traffic information including real-time bus position information and travel time; A pattern extracting unit for extracting a plurality of patterns before the departure stop from the cluster generated by the cluster generating unit, a pattern retrieving unit for retrieving the latest travel history from the inquiry time point and extracting a pattern; A pattern matching unit for extracting a pattern value extracted by the pattern extracting unit and an operation record closest to the Euclidean distance; And a final predicted travel time calculating unit for calculating a median value of the travel time per link for the pattern extracted by the pattern matching unit and calculating a final predicted travel time.
The final predicted traveling time calculating unit calculates the actual physical distance for each link, calculates the first predicted value by dividing the traveling time by the bus average speed according to the physical distance, Calculating a second predicted value according to the number of vehicles, and adding the first and second predicted values to calculate a final predicted traveling time.
[2] The system of claim 2, wherein the service providing device transmits information transmitted through the WEB server to an operator, a user, and a stop display device.

Images