JP2021043641A - Information processing device, system, method and program - Google Patents

Abstract

To provide an information processing device capable of easily estimating, in a facility, a number of people who have passed through a target location where facility users are passing without performing complicated data aggregation processing.SOLUTION: An information processing device that estimates a number of people who have passed through a target location in a facility where facility users are passing is configured: to acquire, from a short-range wireless communication access point device installed at or near the target location, information on a number of detected wireless terminal devices that are detected by the access point device during estimation target time; to store a parameter value of an estimation model for estimating the number of people who have passed through the target location from the number of detected wireless terminal devices; and to estimate the number of people who have passed through the target location during the estimation target time based on the estimation model including the parameter value and the information on the number of detected wireless terminal devices during the estimation target time.SELECTED DRAWING: Figure 1

Description

The present invention relates to an information processing device, a system, a method and a program for estimating the number of people passing through a target location where a user in a facility passes.

Conventionally, there is known a system that aggregates data on the number of people passing through an automatic ticket gate at a station. For example, in Patent Document 1, an admission restriction device provided for each station and an automatic ticket gate that controls the user's entry into the station by opening and closing the gate are connected by an in-station network, and the automatic ticket gate of each station is connected. A system is disclosed in which the number of people passing through a gate in an aircraft is aggregated by an admission restriction device and transmitted as passing personnel data to an admission restriction data generator.

Japanese Unexamined Patent Publication No. 2011-096057

Conventional automatic ticket gates automatically determine the type of ticket such as a magnetic ticket or IC ticket (IC card), the type of ordinary ticket or commuter pass, etc. when a user getting on and off at a station passes by, and use it. Correctly collect the fare from the person. The data read and written by the automatic ticket gate is aggregated and stored in a plurality of data aggregation PC devices different for each type of ticket. Therefore, in order to grasp the number of people who have actually passed through the automatic ticket gate, there is a problem that complicated data processing is required to read and aggregate the data stored in each of the plurality of data aggregation PC devices. ..

The information processing device according to one aspect of the present invention is an information processing device that estimates the number of people passing through a target location where a user in a facility passes. This information processing device acquires information on the number of detections of the wireless terminal device detected by the access point device within the estimated target time from the access point device for short-range wireless communication provided at or near the target location. An acquisition unit, a storage unit that stores parameter values of an estimation model that estimates the number of people passing through the target location from the number of detections of the wireless terminal device, an estimation model that includes the parameter values, and an estimation target time within the estimation target time. It is provided with an estimation unit that estimates the number of people passing through the target location within the estimation target time based on the information on the number of detections of the wireless terminal device.

In the information processing device, the information on the number of detected wireless terminal devices detected by the access point device within a predetermined time and the information on the number of passing persons obtained by totaling the number of people who actually passed the target location within the predetermined time. Based on this, a determination unit for determining the values of the parameters of the estimation model may be further provided. Here, a real number information acquisition unit may be further provided to acquire real number information obtained by totaling the number of people who have actually passed through the target location within the predetermined time.

前記欠損度は、係数ａ，ｂを有する次の（３）式で定義され、前記決定部は、前記アクセスポイント装置が所定時間内に検知した無線端末装置の検知数の情報と前記実数の情報とに基づいて、前記（３）式の欠損度の中の係数ａ，ｂを決定し、前記推定部は、次の（４）式を用いて前記通過人数を推定してもよい。

In the information processing device, the estimated model may be a model in which the number of passing persons is represented by a linear function of the number of detections of the wireless terminal device.
Here, the parameters representing the inclination of the linear function are the carrier share rate, which is the market occupancy rate of the communication carrier corresponding to the wireless terminal device, the short-range wireless function ON rate of the wireless terminal device, and the carrier share rate. The determination unit includes the degree of defect indicating the degree to which the access point device cannot detect the wireless terminal device for reasons other than the short-range wireless function ON rate, and the determination unit detects the access point device within a predetermined time. Based on the information on the number of detections of the wireless terminal device and the information on the real number, the degree of deficiency is calculated and determined using the following equation (1), and the estimation unit uses the following equation (2). The number of people passing by may be estimated.

The degree of deficiency is defined by the following equation (3) having coefficients a and b, and the determination unit determines the information on the number of detections of the wireless terminal device detected by the access point device within a predetermined time and the information on the real number. Based on the above, the coefficients a and b in the degree of deficiency in the equation (3) may be determined, and the estimation unit may estimate the number of passing persons using the following equation (4).

In the information processing apparatus, the value of the parameter of the estimation model is determined for each of a plurality of different conditions that affect the estimation of the number of people passing through the target location, and the estimation unit determines the estimation target from the plurality of conditions. A time-corresponding condition may be selected and an estimation model containing the value of the parameter corresponding to the selected condition may be used to estimate the number of passersby.
The item of the condition may include at least one of a time zone and a day of the week. In addition, the item of the condition may include at least one of the statistical value of the degree of congestion in the vicinity of the target location and the walking speed of a person passing through the target location.

In the information processing device, the estimation model uses information on the number of detections of the wireless terminal device and information on a plurality of types of conditions that affect the estimation of the number of people passing through the target location as environmental variables, and the number of people passing through the target location. This is a machine learning model in which the target variable is information on the number of detected wireless terminal devices detected by the access point device within a predetermined time, and information on a plurality of types of conditions that affect the estimation of the number of people passing through the target location. , The value of the parameter of the machine learning model may be determined by using a plurality of learning data including information on the actual number of people passing through the target location.

The information processing device is further provided with an analysis unit that performs attribute analysis on a group of people who have passed the target location based on the attribute information of the user of the wireless terminal device detected within the estimated target time. May be good.

The information processing device according to another aspect of the present invention is an information processing device that analyzes a person who has passed a target location where a user in a facility passes, and is provided in the vicinity of the target location for short-range wireless communication. With respect to the plurality of wireless terminal devices detected by the access point device, the target location is determined based on the storage unit that stores the attribute information of the user of the wireless terminal device and the attribute information of the user of the plurality of wireless terminal devices. It is equipped with an analysis unit that performs attribute analysis on a group of people who have passed.

The attribute information may include at least one piece of information on the gender, age and place of residence of the user.

The system according to still another aspect of the present invention includes any of the information processing devices and the access point device.

A method according to still another aspect of the present invention is a method of estimating the number of people passing through a target location where a user in a facility passes. In this method, information on the number of detected wireless terminal devices detected by the access point device within the estimated target time is acquired from the access point device for short-range wireless communication provided at or near the target location. The value of the parameter of the estimation model that estimates the number of people passing through the target location from the number of detections of the wireless terminal device is stored, the estimation model including the value of the parameter, and the wireless terminal device within the estimation target time. It includes estimating the number of people passing through the target location within the estimated target time based on the information on the number of detections.
In the method, based on the information on the number of detected wireless terminal devices detected by the access point device within a predetermined time and the information on the number of passing persons obtained by totaling the number of people who actually passed the target location within the predetermined time. , The determination of the values of the parameters of the estimation model may be further included.

A method according to still another aspect of the present invention is a method of analyzing a person who has passed a target portion through which a user in a facility passes. This method is to store the attribute information of the user of the wireless terminal device for a plurality of wireless terminal devices detected by the access point device for short-range wireless communication provided at or near the target location, and the plurality of wireless terminal devices. Includes performing attribute analysis on a group of people who have passed the target location based on the attribute information of the user of the wireless terminal device of the above.

The program according to still another aspect of the present invention is a program executed by a computer or a processor provided in an information processing device for estimating the number of people passing through a target location passed by a user in a facility. This program is a program for acquiring information on the number of detected wireless terminal devices detected by the access point device within the estimated target time from the access point device for short-range wireless communication provided at or near the target location. The code, the program code for storing the parameter values of the estimation model that estimates the number of people passing through the target location from the number of detections of the wireless terminal device, the estimation model including the parameter values, and the estimation target time. Includes a program code for estimating the number of people passing through the target location within the estimated target time based on the information on the number of detections of the wireless terminal device.
In the program, based on the information on the number of detected wireless terminal devices detected by the access point device within a predetermined time and the information on the number of passing persons obtained by totaling the number of people who actually passed the target location within the predetermined time. , The program code for determining the value of the parameter of the estimation model may be further included.

The program according to still another aspect of the present invention is a program executed by a computer or a processor provided in an information processing apparatus for analyzing a person who has passed a target location passed by a user in the facility, and is the target location or the program thereof. For a plurality of wireless terminal devices detected by an access point device for short-range wireless communication provided in the vicinity, a program code for storing attribute information of a user of the wireless terminal device and use of the plurality of wireless terminal devices. Includes a program code for performing attribute analysis on a group of people who have passed the target location based on the attribute information of the person.

According to the present invention, it is possible to easily estimate the number of people who have passed the target location where the user in the facility passes without performing complicated data aggregation processing.

The explanatory view which shows an example of the schematic structure of the whole system including the information processing apparatus which concerns on embodiment. A graph showing an example of time changes in the number of ticket gate users on weekdays. A graph showing an example of time changes in the number of people using ticket gates on holidays. A graph showing an example of time changes in the usage status (number of entries and exits) of magnetic tickets (commuter passes, non-commuter passes) on weekdays. A graph showing an example of time changes in the usage status (number of entries and exits) of magnetic tickets other than commuter passes on weekdays. A graph showing an example of time changes in the usage status (number of entries and exits) of magnetic tickets (commuter passes, non-commuter passes) on holidays. A graph showing an example of time changes in the usage status (number of entries and exits) of magnetic tickets other than commuter passes on holidays. The graph which shows an example of the change of the relationship between the number of ticket gate users per day and the number of terminals detected by the AP device. A scatter plot showing an example of the correlation between the number of ticket gate users and the number of terminal detections. The graph which shows an example of the male ratio and the female ratio by the time zone of the terminal user on a weekday. The graph which shows an example of the male ratio and the female ratio by the time zone of the terminal user on a holiday. A graph showing an example of estimation results of male and female usage status by time zone of ticket gate users on weekdays. The graph which shows an example of the estimation result of the usage situation of men and women by time zone of the ticket gate user on a holiday. (A) and (b) are graphs showing an example of the ratio of terminal users by age group on weekdays and holidays, respectively. (A) and (b) are graphs showing an example of the top five prefectures of the address attributes (contractual addresses) of terminal users on weekdays and holidays, respectively. The graph which shows an example of the change of the relationship between the number of ticket gate users per day, the number of terminals detected by the AP device, and the degree of deficiency. A graph showing an example of the number of ticket gate users and the time transition of the degree of deficiency on weekdays. A graph showing an example of the number of ticket gate users and the time transition of the degree of deficiency on holidays. A scatter plot showing an example of the correlation between the number of ticket gate users and the degree of deficiency on weekdays. A scatter plot showing an example of the correlation between the number of ticket gate users on holidays and the degree of deficiency. The graph which shows an example of the result of having obtained the estimated value of the ticket gate user (real number) from the number of terminal detections using the value of the degree of deficiency (a, b) which is a parameter of a linear model on a weekday. The graph which shows an example of the result of having obtained the estimated value of the ticket gate user (real number) from the number of terminal detections using the value of the degree of deficiency (a, b) which is a parameter of a linear model on a holiday. A graph showing an example of the time transition of the number of ticket gate users and the degree of deficiency on weekdays and holidays created by superimposing FIGS. 17 and 18. The graph which shows an example of the time transition of the ratio of regular users on weekdays and holidays. The block diagram which shows an example of the structure of the main part of the information processing apparatus which concerns on embodiment. The flowchart which shows an example of the determination process of the parameter value of the estimation model in the information processing apparatus which concerns on embodiment.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is an explanatory diagram showing an example of an overall schematic configuration of a system including an information processing device 10 according to an embodiment of the present invention. The information processing device 10 estimates the number of people passing through the automatic ticket gate 90 at the target location where the user 80 passes at the station as a facility. The information processing device 10 receives information on the number of detections of the wireless terminal device 85 detected by the access point device 20 within the estimated target time from the access point device 20 for short-range wireless communication provided at or near the automatic ticket gate 90. get. The information processing device 10 stores in the storage unit (DB) 101 the value of the parameter of the estimation model that estimates the number of people passing through the automatic ticket gate 90 from the number of detections of the wireless terminal device 85. Further, the information processing device 10 has passed through the automatic ticket gate 90 within the estimated target time based on the estimation model including the value of the parameter and the information on the number of detections of the wireless terminal device 85 within the estimated target time. Estimate the number of people passing through (the number of ticket gate users).

The system of this embodiment may include an access point device 20 for short-range wireless communication in addition to the information processing device 10. The information processing device 10 and the access point device 20 can communicate with each other by a predetermined interface such as a mobile communication network 30 to transmit and receive data.

Further, the system of the present embodiment stores subscriber information (for example, gender, age, address, etc.) of communication services such as wide area mobile communication by the wireless terminal device 85 and short-range wireless communication via the access point device 20. A customer management server 15 having a storage unit (database) 151 may be included.

The data read and written by the automatic ticket gate 90 at the target location is aggregated and stored in a plurality of data aggregation PC devices 91 and 92, which are different for each type of ticket. The data of the data aggregation PC devices 91 and 92 are collected in the customer management server 95 via the communication network 35 such as the Internet, stored in the storage unit (DB) 951, and managed as statistical data such as the number of passengers and sales. Will be done.

In this embodiment, the automatic ticket gate 90 owned by the railway operator is based on the background that the individual data held by the operators of different industries are not sufficiently effective and reliable and cannot be used well. Communication log data (access) of 20 access point devices (hereinafter referred to as "AP devices") of data and short-range wireless communication (for example, wireless LAN communication such as Wi-Fi (registered trademark)) owned by a communication company. Log) is integrated and analyzed.

In the present embodiment, based on the result of the above analysis, an estimation model for estimating the number of people passing through the automatic ticket gate 90 from the communication log data of the AP device 20 is constructed. By using this estimation model, only the communication log data of the AP device 20 can be installed by simply installing the AP device 20 without performing complicated data aggregation processing from the plurality of conventional PC devices 91 and 92 for data aggregation. The number of users 80 who have passed through the automatic ticket gate 90 can be easily estimated from the above, and it is possible to grasp the retention status of people in the automatic ticket gate 90 and improve the accuracy of detecting the flow of people.

Further, in the present embodiment, the communication log data of the AP device 20 and various attribute information of the subscriber owned by the communication operator are collated, and various attribute information is given to the user 80 passing through the automatic ticket gate 90. By combining these, it is possible to grasp the accurate usage status of the user 80, which was unknown until now, and use it for disaster prevention, disaster mitigation, urban development, activation of tourist spots and commercial facilities, and the like.

In the present embodiment, the case where the target location through which a person passes is the ticket gate where the automatic ticket gate 90 of the station, which is the facility of the railway company, is installed, but the present invention describes the case of a facility other than the station (for example, It can also be applied to estimate the number of people passing by at leisure facilities such as theme parks and commercial facilities such as department stores, and to analyze the passing users. Further, the AP device 20 in the present embodiment is an AP device that communicates with a user's wireless terminal device 85 by a wireless LAN other than Wi-Fi (registered trademark), or a short-range wireless device other than the wireless LAN. It may be an AP device to perform. Further, the AP device 20 may be a base station of a mobile communication system.

The wireless terminal device 85 of the user 80 of the present embodiment is a device that the user 80 can carry and perform short-range wireless communication, and is a smartphone, a user communication terminal, a user device (UE), a user terminal, a terminal, and a terminal. It may be a wireless communication device called a device, a mobile station, a mobile device, or the like.

First, the wireless LAN communication terminal device (hereinafter referred to as the wireless LAN communication terminal device) detected by communicating between the ticket gate count data counted by the automatic ticket gate 90 of the railway station in the present embodiment and the AP device 20 installed around the ticket gate. It is also referred to as a “terminal”.) The result of examining the relationship with the wireless access data including the information of the number of detections of 85 will be described.

[Ticket gate count data]
The ticket gate count data is owned by, for example, a railway company. The automatic ticket gate 90 automatically determines the types of tickets such as magnetic tickets and IC card tickets (also referred to as "IC tickets") when getting on and off at stations, and the types of ordinary tickets and commuter tickets. However, it is a device for correctly collecting the fare from the user 80. In this embodiment, "ticket gate count data" is used, in which various data that can be acquired from the automatic ticket gate 90 are aggregated for each ticket gate and the number of people passing by for each boarding / alighting is aggregated for each hour.

The automatic ticket gate 90 reads and writes information necessary for fare calculation, such as determining the expiration date, where to get on and where to get off, but the data format is not unified because the contents differ depending on the type of ticket. In addition, the ticket gate count data counts the number of people for magnetic tickets by commuter pass, school commuter pass, and non-commuter pass, but for IC cards, only the total number of people is counted regardless of whether it is regular or non-commuter pass. .. Further, since the holder information of the magnetic ticket or the IC ticket is not held in those tickets, it is not possible to grasp the attributes such as the gender and age of the holder.

The ticket gate count data used in this embodiment is considered to be statistical data, but the so-called "Revised Personal Information Protection Law" that came into effect on May 30, 2017 and the personal information protection of railway operators created based on it. It will be treated according to the policy. For example, in order to make it impossible to identify a specific individual when utilizing the ticket gate count data, the particle size of the statistical value should be 10 or more (for example, if the ticket gate count is 9 or less, it is rounded down to 0). Data with sufficient anonymization was used for the analysis. Furthermore, as consideration for personal information, data in a time zone when there are few users (for example, data in the 5 o'clock range and data in the 24:00 range) are also omitted.

[Wireless access data]
A telecommunications carrier that provides communication services via a short-range wireless (wireless LAN) AP device 20 such as Wi-Fi collects various communication logs for the purpose of grasping the service status, measuring quality, improving, etc. Obtained from device 20. Of the information in this communication log, communication logs such as probe request / response and association request / response exchanged when the terminal 85 for short-range wireless communication such as Wi-Fi connects to the AP device 20 are used as location information. Opportunities to utilize it are expanding with the spread of smartphones.

There are two methods of using the communication log as location information, one is to convert the location information on the terminal 85 side and the other is to convert the location information on the communication carrier side. The former method is a method of extracting the information of the AP device 20 from the probe and beacon information acquired on the terminal 85 side and looking up the matching table showing the relationship between the AP device 20 and the position information to obtain the position information. Is. Functions using this method are installed in most smartphones and are used to acquire location information that is difficult to grasp by GNSS such as GPS in the basement or inside a building. On the other hand, the latter method is a method in which the telecommunications carrier converts the location information by associating it with the location of the AP device 20 installed by the telecommunications carrier. This method is generally considered to have higher accuracy of position information than the former method.

Wireless access data such as communication logs used in this embodiment is also handled in accordance with the above-mentioned "Revised Personal Information Protection Law" and the personal information protection policy of the telecommunications carrier created based on the above-mentioned "Revised Personal Information Protection Law".

Further, when utilizing the position information of the terminal 85 in wireless access data such as a communication log, in most cases, the MAC (Media Access Control) address of the terminal 85 or the AP device 20 is used as an identifier. This MAC address is a 48-bit code for uniquely identifying the device connected to the network (in principle), and of course, when it is easily linked to personal information such as the contractor's name in a telecommunications carrier, etc. , The code alone has personal identification. On the other hand, even if there is no personal identification by itself, it is given to almost all terminals 85 connected to the network, and considering that most individuals always carry some terminals 85, personal information is given from the viewpoint of privacy protection. It is appropriate to protect even if it does not violate the protection law. Therefore, in the present embodiment, the protection of privacy of the user who uses the terminal 85 to which the MAC address is assigned is enhanced by randomizing the MAC address itself or using a different MAC address each time a connection is made. May be good.

The location information of the terminal 85, which corresponds to personal information in the wireless access data such as the communication log used for statistical analysis in the present embodiment, consents to the handling of personal information when using the public wireless LAN service provided by the telecommunications carrier. It is based on the probe and the association log obtained from the terminal 85, excluding the users who refused to utilize the users after the fact. In addition, when using it, it will be handled in accordance with the privacy policy of the telecommunications carrier, and if the particle size of the statistical value is 10 or more (for example, if the number of detections of the terminal 85 is 9 or less, it will be rounded down to 0) within the telecommunications carrier. Only the information that was sufficiently anonymized so as to be processed) was used for analysis as wireless access data (Wi-Fi data).

[Examination of matching between ticket gate count data and wireless access data]
Through a demonstration experiment at an actual station, we examined the matching between ticket gate count data and wireless access data. The demonstration experiment was conducted at Kintetsu Nara Station. Kintetsu Nara Station is (a) a tourist station with world heritage sites such as Kasuga Taisha, Todaiji Temple, and Kofukuji Temple, (b) an arrival station that is central to the administration such as Nara Prefectural Office and has many schools such as Nara Women's University, (c) There are many residential areas in Nara, and it is thought that it is possible to analyze the trends of various users at stations with different usage aspects such as departure stations for commuting to work and school in the direction of Osaka.

The demonstration experiment was conducted from November 2017 to March 2018 by installing AP devices 20 at each of the east and west ticket gates on the premises of Kintetsu Nara Station. Kintetsu Nara Station is an underground station, and most of the terminals other than the terminal 85 of the user who actually uses the station are not detected by the AP device 20.

Table 1 shows an example of ticket gate count data of Kintetsu Nara Station owned by a railway company. In addition, FIGS. 2 and 3 are graphs showing an example of time changes in the number of ticket gate users on weekdays and holidays at Kintetsu Nara Station, respectively. 2 and 3 show entrance and exit separately for the total number of ticket gate users at the east and west ticket gates regardless of the ticket type (IC ticket, magnetic ticket, etc.) from the ticket gate count data. The average number of passengers getting on and off each day is 55,000 on weekdays and 50,000 on holidays, which is not so different, but the usage status by hour is clearly different.

FIG. 4 is a graph showing an example of a time change in the usage status (number of entries / exits) of the magnetic ticket type (commuter pass, other than the commuter pass) on weekdays. FIG. 5 is a graph showing an example of time changes in the usage status (number of entries and exits) of magnetic tickets other than commuter passes on weekdays. In this demonstration experiment, since the ticket gate count data cannot distinguish between commuter passes and commuter passes for IC card users, FIGS. 4 and 5 show the results of analysis using only magnetic tickets (described later). The same applies to FIGS. 6 and 7). The ratio of Kintetsu's magnetic ticket to IC ticket is about 4: 6.

In FIG. 4, the peaks of regular use are from 6:00 to 8:00 and from 17:00 to 19:00, and it is estimated that there are many users who commute to work or school. On the other hand, the peak of non-regular use is from 9:00 to 18:00, but as shown in Fig. 5, the peak of participation is from 9:00 to 11:00, the peak of admission is from 14:00 to 18:00, and in the morning. It is presumed that users who visit Nara will leave Nara after noon, and it is thought that many tourist users will be included outside the commuter pass.

FIG. 6 is a graph showing an example of a time change in the usage status (number of entry / exit) of the magnetic ticket type (commuter pass, other than the commuter pass) on holidays. FIG. 7 is a graph showing an example of time changes in the usage status (number of entrances and exits) of magnetic tickets other than commuter passes on holidays.
Although there is no big difference in the total number of users between weekdays and holidays, as shown in Fig. 6, most of the users use the commuter pass outside the commuter pass, and the peak is from 9:00 to 18:00, which is the same as weekdays. In addition, regarding the number of visitors and participants outside the commuter pass on holidays in Fig. 7, there is a tendency for users who visited Nara in the morning to leave Nara after noon, as on weekdays, but the number is higher than on weekdays. In many cases, it can be estimated that Kintetsu Nara Station on holidays is a tourist center.

As shown in FIGS. 2 to 7 above, the number of users (real numbers) passing through the ticket gate can be grasped from the ticket gate count data of the railway operator, but the attributes of the users are not known, but the following As shown in the above, the attributes of the user can be known by analyzing the data in combination with the wireless access data acquired from the AP device 20 of the telecommunications carrier.

FIG. 8 shows an example of a change in the relationship between the number of ticket gate users (right scale) per day at Kintetsu Nara Station and the number of terminals (left scale) detected by the AP device (Wi-Fi access point) 20. It is a graph. The number of ticket gate users in FIG. 8 is the total of entry and exit of all ticket types such as magnetic tickets and IC tickets, and is considered to be a "real number" using an automatic ticket gate. In fact, the number of ticket gate users increased on December 9th and 10th, 2017, the Nara Marathon, the beginning of 2018, and the day of Wakakusayama-yaki on January 27th, reflecting the movement of users. It seems to be doing.

Since the terminal 85 such as a smartphone communicates with the same AP device in a timely manner, the same terminal may be detected a plurality of times in the terminal detection number data (Wi-Fi data) which is the wireless access data. Therefore, the number of terminals detected (Wi-Fi detection number) is set to the number of unique terminals within a certain period, so even if you enter from Kintetsu Nara Station in the morning and participate in the evening, the number of terminals detected per day is one terminal. It is counted. On the other hand, since entry and entry are counted in the ticket gate count data, respectively, in FIG. 8, a value twice the actual number of detected terminals is adopted.

Looking at FIG. 8, since the increase / decrease in the number of ticket gate users and the number of terminal detections are similar, a correlation analysis between the two was performed.

FIG. 9 is a scatter diagram showing an example of the correlation between the number of ticket gate users and the number of terminal detections (daily) at Kintetsu Nara Station. From the results of the correlation analysis in FIG. 9, a high correlation of R2 (coefficient of determination) = 0.7179 was observed between the number of ticket gate users and the number of terminal detections. From the result of this correlation, the present inventors do not always grasp all the personnel for the number of terminal detections, but it is possible to estimate (estimate) the actual number of ticket gate users from the number of terminal detections. , The attribute of the user or terminal 85 possessed by the wireless access data, or the user or terminal 85 associated (linked) with the wireless access data to the user who uses the automatic ticket gate. I thought it was possible to apply the attributes.

[Application of terminal user attribute information]
As the attribute information of the wireless access data applied to the user who uses the automatic ticket gate, the contractor information (gender) of the wireless LAN communication (Wi-Fi) which is a short-range wireless communication linked to the MAC address of the terminal 85. , Age, place of residence).

As a feature of Kintetsu Nara Station in the above-mentioned demonstration experiment, the railway operator (transportation operator) intuitively understands that there are many men during the morning commuting time and many women during the daytime sightseeing time. However, it is difficult to quantitatively grasp the actual situation, and the actual situation is that we are trying to grasp the actual situation by questionnaires. Therefore, we tried to visualize the attributes of the users who use Kintetsu Nara Station by using the attribute information possessed by the wireless access data (Wi-Fi data).

10 and 11 are graphs showing an example of the hourly gender ratio of terminal users held by wireless access data (Wi-Fi data) on weekdays and holidays at Kintetsu Nara Station, respectively. 12 and 13 show men and men by time zone obtained by applying the hourly gender ratio held by wireless access data (Wi-Fi data) on weekdays and holidays at Kintetsu Nara Station to the number of ticket gate users, respectively. It is a graph which shows an example of the estimation result of the usage situation of a woman. 12 and 13 were obtained by adding the "attribute" (gender of the contractor) of the wireless access data (Wi-Fi data) to the "real number" of the ticket gate count data.

The weekday results shown in FIGS. 10 and 12 show that there are many men during commuting hours at 6 and 7 o'clock, and the proportion of women increases from around 10 o'clock when the surrounding tourist facilities open, and women's during the daytime. This is consistent with the feeling of railway operators (transportation operators) that there will be more. On the other hand, according to the results of the holidays shown in FIGS. 11 and 13, the male-female ratio of the holiday ticket gate users per day is 44% for males and 56% for females. It can be seen that the percentage of female tourists is high in Nara, with the number of female tourists surpassing all day long.

14 (a) and 14 (b) are graphs showing an example of the ratio of terminal users by age group held by wireless access data (Wi-Fi data) on weekdays and holidays at Kintetsu Nara Station, respectively. As shown in Fig. 14, when comparing the actual usage on weekdays and holidays by age group, there was no change in the ratio of young people (teens, 20s) and pre-seniors (50s). , Family members (30's and 40's) tended to be on holidays, and seniors (60's and 70's) tended to be on weekdays.

15 (a) and 15 (b) are examples of the top five prefectures of terminal user address attributes (contractual addresses) held by wireless access data (Wi-Fi data) on weekdays and holidays at Kintetsu Nara Station, respectively. It is a graph which shows. From FIG. 15, it can be seen that Nara Prefecture has the largest number of users on weekdays and holidays, but unexpectedly there are many users in distant places such as Tokyo and Kanagawa Prefecture on weekdays. In addition, since the number of users in Kyoto, Osaka, Hyogo, etc. increases on holidays, it can be seen that the inflow from other neighboring prefectures will increase. In any case, it can be inferred that Kintetsu Nara Station has a considerable number of tourists as well as local users on weekdays and holidays.

[Model for estimating the actual number of ticket gate users from wireless access data]
As shown in Fig. 9 above, there is a high correlation between the number of ticket gate users and the number of terminal detections, so it is possible to build an estimation model that estimates the actual number of ticket gate users from the number of terminal detections in wireless access data. it can.

[Linear model]
As an estimation model, for example, there is a linear model (also referred to as a "linear regression model") in which the real number of ticket gate users is represented by a linear function of the number of terminal detections. In examining the estimation accuracy of this estimation model, how many terminals 85 are detected by the AP device 20 among the terminals 85 such as smartphones located around the AP device (for example, Wi-Fi access point) 20? The "detection rate" of terminal detection by the AP device 20 and its reciprocal "deficiency" are important. The degree of defect and the detection rate are parameters of the estimation model (linear model).

The degree of deficiency and detection rate of terminal detection by the AP device 20 is the number of people actually passing through the "number of ticket gate users" obtained from the ticket gate count data, that is, the number of people passing through the target location (point). It can be evaluated by setting it as a "real number" and comparing it with the number of terminal detections (Wi-Fi detection number) detected by the AP device 20. Generally, the actual number of ticket gate users (passing number) is the number of terminals detected in consideration of the carrier share rate, which is the market share of the communication carrier corresponding to the terminal 85, and the short-range wireless function ON rate of the terminal 85. It can be calculated from. Here, considering the "deficiency", which is the degree to which the AP device 20 cannot capture the terminal 85 for some reason that cannot be explained by the carrier share rate and the short-range wireless function ON rate, the number of detected terminals and the real number (actual number of people counted). The relationship with is as shown in the following equation (5). When the AP device 20 can capture all the peripheral terminals 85, the degree of deficiency = 1.

Therefore, the degree of deficiency can be calculated by the following equation (6).

FIG. 16 is a graph showing an example of a change in the relationship between the number of ticket gate users per day at Kintetsu Nara Station, the number of terminals detected by the AP device (Wi-Fi access point) 20, and the degree of deficiency. Table 2 is an example of calculation of the degree of deficiency, the detection rate, and the number of terminals detected when the real number is 1000. The degree of deficiency in FIGS. 16 and 2 was calculated using the above equation (6), with the carrier share rate of the telecommunications carrier in the above-mentioned demonstration experiment being 22.3% and the short-range wireless function ON rate being 20%. .. As shown in FIG. 16, there is a detection loss that cannot be explained by the carrier share rate and the short-range wireless function ON rate in the terminal detection by the AP device 20, and the loss rate indicating the detection loss is about 1.5 to 2.25. .. Further, as shown in Table 2, the number of terminals detected by the AP device 20 (persons) is 156 at the maximum, while the actual number is 1000.

[Daily deficiency]
Based on the above formula (6), the carrier share rate is 22.3%, the short-range wireless function ON rate is 70%, and the number of terminal detections and ticket gate users during the above-mentioned demonstration experiment at Kintetsu Nara Station ( When the degree of deficiency is calculated on a daily basis from (= real number), the degree of deficiency on a daily basis is 1.94 on average (maximum value: 2.34, minimum value: 1.54). When the degree of defect is 1.94, the detection rate of terminal detection by the AP device 20 is 51.5%, which means that only about half of the terminals 85 can be detected.

Looking at the degree of deficiency by day, a certain degree of periodicity was observed, so the degree of deficiency was calculated by dividing it into weekdays and holidays on calendar days. As a result, the average deficiency on weekdays is 2.05 (maximum value: 2.34, minimum value: 1.74), and the average deficiency on holidays is 1.85 (maximum value: 2.04, minimum value: 1.74). It becomes 1.54), and it can be seen that the terminal 85 is more easily detected on holidays than on weekdays.

[Deficiency by time zone]
17 and 18 are graphs showing an example of the time transition of the number of ticket gate users and the degree of deficiency on weekdays and holidays of Kintetsu Nara Station, respectively. The average degree of deficiency on weekdays in FIG. 17 is 2.99 (maximum is 4.03 at 6 o'clock, minimum is 2.01 at 23:00), and the average degree of deficiency on holidays in FIG. 18 is 2.69 (. The maximum is 3.09 at 6 o'clock and the minimum is 1.99 at 23:00), and like the daily average, the degree of deficiency on holidays is smaller, and the terminal 85 is more likely to be detected on holidays than on weekdays.

[Linear model]
As shown in FIGS. 17 and 18, the degree of deficiency was highest during commuting on weekday mornings when the number of ticket gate users was the highest, and the degree of deficiency gradually decreased after 18:00. It is considered that the congestion rate affects the degree of deficiency. Therefore, we conducted a correlation analysis between the number of ticket gate users and the degree of deficiency on weekdays and holidays.

19 and 20 are scatter plots showing an example of the correlation between the number of ticket gate users and the degree of deficiency on weekdays and holidays at Kintetsu Nara Station, respectively. From the results of the correlation analysis of FIGS. 19 and 20, a correlation of R2 (coefficient of determination) = 0.46 on weekdays and R2 = 0.57 on holidays is seen, and the degree of deficiency tends to increase as the number of ticket gate users increases. It was observed. In this analysis, both weekdays and holidays, the 6 o'clock level, which has a large degree of deficiency despite the small number of ticket gate users, is excluded.

From the results of FIGS. 19 and 20, there is a relationship between equations (7) and (8) consisting of the following linear equations between the real number (number of ticket gate users) and the degree of deficiency.

Temporarily, the relationship between the degree of deficiency and the actual number of passers-by of ticket gate users is defined by equation (9) consisting of the following linear equation, and equation (9) is substituted for the degree of deficiency in equation (1) above to organize. By doing so, the equation (10) of the linear model for calculating the estimated value of the real number of passing people can be derived from the number of detected terminals. A and b in Eqs. (9) and (19) are coefficients of the degree of deficiency defined by the linear equation.

21 and 22 are graphs showing an example of the result of obtaining the estimated value of the ticket gate user (real number) from the number of detected terminals using the parameter values of the linear model on weekdays and holidays of Kintetsu Nara Station, respectively. For the calculation of equation (10) on weekdays in FIG. 21, the value of the degree of deficiency (a = 0.0002, b = 2.2274), which is the parameter of equation (7) described above, was used. On the other hand, in the calculation of the holiday equation (10) in FIG. 22, the value of the degree of deficiency (a = 0.0002, b = 2.0540), which is the parameter of the above equation (8), was used.

As shown in FIGS. 21 and 22, the actual number of ticket gate users and the estimated value estimated from the number of terminal detections based on the estimation model (linear model) are very consistent with weekdays and holidays. For example, the correlation function r between the actual number of ticket gate users and the estimated value in the case of weekdays in FIG. 21 is 0.97, and the correlation function between the actual number of ticket gate users and the estimated value in the case of holidays in FIG. 22. r is 0.98. Therefore, the above equation (10) is considered to be a "model that predicts a real number from the number of terminal detections (Wi-Fi detection number) at Kintetsu Nara Station", although it is divided into weekdays and holidays.

[Other factors that affect the degree of deficiency]
The above model was built on the premise that the degree of deficiency is affected by the number of people (density), that is, the deficiency rate increases as the number of ticket gate users increases, but the deficiency rate is not limited to ticket gate users but other factors. May be affected by.

FIG. 23 is a graph showing an example of the time transition of the number of ticket gate users and the degree of deficiency on weekdays and holidays created by superimposing FIGS. 17 and 18 described above. From FIG. 23, the degree of deficiency is higher on weekdays as a whole, and especially from 10:00 to 16:00, the degree of deficiency is lower on holidays with a large number of ticket gate users than on weekdays. From this, the degree of deficiency is influenced not only by the number of ticket gate users but also by other factors (for example, behavioral characteristics such as regular or temporary use of stations, walking speed, movement route, etc.). it is conceivable that.

FIG. 24 is a graph showing an example of the time transition of the ratio of regular users on weekdays and holidays at Kintetsu Nara Station. As for the difference between weekdays and holidays, considering the possibility of a difference between users who regularly use the station and users who temporarily use the station such as sightseeing, the ratio of commuter pass users on weekdays and holidays Was compared. The percentage of commuter pass users uses the data of magnetic commuter passes. Looking at FIG. 24, it is considered that the usage rate of the commuter pass is higher on weekdays as a whole, and the behavioral characteristics of the commuter pass users (for example, walking speed and route) may affect the degree of deficiency. Be done. That is, it is possible that the degree of deficiency changes depending on the characteristics of the station, such as a tourist station or a commuting station. From this result, the above estimation model (estimation model) may be constructed according to the types of stations having different characteristics. For example, the above analysis is performed on a plurality of types of stations having different characteristics, such as a station with few passengers, a station with a large number of passengers, a station with many communication users, and a station with many tourist users. The number of ticket gate users may be estimated by constructing a real number prediction model, which is the above estimation model (linear model) different from each other for the stations. In addition, different real number prediction models are constructed according to the type of terminal 85 (for example, whether it is a smartphone or another legacy terminal), the age of the user, the difference in the short-range wireless function ON rate, and the like. The number of ticket gate users may be estimated.

[Machine learning model]
The estimation model used for estimating the number of ticket gate users uses information on the number of terminals detected by the AP device 20 and information on a plurality of types of conditions that affect the estimation of the number of people passing through the automatic ticket gate 90 as environment variables, and is an automatic ticket gate. It may be a machine learning model in which the number of passing people of 90 is used as an objective variable.

The algorithm used for the machine-learned model of this embodiment is not limited to a specific algorithm. For example, as an algorithm of a machine-learned model that learns using supervised learning data, it is possible to use SVR (support vector regression) classified as "Regression" that learns numerical data and predicts numerical values. it can. Instead of this SVR, Linear (Ordinary) Regression, Bayesian Linear Regression, Randam (Decision) Forest, Boosed decision tree, Fast forest quantile, Neural network, Poisson Regression, Support Vector Ordinal Regression, Ridge Regression, Lasso Regression, etc. may be used.

The machine learning model includes, for example, information on the number of detected terminals 85 detected by the AP device 20 within a predetermined time, information on a plurality of types of conditions that affect the estimation of the number of people passing through the automatic ticket gate 90, and an automatic ticket gate. Machine learning is performed using a plurality of learning data including information on the actual number of people passing through 90 (the actual number of ticket gate users). By this machine learning, a machine-learned model can be constructed by determining the values of the parameters of the machine-learning model. By inputting information on the number of terminals detected by the AP device 20 within the estimation target time and information on a plurality of types of conditions that affect the estimation of the number of people passing through the automatic ticket gate 90 into this machine-learned model, the estimation target It is possible to estimate the number of people passing through the automatic ticket gate 90 (the actual number of ticket gate users) within the time.

FIG. 25 is a block diagram showing an example of the configuration of the main part of the information processing apparatus 10 according to the present embodiment. In FIG. 25, the information processing device 10 includes an information acquisition unit 102, a storage unit (DB) 101, and an estimation unit 103. The information acquisition unit 102 provides information on the number of detected terminals 85 detected by the AP device 20 within the estimated target time from the short-range wireless communication access point device (AP device) 20 provided at or near the automatic ticket gate 90. To get.

Further, the information acquisition unit 102 is a ticket gate that includes data of a ticket read / written by the automatic ticket gate 90 in the predetermined period from a plurality of data aggregation PC devices 91 and 92 connected to the automatic ticket gate 90. Information on the number of people passing through the automatic ticket gate 90 (real number) such as count data may be acquired. By automatically and in real time aggregating the ticket gate count data, the parameter values in the above estimation model can be constantly updated. The information acquisition unit 102 may read the data from the data aggregation PC devices 91 and 92 via a recording medium such as a memory (for example, a USB memory), or may read the data from the data aggregation PC devices 91 and 92 from the data aggregation PC devices 91 and 92 to the communication network 35. It may be received and read via such as. The data of the data aggregation PC devices 91 and 92 may be acquired by a real number information acquisition unit provided separately from the information acquisition unit 102. Further, the data of the data aggregation PC devices 91 and 92 may be manually input by the operator by operating the information processing device 10.

The storage unit (DB) 101 stores the values of the parameters (deficiency degree or the coefficients a and b in the calculation formula) of the estimation model that estimates the number of people passing through the automatic ticket gate 90 from the number of detections of the terminal 85. The estimation unit 103 estimates the number of people passing through the automatic ticket gate 90 within the estimation target time based on the estimation model including the value of the parameter and the information on the number of detections of the terminal 85 within the estimation target time.

In the information processing device 10 of FIG. 25, information on the number of detected terminals 85 detected by the AP device 20 within a predetermined time and information on the number of passing persons by totaling the number of people who actually passed through the automatic ticket gate 90 within the predetermined time. Based on the above, a determination unit 104 for determining the value of the parameter of the estimation model may be further provided.

The estimated model may be a linear model in which the number of people passing through is represented by a linear function of the number of detections of the wireless terminal device. Here, the parameter representing the slope of the linear function is the ratio of the number of detected terminals 85 to the actual number of people passing through (counted number of people), and the carrier share rate, which is the market share of the communication carrier corresponding to the terminal 85. , The short-range radio function ON rate of the terminal 85 and the above-mentioned degree of deficiency may be included.

Further, the value of the parameter of the linear model is determined for each of a plurality of different conditions that affect the estimation of the number of people passing through the automatic ticket gate 90, and the estimation unit 103 determines the condition corresponding to the estimation target time from the plurality of conditions. May be selected and an estimation model including the values of the parameters corresponding to the selected conditions may be used to estimate the number of people passing through the automatic ticket gate 90. Here, the item of the above condition may include at least one of the time zone and the day of the week, and includes at least one of the statistical value of the congestion degree near the automatic ticket gate 90 and the walking speed of the person passing through the automatic ticket gate 90. But it may be.

The estimation model has a plurality of types of conditions (for example, time zone, day of the week, congestion degree near the automatic ticket gate 90, automatic ticket gate) that affect the information on the number of detections of the terminal 85 and the estimation of the number of people passing through the automatic ticket gate 90. It may be a machine learning model in which the information of the walking speed of a person passing through the machine 90) is used as an environment variable and the number of people passing through the target location is used as an objective variable. In this case, the information on the number of detected terminals 85 detected by the AP device 20 within a predetermined time, the information on the plurality of types of conditions that affect the estimation of the number of people passing through the automatic ticket gate 90, and the actual number of the automatic ticket gate 90. The value of the parameter of the machine learning model is determined by using a plurality of learning data including information on the number of passing people.

Further, in the information processing device 10 of FIG. 25, a person who has passed through the automatic ticket gate 90 based on the attribute information (for example, gender, age, place of residence) of the user of the terminal 85 detected within the estimated target time. An analysis unit 105 that performs the above-mentioned attribute analysis on the group of Further, in the information processing apparatus 10 of FIG. 25, an analysis unit 105 may be provided instead of the estimation unit 103.

FIG. 26 is a flowchart showing an example of an estimation model determination process (learning phase: S100) in the information processing apparatus 10 according to the present embodiment and an estimation process (estimation phase: S200) of the number of people passing through the automatic ticket gate 90. The learning phase and the estimation phase in the figure may be executed separately at arbitrary timings or continuously.

In the learning phase (S100) of FIG. 26, the information processing device 10 receives a communication log from the AP device 20 via a predetermined interface such as the mobile communication network 30, and the number of terminals detected by the AP device 20 during a predetermined period. (S101).

Further, the information processing device 10 automatically performs the above-mentioned predetermined period based on the ticket data read / written by the automatic ticket gate 90 from the plurality of data aggregation PC devices 91 and 92 connected to the automatic ticket gate 90. Information on the number of people passing through the ticket gate 90 (real number) is acquired (S102). The information processing device 10 may read the data from the data aggregation PC devices 91 and 92 via a recording medium such as a memory, or may receive the data from the data aggregation PC devices 91 and 92 via the communication network 35 and the like. You may read it. Further, the data of the data aggregation PC devices 91 and 92 may be manually input by the operator by operating the information processing device 10.

The information acquisition process of the terminal detection number of the AP device 20 may be performed after the acquisition process of the information on the number of people passing through the automatic ticket gate 90 (real number), or both information acquisition processes may be performed in parallel. You may go with.

Next, the information processing device 10 is based on the information on the number of terminals detected by the AP device 20 and the information on the number of people passing through the automatic ticket gate 90 (real number), and the automatic ticket gate is based on the information on the number of terminals detected by the AP device 20. An estimation model is constructed by determining and storing the values of the parameters of the estimation model such as the above-mentioned linear model or machine learning model that estimates the number of people passing through 90 (S103).

By determining the values of the parameters of the estimation model for each of a plurality of different conditions that affect the estimation of the number of people passing through the automatic ticket gate 90, a plurality of estimation models different for each condition may be constructed. The item of the above condition may include at least one of the time zone and the day of the week, and includes at least one of the statistical value of the congestion degree near the automatic ticket gate 90 and the walking speed of a person passing through the ticket gate of the automatic ticket gate 90. It may be.

In the estimation phase (S200) of FIG. 26, the information processing apparatus 10 estimates the number of people passing through the automatic ticket gate 90 from a plurality of estimation models that are different from each other and are constructed in advance for each condition in the learning phase. (For example, time zone, day of the week, congestion degree near the automatic ticket gate 90, statistical value of walking speed of a person passing through the ticket gate of the automatic ticket gate 90, etc.), one estimation model is selected (for example). S201). When only one estimation model is constructed in the learning phase (S100), the estimation model selection process (S201) may be omitted.

Next, the information processing device 10 receives a communication log from the AP device 20 via a predetermined interface such as the mobile communication network 30, and acquires information on the number of terminals detected during the estimated target period detected by the AP device 20 ( S202).

Next, the information processing device 10 passes through the ticket gate of the automatic ticket gate 90 within the estimation target time based on the selected estimation model and the information on the number of terminals detected during the estimation target period acquired from the AP device 20. The number of passing people is estimated (S203), and the estimation result of the number of passing people is stored in the storage unit (DB) 101 (S204).

The information processing device 10 may output an estimation result of the number of passing people, if necessary. For example, the information processing device 10 may display the estimation result of the number of passing people on the display device or transmit it to an external transmission destination via the communication network.

As described above, according to the present embodiment, complicated data from the automatic ticket gate 90 is obtained by acquiring the number of detections of the terminal 85 based on the communication log of the automatic ticket gate 90 of the station or the AP device 20 installed in the vicinity thereof. The number of people who have passed through the automatic ticket gate 90 can be easily estimated without performing the counting process.
In particular, according to the present embodiment, by constructing a plurality of estimation models different for each condition and using an estimation model suitable for the conditions in the estimation target period, the automatic ticket gate 90 is based on the information on the number of terminals detected in the estimation target period. It is possible to further improve the estimation accuracy when estimating the number of people passing through the ticket gate.

The processing process described in the present specification and components such as a server (information processing device) and a system can be implemented by various means. For example, these steps and components may be implemented in hardware, firmware, software, or a combination thereof.

Regarding hardware implementation, a processing unit or the like used to realize the above steps and components in an entity (for example, a computer device, a seed wireless communication device, a Node B, a terminal, a hard disk drive device, or an optical disk drive device) Means include one or more application-specific ICs (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs). ), Processors, controllers, microcontrollers, microprocessors, electronic devices, other electronic units designed to perform the functions described herein, computers, or combinations thereof. Good.

For firmware and / or software implementation, means such as processing units used to implement the above components are programs (eg, procedures, functions, modules, instructions) that perform the functions described herein. , Etc.) may be implemented. Generally, any computer / processor readable medium that clearly embodies the firmware and / or software code is a means such as a processing unit used to implement the steps and components described herein. May be used to implement. For example, the firmware and / or software code may be stored in memory and executed by a computer or processor, for example, in a control device. The memory may be implemented inside the computer or processor, or may be implemented outside the processor. Further, the firmware and / or software code includes, for example, a random access memory (RAM), a read-only memory (ROM), a non-volatile random access memory (NVRAM), a programmable read-only memory (PROM), and an electrically erasable PROM (EEPROM). ), FLASH memory, floppy (registered trademark) discs, compact discs (CDs), digital versatile discs (DVDs), magnetic or optical data storage devices, etc. Good. The code may be executed by one or more computers or processors, or the computers or processors may be made to perform functional embodiments described herein.

Further, the medium may be a non-temporary recording medium. Further, the code of the program may be read and executed by a computer, a processor, or another device or device machine, and the format thereof is not limited to a specific format. For example, the code of the program may be any of source code, object code, and binary code, or may be a mixture of two or more of these codes.

Also, the description of the embodiments disclosed herein is provided to allow one of ordinary skill in the art to manufacture or use the disclosure. Various amendments to this disclosure will be readily apparent to those of skill in the art and the general principles defined herein are applicable to other variations without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not limited to the examples and designs described herein, but should be accepted in the broadest range consistent with the principles and novel features disclosed herein.

10: Information processing device 15: Customer management server 20: AP device (access point device)
30: Mobile communication network 35: Communication network 80: User 85: Terminal 85: Wireless terminal device 90: Automatic ticket gates 91, 92: Data aggregation PC device 95: Customer management server 102: Information acquisition unit 103: Estimating unit 104 : Decision unit 105: Analysis unit

Claims (20)

An information processing device that estimates the number of people passing through a target location in a facility.
A detection information acquisition unit that acquires information on the number of detections of the wireless terminal device detected by the access point device within the estimated target time from the access point device for short-range wireless communication provided at or near the target location.
A storage unit that stores the parameter values of the estimation model that estimates the number of people passing through the target location from the number of detections of the wireless terminal device.
An estimation unit that estimates the number of people passing through the target location within the estimation target time based on an estimation model including the value of the parameter and information on the number of detections of the wireless terminal device within the estimation target time. An information processing device characterized by being provided with. In the information processing device of claim 1,
The estimation model is based on information on the number of detected wireless terminal devices detected by the access point device within a predetermined time and information on the number of people who have actually passed through the target location within the predetermined time. An information processing apparatus further comprising a determination unit for determining a value of a parameter of. In the information processing device of claim 2,
An information processing apparatus further comprising a real number information acquisition unit that acquires real number information obtained by totaling the number of people who have actually passed through the target location within the predetermined time. In the information processing device according to any one of claims 1 to 3,
The information processing device is characterized in that the estimated model is a model in which the number of passing persons is represented by a linear function of the number of detections of the wireless terminal device. In the information processing device of claim 4,
The parameters representing the inclination of the linear function are the carrier share rate, which is the market share of the communication carrier corresponding to the wireless terminal device, the short-range wireless function ON rate of the wireless terminal device, the carrier share rate, and the near field. Includes a degree of defect indicating the degree to which the access point device cannot detect the wireless terminal device for reasons other than the distance radio function ON rate.
Based on the information on the number of detected wireless terminal devices detected by the access point device within a predetermined time and the information on the real number, the degree of deficiency is calculated and determined using the following equation (1).
The information processing device is characterized in that the estimation unit estimates the number of people passing by using the following equation (2).

In the information processing device of claim 5,
The degree of deficiency is defined by the following equation (3) having coefficients a and b, and is defined by the following equation (3).
Based on the information on the number of detected wireless terminal devices detected by the access point device within a predetermined time and the information on the real number, the coefficients a and b in the degree of deficiency in the equation (3) are determined.
The estimation unit is an information processing device characterized in that the number of passing persons is estimated using the following equation (4).

In the information processing device according to any one of claims 1 to 6,
The values of the parameters of the estimation model are determined for each of a plurality of different conditions that affect the estimation of the number of people passing through the target location.
The estimation unit is characterized in that a condition corresponding to the estimation target time is selected from the plurality of conditions, and an estimation model including the value of the parameter corresponding to the selected condition is used for estimating the number of passing persons. Information processing device. In the information processing device of claim 7,
An information processing apparatus characterized in that the item of the above condition includes at least one of a time zone and a day of the week. In the information processing device of claim 7 or 8,
The item of the condition is an information processing apparatus including at least one of a statistical value of a degree of congestion in the vicinity of the target location and a walking speed of a person passing through the target location. In the information processing device according to any one of claims 1 to 3,
The estimation model is a machine in which information on the number of detections of the wireless terminal device and information on a plurality of types of conditions affecting the estimation of the number of people passing through the target location are used as environment variables, and the number of people passing through the target location is used as an objective variable. It ’s a learning model,
Information on the number of detected wireless terminal devices detected by the access point device within a predetermined time, information on a plurality of types of conditions that affect the estimation of the number of people passing through the target location, and information on the actual number of people passing through the target location. An information processing apparatus characterized in that the value of a parameter of the machine learning model is determined using a plurality of learning data including the above. In the information processing device according to any one of claims 1 to 10.
Information processing characterized by further including an analysis unit that performs attribute analysis on a group of people who have passed the target location based on the attribute information of the user of the wireless terminal device detected within the estimated target time. apparatus. It is an information processing device that analyzes people who have passed the target location where the user in the facility passes.
A storage unit that stores attribute information of a user of the wireless terminal device for a plurality of wireless terminal devices detected by an access point device for short-range wireless communication provided in the vicinity of the target location.
An analysis unit that performs attribute analysis on a group of people who have passed the target location based on the attribute information of the users of the plurality of wireless terminal devices.
An information processing device characterized by being equipped with. In the information processing device of claim 11 or 12,
The information processing device is characterized in that the attribute information includes at least one information of the user's gender, age, and place of residence. A system including the information processing device according to any one of claims 1 to 13 and the access point device. It is a method of estimating the number of people passing through the target location where the user in the facility passes.
Obtaining information on the number of detected wireless terminal devices detected by the access point device within the estimated target time from the access point device for short-range wireless communication provided at or near the target location.
To store the parameter values of the estimation model that estimates the number of people passing through the target location from the number of detections of the wireless terminal device.
Based on the estimation model including the value of the parameter and the information on the number of detections of the wireless terminal device within the estimation target time, the number of people passing through the target location within the estimation target time is estimated.
A method characterized by including. In the method of claim 15,
The estimation model is based on information on the number of detected wireless terminal devices detected by the access point device within a predetermined time and information on the number of people who have actually passed through the target location within the predetermined time. A method further comprising determining the value of a parameter of. It is a method of analyzing people who have passed the target place where the user in the facility passes.
To store the attribute information of the user of the wireless terminal device for a plurality of wireless terminal devices detected by the access point device for short-range wireless communication provided at or near the target location.
Based on the attribute information of the users of the plurality of wireless terminal devices, attribute analysis is performed on a group of people who have passed the target location, and
A method characterized by including. A program executed by a computer or processor provided in an information processing device that estimates the number of people passing through a target location in a facility.
A program code for acquiring information on the number of detected wireless terminal devices detected by the access point device within the estimated target time from the access point device for short-range wireless communication provided at or near the target location.
A program code for storing the parameter values of the estimation model that estimates the number of people passing through the target location from the number of detections of the wireless terminal device, and
A program for estimating the number of people passing through the target location within the estimated target time based on the estimation model including the value of the parameter and the information on the number of detections of the wireless terminal device within the estimated target time. Code and
A program characterized by including. In the program of claim 18.
The estimation model is based on information on the number of detected wireless terminal devices detected by the access point device within a predetermined time and information on the number of people who have actually passed through the target location within the predetermined time. A program comprising further program code for determining the value of a parameter of. A program executed by a computer or processor provided in an information processing device that analyzes a person who has passed a target location where a user in a facility passes.
A program code for storing attribute information of a user of the wireless terminal device for a plurality of wireless terminal devices detected by an access point device for short-range wireless communication provided at or near the target location.
Based on the attribute information of the users of the plurality of wireless terminal devices, a program code for performing attribute analysis on a group of people who have passed the target location, and
A program characterized by including.

Images