CN105303245A - Traffic analysis system and traffic analysis method - Google Patents

Abstract

The present invention relates to a traffic analysis system and a traffic analysis method. The traffic analysis system is configured to collect mobile daily records of passengers using vehicles and perform analysis by using collected mobile daily records. The traffic analysis system comprises a path presumption part configured to presume a plurality of mobile paths of the mobile daily records from a departure station to a destination station, a utilization presumption part configured to determine whether the mobile daily records use the mobile daily record of the path arrived at the destination station passing through a previous station when the previous station of the destination station is the same in the paths having a preset threshold value larger than the plurality of mobile paths; and a real schedule presumption part configured to presume the real schedule of each path through the get-off time of the mobile daily record of each path generated by the path presumption part.

Description

Traffic analysis system and traffic analysis method

technical field

The invention relates to a traffic analysis system and a traffic analysis method. More specifically, it relates to a traffic analysis system, a traffic analysis program, and a traffic analysis method for estimating arrival times and departure times of trains and the like from user movement logs.

Background technique

Aiming at improving the convenience and efficiency of public transportation, various measures are being implemented by transportation companies. For example, in recent years, it is possible to estimate the congestion rate and flow prediction by estimating the travel route of passengers using the history of IC tickets in general transportation systems. Timetable (schedule) information is necessary to accurately estimate the movement route of passengers. Transportation companies have a pre-set schedule, but in practice it is difficult to operate trains and cars according to the plan. Due to the influence of accidents and traffic congestion, small-scale delays often occur, and there are many cases where deviations from the plan must be carried out. running situation. Therefore, it is considered that more accurate analysis can be performed by using an actual timetable in which the actual arrival time and departure/arrival time are counted instead of the planned travel timetable when estimating the travel route of the passenger. However, there is a problem that in order to collect the actual time table mechanically, special devices must be installed on vehicles, lines, roads, etc., and extra cost and time are required.

Patent Document 1 discloses an operation management system with a train delay prediction display function including an actual timetable generation means.

Also, Patent Document 2 discloses a system for estimating an actual timetable based on a distribution of the number of people getting off generated from a travel log of passengers estimated to be one-way route travelers who do not transfer.

Patent Document 1: Japanese Patent Laid-Open No. 2000-1168

Patent Document 2: PCT/JP2012/076750

The technology described in Patent Document 1 is based on the premise that train online information, station congestion information, garage information, weather information, and other information related to current train operation can be received in order to generate an actual timetable. The technique for generating the actual timetable is not documented.

The technology described in Patent Document 2 estimates the actual timetable based on the distribution of the number of people getting off the bus, but in order to eliminate the movement logs of passengers who got off using other routes, only those who are estimated to be moving on one-way routes without transferring are used. Passenger's movement log. Therefore, there is a problem that the movement log used for estimation becomes smaller than the log of the passenger who actually used the route and got off.

In addition, the peak of the distribution of the number of people getting off the train generated by the above-mentioned means has a missing or false peak when comparing the actual arrival time of the train, so it is not easy to estimate the actual time table from the distribution of the number of people getting off the train.

Contents of the invention

The present invention was made in view of the above circumstances, and an object of the present invention is to estimate the arrival time and departure time of trains and the like with higher accuracy based on user movement logs.

A typical example of the means for solving the problems of the present invention is a traffic analysis system that estimates the actual timetable of a predetermined route using a movement log of a passenger using a vehicle. The traffic analysis system has: a route estimating unit that estimates a plurality of travel routes from a departure station to a destination station in the travel log; and a route estimation unit that uses a route that passes a station immediately before the arrival station of the route among the plurality of travel routes as predetermined When the above-mentioned movement log is above the threshold value, it is estimated that the above-mentioned movement log is a log of arriving at the above-mentioned arrival station by the above-mentioned route; actual schedule.

Alternatively, a traffic analysis program for estimating an actual timetable of a predetermined route using a travel log of passengers using a vehicle, causing a computer to execute a step of estimating a plurality of travel routes from a departure station to an arrival station in the above travel log The step of: when the route of the previous station of the above-mentioned arrival station via the above-mentioned route among the above-mentioned multiple movement routes is more than a predetermined threshold value, the step of inferring that the above-mentioned movement log is a log of the above-mentioned arrival station by the above-mentioned route; according to the above-mentioned utilization A step of estimating the actual timetable by the route estimating unit estimating the disembarkation time using the movement log of the route.

Alternatively, a traffic analysis method for estimating an actual timetable of a predetermined route using a movement log of a passenger using a vehicle includes the step of estimating a plurality of movement paths from a departure station to an arrival station in the movement log. ; When the route of the previous station of the above-mentioned arrival station via the above-mentioned route among the above-mentioned multiple moving routes is more than a predetermined threshold value, the step of estimating that the above-mentioned movement log is a log of using the above-mentioned route to the above-mentioned arrival station; A step of estimating the actual timetable by estimating the disembarkation time using the movement log of the route.

According to one embodiment of the present invention, the arrival time and departure time of a train or the like can be estimated with higher accuracy from the user's movement log.

Description of drawings

FIG. 1 is a system configuration diagram of a traffic analysis system according to an embodiment of the present invention.

Fig. 2 is a diagram illustrating the structure of transportation system IC card data according to the embodiment of the present invention.

3 is a diagram illustrating the structure of master data storing basic information such as stations and routes according to the embodiment of the present invention.

FIG. 4 is a diagram illustrating the structure of movement log data according to the embodiment of the present invention.

FIG. 5 is a diagram illustrating a data structure of a schedule schedule according to the embodiment of the present invention.

FIG. 6 is a flowchart showing a movement log generation processing procedure according to the embodiment of the present invention.

FIG. 7 is a flowchart showing the procedure of the used route estimation process according to the embodiment of the present invention.

FIG. 8 is an explanatory diagram showing an example of route estimation according to the embodiment of the present invention.

FIG. 9 is an explanatory diagram showing an example of the estimated result of the used route according to the embodiment of the present invention.

FIG. 10 is a flowchart showing the actual schedule estimation processing procedure according to the embodiment of the present invention.

FIG. 11 is an explanatory diagram showing an example of the estimated result of the route used and the presentation of the estimated result of the planned timetable and the actual timetable according to the embodiment of the present invention.

12 is a flowchart showing another example of the actual schedule estimation processing procedure according to the embodiment of the present invention.

13 is an explanatory diagram showing an example of representative alighting time estimation processing according to the embodiment of the present invention.

14 is an explanatory diagram showing an example of calculation processing of the average required time between stations according to the embodiment of the present invention.

15 is an explanatory diagram showing another example of actual schedule estimation processing according to the embodiment of the present invention.

16 is a flowchart showing an example of on-board train estimation processing according to the embodiment of the present invention.

FIG. 17 is an explanatory diagram showing an example of a seat priority route according to the embodiment of the present invention.

FIG. 18 is a detailed explanatory diagram showing on-board train estimation processing according to the embodiment of the present invention.

FIG. 19 is an explanatory diagram showing an example of an estimated result of a boarding train according to the embodiment of the present invention.

Explanation of symbols

101: user; 102: reading terminal; 103: portable terminal; 104: network; 105: server group; 107: traffic analysis system; 111: data server; 112: computing server; 113: information sending server; 114: network; 115, 117: transportation enterprise; 121: data storage department; 122: traffic system IC card data; 123: master data; 124: movement log data; 125: planning timetable; 126: actual timetable, route preference mode data; 130 : I/F; 131: CPU; 132: memory; 133: storage unit; 134: travel log generation program; 135: route estimation program; 136: utilization route estimation program; 137: actual timetable estimation program; 138: boarding Train estimation program; 139: data storage unit; 141: display screen creation program; 142: information transmission program; 145: I/F; 146: CPU; 147: memory; 201: log ID; 202: user ID; 203: station /bus station ID; 204: time of use; 205: use category; 300: main station/bus station; 301: station/bus station ID; 302: station/bus station name; 303: holding company; 304: location; 305 : latitude and longitude; 310: main route; 311: route ID; 312: route name; 313: operating company; 314: route type; 320: main station/bus station/route relationship; 321: route ID; 322: station/bus Station ID; 323: order; 324: category; 325: required time from the starting point; 401: log ID; 402: user ID; 403: ride date and time; 404: get off date and time; : boarding station/bus station ID1; 407: getting off station/bus station ID1; 408: boarding station/bus station ID2; 409: getting off station/bus station ID2; 501: schedule ID; 502: route ID; 503: Parking station; 504: arrival time; 505: departure time; 601-1005: steps; 1101: screen; 1102: pull-down menu; 1103: display screen; Table estimation result; 1107: cursor; 1201-1207: steps; 1501: representative alighting time; 1502: representative alighting time of the next stop; 1503: time difference; 1504: interpolation; 1601-1605: steps; 1701: boarding station; 1702: get off the station; 1703: route; 1704: departure station; 1705: seat priority route; 1801: boarding time; 1802: departure time; 1803: next departure time; 1804: boarding station; 1805: departure Station time; 1806: Arrival time; 1807: Departure time; 1901: Screen; 1902: Station; 1903: picture; 1904: diagram.

detailed description

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a system configuration diagram of a traffic analysis system according to an embodiment of the present invention.

In recent years, most users (101) who use vehicles use non-contact IC cards or portable terminals (103) that have the same function, and pass through the ticket gates for using vehicles and the reading terminals (102) installed in vehicles. . Data acquired by these ticket gates and in-vehicle terminals are sent to server groups (105) managed by respective transportation companies via a network (104).

The traffic analysis system (107) is composed of a data server (111), a calculation server (112), and an information sending server (113), and stores and aggregates the use data of a non-contact IC card or a portable terminal (103) with equivalent functions. The obtained mobile data is analyzed and processed. In addition, descriptions of functions and structures of non-contact IC cards, ticket gates, and information processing techniques that are not directly related to the description of the present invention are omitted.

If a user (101) holding a non-contact IC card (103) passes through the ticket gate, the user ID for identifying the IC card (103) and positional information including passing time and date are stored in the ticket gate (102), Stored in the server (105) managed by the transportation company as raw data. These data are transmitted to the data server (111) via the network (104) at an appropriate timing such as every hour or every day while storing these data. A traffic analysis system (107) composed of a data server (111), a computing server (112), and a server group of an information sending server (113) is connected to the network (104), and can communicate with users and traffic companies (115, 117). communication. Therefore, for example, the transportation company (115) can make the traffic analysis system (107) cooperate with other systems such as the operation management system and the IC card data management system. In addition, in this embodiment, the server group of the data server (111), the calculation server (112), and the information delivery server (113) has been described, but it can be configured that one or more servers perform the functions of these server groups. .

The data server (111) receives user data read by an IC card reading terminal such as a ticket gate via a network (104), and records it in a data storage unit (121) in the server. The collected and stored data include traffic system IC card data (122), basic master data (123) related to stations/bus stations or routes, and the like. Then store the mobile log data (124), the planned timetable (125), the actual timetable (126) generated by the traffic analysis system (107) once processed to the traffic system IC card data (122), etc. Route preference pattern data (127) etc. used for estimating the route of the passenger. Regarding the basic master data (123) related to the station and the route, when there is a change or when an update is performed, it is appropriately input from the outside of the system and updated/recorded.

In the calculation server (112), processing of generating movement data from data stored in the data server (111), estimating the moving route of the passenger, and estimating the actual timetable are performed. The calculation server (112) is mainly composed of a network interface (I/F(A)) (130), a CPU (131), a memory (132), and a storage unit (133). A network interface is an interface for connecting to a network. The storage unit (133) includes a data storage unit (139) storing a travel log generation program (134), a route estimation program (135), a route estimation program (136), an actual schedule estimation program ( 137), a group of programs such as the boarding train estimation program (138), the results of calculation processing, the obtained statistical values and index values, and the like. The storage unit is, for example, a hard disk drive, a CD-ROM drive, a flash memory, or the like. In addition, various programs and various data can be divided and recorded in a plurality of recording devices.

When each program group is executed, the data to be analyzed is read from the data server (111) and temporarily stored in the memory (132), and each program (134, 135, 136, 137, 138) is read by the CPU (131). Execute after output to the memory, thereby realizing various functions. The execution timing of these programs can be performed at the timing of requests from operators (119), users, and transportation companies (115, 117), or every time new data is added to the data server (111), or as batch processing, it can be Processing is automatically performed at the time determined each day.

The information distribution server (113) includes a network interface (I/F(B)) (145), a CPU (146), a memory (147), and a recording device (148). A network interface is an interface for connecting to a network. The recording device records various programs and various data, and is, for example, a hard disk drive, a CD-ROM drive, a flash memory, or the like. In addition, various programs and various data may be divided and recorded in a plurality of recording devices.

The information sending server (113) is used for transportation companies and users (115, 117) to refer to the actual time table via the network (114) from the mobile information terminal (116) or the fixed information terminal (118), and use the generated actual schedule. A server for analysis results such as timetable flow analysis and delay status. The recording device (148) includes a display screen creation program (141) and an information transmission program (142). The CPU (146) executes various functions by reading and executing various programs recorded in the recording device (148) into the memory.

Fig. 2 is a diagram showing the structure of transportation system IC card data (122), which is representative data stored in the data server (111). First, the traffic system IC card data (122) includes the log ID (201), the target user ID (202), and the station and bus station IDs (203) linked by information of which data reading terminal has passed. , information such as the use time (204) and the use category (205) such as entering or leaving the station have been passed through the reading terminal. The category used here is information indicating, for example, processing categories such as "entry" or "exit" in the case of ticket gates or gates, and "purchase" in the case of merchandise sales terminals. The traffic system IC card data (122) may be sent every time data is newly generated, or may be sent together late at night when usage decreases. On the side of the data server (111), it is only necessary to perform storage processing in accordance with the transmission timing.

Fig. 3 is a diagram showing the types of master data (123) stored in the data server (111) and their respective data structures. First of all, the basic data related to places where transportation means can be used, such as stations, bus stations, and roads, that is, the main location (300) includes station/bus station ID (301), station/bus station name (302), holding company (303 ), location (304), latitude and longitude information (305) and other information. When there is a change in the structure of stations, bus stations, routes, and roads, data can be added or corrected at any time. The basic data about the route, that is, the main route (310) includes information such as route ID (311), route name (312), operating company (313), and route type (314) for distinguishing whether it is a railway route or a car route for identifying the route . The basic data for associating stations and routes, that is, the main station/bus station-route relationship (320) includes the route ID (321) for identifying the route, the station/bus station ID (322) included in the route, and the management station/bus station ID (322). Information such as the sequence number (323) of the sequence of the bus stop, the category (324) identifying whether to stop or pass, and the required time (325) from the starting point. In the case of changes such as stations and bus stops, routes and roads, master data is input and updated/recorded from outside the system shown in FIG. 1 at each change (123).

FIG. 4 is a diagram showing a data structure for storing movement log data (124) stored in the data server (111). The movement log data (124) includes the log ID (401) for identifying the log and the target user ID (402), the boarding date and time (403) indicating the time when the use of the transportation means was started at the departure point, and the date and time (403) indicating the time at the arrival point Get off date and time (404) of the time when the use of the transportation means is finished, the payment amount (405) indicating the cost of moving, the boarding station/bus station ID1 (406), the getting off station/bus station ID1 (407) , boarding station/bus station ID2 (408), getting off station/bus station ID2 (409) and other information. The travel log data (124) is primary processed data generated using the transportation system IC card data (122) and the like.

FIG. 5 is a diagram showing a data structure for storing a planning schedule (125) stored in a data server (111). Planned schedule data (125) includes information on schedule ID (501) for identifying a schedule, its route ID (502), stop or bus station (503), arrival time (504), and departure time (505) .

FIG. 6 is a diagram illustrating a processing procedure of a movement generation program (134) that generates movement log data (124) based on the transportation system IC card data (122) and stores it in the data server (111). Here, the storage process to the data server (111) will be described as a batch process performed at a time determined every day. First, with reference to the user ID (202) and usage time (204) included in the newly collected traffic system IC card data (122), all data are sorted in user ID order and time order (processing step 601). Next, the following same processing for the number of user IDs is repeated for the data sorted in processing step 601 (processing step 602 ). First, list-type variables corresponding to boarding station/bus station ID, boarding date and time, alighting station/bus station ID, and alighting date and time are initialized (processing step 603). Next, the same processing as below is repeated for the data arranged in chronological order (processing step 604). First, by using the value of the category (205) to distinguish between situations (processing step 605), each process is performed. When the value of use category (205) is inbound, at first confirm whether there is previous outbound diary (processing step 606) in the diary of the same user and the same day, when there is outbound diary, judge the date and time of getting off the bus Whether the difference with the boarding date and time of the current log is within a predefined threshold (processing step 607). This threshold is a value for judging the transfer of a plurality of vehicles, and is preferably set within a range of several minutes to several tens of minutes, for example. If the difference between the alighting time and date of the previous outbound log and the boarding time and date of the current log is within the threshold, it is considered to continue a series of movements, and the list of boarding station/bus station ID and boarding date and time is appended value (process step 608). When the threshold value is exceeded, it is considered that the time since the previous movement is very free, so it can be judged that the previous movement information should be divided here. The value of the variable is therefore stored in the movement log data (124) (process step 609), and the variable is initialized again (process step 610). If there is no corresponding previous departure log, a value is added to the list of boarding station/bus station ID and boarding date and time (processing step 611). When the value of the usage category (205) is outbound, a value is added to the variables of the getting off station/bus stop ID and the getting off date and time (processing step 612). When a value is set to a variable at the end of the repeated processing for one user ID, the value of the variable is stored in the movement log data (processing step 613). Here, the log ID (201) is held as a serial number. Here, the threshold value t for judging whether to perform a series of movements is set in advance as a standard transfer time. The allowable range of transfer time can be adjusted by this threshold t. The threshold value related to the standard transfer time is a positive value, and it can be used as a common value in all traffic networks, or a different value can be set for each area.

FIG. 7 is a diagram illustrating a processing procedure of a usage route estimation program (136) for estimating a usage route of each movement log using movement log data (124). Processing is repeated for the number of movement logs (processing step 701). First, a route from a boarding station to an alighting station is searched using a plurality of search criteria (processing steps 702, 703). In the present embodiment, the route selection criteria will be described as three types: time priority criteria, number of transfer times priority criteria, and cost priority criteria. Usable route selection criteria are not limited to these three. Next, it is determined whether or not there is a route with the same moving direction equal to or greater than a predetermined threshold for all the searched routes (processing step 704 ).

Here, the details of the processing step 704 will be described in detail using FIG. 8 . FIG. 8 shows the results of searching for a moving route from a boarding station (801) to an alighting station (802) based on time priority criteria, transfer times priority criteria, and cost priority criteria. The time-priority route (803) moves to station C (808) via route 1 (809), and transfers at route 2 (811) to get off at station (802). The transfer number priority route (804) is the same as the cost priority route (805), and uses the route 3 (810) to arrive at the getting off station (802). In the case where there is the above-mentioned route, the actual timetable is estimated by considering route 3 as an object.

At this time, in the technique described in Patent Document 2, the movement log from the boarding station to the disembarking station in FIG. 8 cannot be used for estimating the actual timetable of the route 3 . The route from the boarding station includes the routes via the stations C and B using the routes 1 and 2, so it is not always possible to uniquely identify that the log using the route 3 is used.

On the other hand, in the traffic analysis system of the present embodiment, in processing step 704, in order to determine whether the moving direction of the searched route is the same, the stop immediately before the stop of each route (station A in the estimation of route 3) is confirmed. . It is estimated which route was used to arrive at the getting off station by checking whether the probability of which station is the previous stop of the getting off station is high. As a specific method, on the basis of setting a predetermined threshold in advance, by judging a suitable route (a route that has some reasonable advantages compared with other routes regarding the number of transfers, costs, or time, etc.) via station A Whether the probability exceeds this threshold to proceed. In the example of Fig. 8, the station A (806) before the getting off station of 2 routes (transfer times priority route (804) and cost priority route (805)) in 3 routes is station A (806), use route 3 (810 ) to get off at station (802), the previous stop of the get off station of 1 path (time priority path (803)) in 3 paths is station B, use route 2 (811) to arrive at get off station (802). At this time, for example, if the predetermined threshold value is 0.5, two routes (0.67) among the three routes exceed the threshold value, so the probability of moving at least via the station A is high. In other words, the probability of using route 3 (810) is high. Therefore, it is presumed that this log is the log of the user who arrived at the alighting station (802) using the route 3 (810). At this time, the alighting time of the log is recorded as the arrival time candidate of the alighting station of the route (processing step 705).

In this way, the traffic analysis system of the present embodiment is a traffic analysis system that estimates the actual time table of a predetermined route using the passenger movement log of the vehicle, and is characterized in that it has a route estimation unit (135) that estimates the time table from the movement log. A plurality of movement paths from the departure station to the arrival station; using the route estimation unit (136), when the route of the previous station passing through the arrival station of the route is above a predetermined threshold, it is estimated that the movement log uses the route to arrive at The log of the arrival station; the actual timetable estimation unit (137) which estimates the actual timetable based on the alighting time estimated by the used route estimation unit from the travel log which used the route.

Or, a traffic analysis method (and including a program for executing the traffic analysis method and a storage medium storing the program) for estimating an actual timetable of a predetermined route using a movement log of a passenger using a vehicle, characterized in that it includes The following steps: the step of estimating a plurality of movement routes from the departure station to the arrival station in the movement log (702), when the route of the previous station passing through the arrival station of the route among the plurality of movement paths is greater than or equal to a predetermined threshold value, the step of estimating the Step (704) where the travel log is a log of the arrival station using the route; and a step (1001-1005) of estimating the actual timetable based on the alighting time of the travel log that used the route estimated by the route estimating unit.

According to the above-mentioned features, compared with the estimation using the log of only the route using the route as in Patent Document 2, the number of logs used for estimation can be further increased, and as a result, the accuracy of estimation can be further improved.

Here, when the alighting time of the movement log records the actual ticket check-out time, an error occurs with the arrival time of the train. Therefore, in processing step 705, the time of getting off the vehicle in the movement log can be corrected according to predetermined information. As the predetermined information, for example, the average travel time between the route from the parking station to the platform and the ticket inspection can be used.

Fig. 9 is a diagram showing an example of the result of the route estimation process performed by the route estimation program (136). The number of movement logs estimated to get off at each station using the route at each time at each station of each route is recorded. Generally, passengers expect to leave the station through the ticket gate without staying after getting off the train, so it is considered to collect a plurality of movement logs around the train arrival time.

Fig. 10 is a diagram illustrating an example of a processing procedure of the actual schedule estimation program (137). The actual timetable estimation step is processed for each route (processing step 1001). First, a minor correction is added to the planned timetable for the route (process step 1002). As a method of increasing the correction, it is conceivable to shift the departure and arrival times by several seconds to several minutes for the entire timetable, or for each train, or for each station. Next, the degree of coincidence between the corrected planned timetable and the alighting time estimated by using the route estimation program (136) using the travel log of the route is calculated (processing step 1003). As the degree of coincidence, the number of movement logs in which the train arrival time and alighting time of the corrected planned timetable match can be used. If a predetermined threshold is set because an error is included in the alighting time of the movement log and the difference between the train arrival time and the alighting time of the corrected planned timetable is within the threshold, it can be determined to be coincidence. If the coincidence degree is the largest among the coincidence degrees calculated so far, the revised planned schedule is recorded as the actual schedule (processing step 1005). Finally, the corrected planned schedule with the highest degree of consistency is recorded in the data server (111) as the actual schedule (126). In general, there is rarely a large deviation between the planned timetable and the actual timetable. Therefore, the above-mentioned method can estimate the actual timetable with higher accuracy.

Fig. 11 is a diagram showing an example of an estimation result presentation screen by the route estimation program (136) and the actual timetable estimation program (137) performed by the display screen creation program (141). An operator (119) can select a route through a pull-down menu (1102) in the screen (1101). A used route estimation result (1104), a planned timetable (1105), and an actual timetable estimation result (1106) of the selected route are displayed on the display screen (1103). When the actual timetable estimation result has an error, the operator (119) can use the cursor (1107) to correct the actual timetable (1106). When correcting the actual time table, the operator can easily grasp whether it should be corrected in the forward direction or in the opposite direction on the time axis, and there is information to improve the accuracy of the correction, so the operator can intuitively correct the actual time table as shown in FIG. 11 . A timetable GUI is useful.

Fig. 12 is a diagram illustrating another example of the processing procedure of the actual schedule estimation program (137). The actual timetable estimation step is processed for each route (processing step 1201). Next, processing is performed for each station in the route (processing step 1202). A representative alighting time is determined from the alighting time of the usage log of each station (processing step 1203). The representative alighting time refers to the alighting time representing the partial set when the movement log is divided such that logs with close alighting times are included in the same partial set.

Here, the method of determining the representative alighting time will be described in detail using FIG. 13 . As a decision method representing the drop-off time, a clustering method is most suitable. When the number of daily runs of the route is known, clustering methods such as the K-means method and the Gaussian mixture model can be used, and when the number of runs is unknown, a Dirichle process mixture model can be used clustering method. Known techniques can be used for these clustering methods, so detailed descriptions are omitted. By applying the clustering method, the alighting time of the usage log of the route can be divided into a plurality of clusters (from the upper row to the middle row in FIG. 13 ). Next, by determining the representative values of the divided clusters, it is possible to obtain the representative alighting time (lower row of FIG. 13 ). As for the method of determining the representative value, an average value, a median value, a mode value, and the like of the alighting time within the cluster are suitable. In this way, by obtaining the representative alighting time from the alighting time in the usage log of the route, it is possible to eliminate an error in the estimated result of the used route included in the usage log.

Next, the actual timetable estimation program (137) performs processing between each station (processing step 1204). The correlation between the representative alighting times at a certain station and the next station is obtained, and the time difference τ at which the correlation becomes maximum is obtained, and the time difference τ is set as the average required time between the station and the next station (processing step 1205). The correlation when the time difference is τ is calculated by Equation 1.

[Formula 1]

Here, X(t) is a sequence in which the representative alighting time of the station is stored, and X(t) is 1 when time t is the representative alighting time. Y(t) is a sequence in which the representative alighting time of the station next to the station is stored. When the average required time between the station and the next station is τ, if the time t is the representative alighting time at the station, the expected time t+τ becomes the representative alighting time at the next station. Therefore, the correlation is greatest when τ is the average time required for this station and the next station. An example of the calculation result of the correlation is shown in FIG. 14 .

Next, the actual timetable estimation program (137) searches for a representative alighting time whose time difference is close to the average required time between stations calculated in processing step 1205, thereby generating an actual timetable (processing step 1206). Since the representative alighting time may include gaps, the actual time table can be corrected by interpolating the gaps (processing step 1207). FIG. 15 shows a processing conceptual diagram of processing step 1206 and processing step 1207. An actual time table is generated by linking the representative alighting time (1502) of the next station whose time difference with the representative alighting time (1501) of a certain station is the average required time between stations calculated in the processing step 1205. Interpolation is performed for the missing portion representing the time of alighting (1504).

In this way, the actual timetable estimation program (137) estimates the actual timetable using the planned timetable when the planned timetable can be used, and can calculate the representative alighting time and The average time required between stations is used to estimate the actual timetable.

Fig. 16 is a flowchart showing an example of the processing procedure of the boarding train estimation program (138). Processing is repeated for the number of movement logs (processing step 1601). First, a route search is performed for each route preference pattern recorded in the route preference pattern data (127) stored in the data server (111) (processing steps 1602, 1603). The route selection criteria of passengers considered for route search are stored in the route preference pattern data (127). In addition to the "time priority criterion" for selecting the route that arrives at the earliest station, the "transfer priority criterion" for selecting the route with the fewest number of transfers, the "cost priority criterion" for selecting the cheapest route, and the least expensive train In addition to the "congestion avoidance criteria" for empty routes, the "seat priority criteria" is also considered for prioritizing securing seats by going from the boarding station in the opposite direction to the disembarking station in order to sit on a crowded train and returning to the starting station of the route. "Wait. By setting the stations as nodes and the lines connecting the stations as edges, expressing the railway network through a graph structure, and setting the weights between stations as required time, number of transfers, fare, and congestion respectively, through Dijie Algorithms such as Dijkstra's algorithm solve the shortest path problem, thereby realizing path search on a time priority basis, transfer priority basis, cost priority basis, and congestion avoidance basis. Known techniques can be used to solve the shortest path problem, so detailed description is omitted.

Here, the route search based on seat priority will be described in detail using FIG. 17 . When the path (1703) from the boarding station (1701) to the getting off station (1702) was crowded, the path (1705) of the downward station (1702) after turning back by selecting the departure station (1704) in the opposite direction once, Seating at the origin station (1704) is prioritized by increasing travel time. Such a seat priority route (1705) can be searched by adding a round-trip route between the boarding station (1701) and the departure station (1704) to the searched general route (1703). When the number of stations between the boarding station (1701) and the departure station (1704) is small, or when the number of stations between the boarding station (1701) and the alighting station (1704) is large, it is appropriate to select the seat priority route (1705) Therefore, when the number of stations between the boarding station (1701) and the originating station (1704) is below a predetermined threshold, or when the number of stations between the boarding station (1701) and the disembarking station (1704) is predetermined The search can be performed when it is above the threshold. In addition, depending on the transportation company, such a return ride may be prohibited. At this time, the purpose of the present invention is not to recommend the above-mentioned prohibited behavior to the user, but to allow the transportation company to grasp the actual situation of such a return ride by adding seat priority criteria to the route search candidates.

Next, the on-board train estimation program (138) performs on-board train estimation for each searched route (processing steps 1604, 1605). The details of the on-board train estimation will be described using FIG. 18 . First, use the boarding time (1801) of the movement log and the actual timetable (126) estimated by the actual timetable estimation program (137) to determine the train with the latest departure time (1802) after the boarding time (1801). When the boarding time (1801) in the travel log is actually the check-in time, a time table in which a predetermined estimated travel time is added can be used. In fact, there are fluctuations in the travel time from the passenger's ticket check to the boarding place, so there is no limit to the train that can be boarded at the departure time (1802) closest to the boarding time (1801). Therefore, the train of the next departure time (1803) is also added to the candidate train. Even at the transfer station (1804), trains with a plurality of departure times are added to the boarding train candidates. The number of trains added to the boarding train candidates can be given by a threshold or the like. Next, when the alighting time (1805) of the movement log is the departure time, compare the departure time (1807) obtained by adding the estimated estimated movement time to the arrival time of the train candidate (1806) and the departure time of the movement log (1807). Station time (1805), the nearest train is set as the boarding train. Finally, in the search route, the nearest departure time (1805) of the movement log and the estimated departure time (1807) of the boarding train is output as the boarding train. In this embodiment, the nearest departure time (1805) of the movement log and the estimated departure time (1807) of the boarding train candidate is output as the boarding train, but the boarding train estimation method is not limited to this, for example It is also possible to determine the boarding train from a plurality of boarding train estimation candidates whose difference between the departure time (1805) of the movement log and the departure time (1807) of the boarding train estimation candidate is within a predetermined threshold according to a predetermined priority order. . In the present embodiment, the on-board train estimation (processing step 1605) using the route search (processing step 1603) and the actual time table (126) are performed separately, but it is also possible to perform a route search (processing step 1603) considering the actual Timetable (126) search.

Fig. 19 is a diagram showing an example of an estimation result presentation screen of the boarding train estimation program (138) of the display screen generation program (141). When the operator (119) selects any station (1902) on the route map in the screen (1901), the route selection ratio of passengers at that station is displayed as a stacking surface graph (1904) in the screen (1903). The horizontal axis of the accumulation surface graph (1904) represents time, and the vertical axis represents the details of the route selection criteria estimated from the movement logs of passengers at the station at that time, the number of people passing by. The GUI of FIG. 19 is useful because it is possible to easily grasp the ratio of the route selection criteria of a desired station or its change. The presentation method of the ratio of the route selection reference is not limited to the stacked surface graph, and a circular graph or the like may be used. In addition, not only the route selection ratio of passengers but also the route selection ratio of passengers getting off can be presented, and the operator (119) can select not only a station but also a route.

Claims (9)

1. A traffic analysis system for estimating an actual timetable of a predetermined route using a movement log of a passenger using a vehicle, the traffic analysis system being characterized in that it has: a route estimation unit that estimates a plurality of travel routes from a departure station to an arrival station in the travel log; a route estimating unit that estimates that the travel log is a log of arriving at the destination station by the route when a route preceding the destination station via the route among the plurality of travel routes is equal to or greater than a predetermined threshold value; and The actual timetable estimating unit estimates the actual timetable based on the disembarkation time estimated by the used route estimating unit in the travel log in which the route is used. 2. The traffic analysis system according to claim 1, wherein: The above-mentioned planned timetable is corrected so that the degree of agreement between the alighting time of the movement log of each of the above-mentioned routes and the planned timetable is maximized, and the above-mentioned revised planned timetable is set as the above-mentioned actual timetable. 3. The traffic analysis system according to claim 1, wherein: Divide the movement log for each of the above-mentioned routes so that the logs whose alighting times are close to each other are included in the same partial set, perform calculations to obtain the representative alighting time representing the above-mentioned partial set, and set the above-mentioned representative alighting time as the time of the above-mentioned route Estimated result of train arrival time. 4. The traffic analysis system according to claim 1, wherein: A calculation is performed to obtain the correlation between adjacent stations in the movement log for each of the above-mentioned lines, and the time difference at which the above-mentioned correlation is the largest is set as the average required time between the above-mentioned adjacent stations. 5. The traffic analysis system according to claim 1, wherein: Furthermore, it has a display part which displays the alighting time of the movement log of each said route, a planned timetable, or the said actual timetable on the same screen. 6. The traffic analysis system according to claim 5, wherein: It also has a correcting unit for correcting the actual timetable on the above-mentioned screen. 7. The traffic analysis system according to claim 1, wherein: Each of the above-mentioned movement logs also has a boarding train estimation unit that searches for arrivals from the above-mentioned departure station with at least one criterion including time priority criteria, transfer priority criteria, fare priority criteria, congestion avoidance criteria, or seat priority criteria. In the case of a plurality of routes to the arrival station, the train on which the passenger rides corresponding to the movement log is estimated by using the actual timetable. 8. The traffic analysis system according to claim 7, wherein: A display unit is further provided for displaying a ratio of an estimated route selection criterion which is a result of on-board train estimation by the on-board train estimating unit. 9. A traffic analysis method for estimating an actual timetable of a predetermined route using a movement log of a passenger using a vehicle, wherein the traffic analysis method is characterized in that with the following steps: A step of estimating a plurality of movement routes from the departure station to the arrival station of the movement log; A step of estimating that the movement log is a log of arriving at the destination station via the route when a route preceding the destination station via the route among the plurality of travel routes is equal to or greater than a predetermined threshold value; and A step of estimating the actual timetable based on the alighting time estimated by the used route estimating unit in the movement log that used the route.