JP2021047673A - Estimation system and estimation method - Google Patents

Abstract

To estimate a movement method suitable for a moving person without provision of information from the moving person.SOLUTION: An estimation system comprises a processor which executes a program, and a storage device which stores the program. The processor executes analysis processing for analyzing an attribute of a moving person within a region on the basis of observation data of the moving person, estimation processing for estimating a specific movement method to be utilized in the case where the moving person moves from the inside of the region to the outside of the region, from among a plurality of movement methods based on the attribute of the moving person analyzed by the analysis processing, and output processing for outputting an estimation result by the estimation processing.SELECTED DRAWING: Figure 1

Description

The present invention relates to an estimation system and an estimation method for estimating information.

With the progress of aging, the number of people with mobility restrictions such as persons with disabilities, the elderly, and people with small children is increasing. Welfare-oriented transportation services for people with mobility restrictions are increasing, but it is required to provide highly convenient services regardless of mobility restrictions, even for railways and buses, which are the main transportation methods. ing. Under these circumstances, route search services that take transfer time into consideration are provided for the boundary areas where transfers occur between transportation services such as stations, but there is no support that is suitable for individual travelers. However, this transfer between transportation services is a heavy burden for people with mobility restrictions and other transportation service users. In order to support such a transfer, a technique for searching a moving method suitable for a moving person based on the preference of the moving person is disclosed (Patent Document 1).

The transportation means information providing server of Patent Document 1 receives a search request for transportation means including at least a departure place, a destination, and a desired arrival time from a user terminal connected via a network. In addition, an evaluation function will be provided for transportation means that uses conditions such as destination, desired arrival time, probe information, and traffic information acquired from the outside as input parameters. Then, the evaluation items of the means of transportation are weighted according to the individual situation of the user, and the evaluation value for each means of transportation is calculated by using the weighted evaluation function, and the means of transportation are selected in descending order of the evaluation value. Generate the arranged transportation means selection screen and present it to the user. The evaluation value of each means of transportation is recalculated and updated in real time each time the value of the input parameter changes.

Japanese Unexamined Patent Publication No. 2016-176903

In Patent Document 1, when a mover searches for a route to a destination, an individual preference when selecting a move method such as time priority or charge priority is set in advance, so that the optimum move method for the mover is set. Can be searched. However, depending on the destination, the migrant has to enter the preference in advance, search for the travel method, and arrange the travel method, so the mobility-restricted person switches to the optimal travel method according to his / her current situation. It's difficult. In addition, although it is possible to reduce the time and effort of migrants by preparing various transportation methods at transit points such as stations in advance, it is necessary to always provide transportation methods that do not know whether the transportation company will be used. Yes, it wastes the management resources of the transportation company.

Therefore, an object of the present invention is to estimate a moving method suitable for a moving person without having the moving person provide information.

An estimation system that is one aspect of the invention disclosed in the present application includes a processor that executes a program and a storage device that stores the program, and the processor is based on observation data of a mover in an area. , The mover moves from the inside of the area to the outside of the area from a plurality of movement methods based on the analysis process for analyzing the attributes of the mover and the attributes of the mover analyzed by the analysis process. It is characterized in that an estimation process for estimating a specific movement method to be used in the case of the above and an output process for outputting the estimation result by the estimation process are executed.

According to a typical embodiment of the present invention, it is possible to estimate a moving method suitable for a moving person without having the moving person provide information. Issues, configurations and effects other than those described above will be clarified by the description of the following examples.

FIG. 1 is an explanatory diagram showing a system configuration example of a transportation service cooperation system. FIG. 2 is an explanatory diagram showing an example of a method of observing a moving person using a plurality of image sensors and acoustic sensors. FIG. 3 is a block diagram showing a hardware configuration example of the transportation service cooperation system. FIG. 4 is a flowchart showing an example of the attribute analysis processing procedure by the attribute analysis unit. FIG. 5 is an explanatory diagram showing an example of the attribute environment information DB of the migrant output by the attribute analysis unit. FIG. 6 is a flowchart showing an example of a processing procedure for defining a movement method estimation model used for estimating the movement method. FIG. 7 is an explanatory diagram showing an example of a movement method estimation model. FIG. 8 is an explanatory diagram showing an example of the stored contents of the pattern DB. FIG. 9 is an explanatory diagram showing an example of usage results. FIG. 10 is an explanatory diagram showing a calculation example of the selection probabilities of all the movement methods for all the patterns calculated by the movement method estimation unit. FIG. 11 is a flowchart showing an example of a movement method estimation processing procedure by the movement method estimation unit.

Hereinafter, the movement method required by the migrant is estimated based on the observation data from a plurality of image sensors and acoustic sensors and provided to the transportation company, and the transportation company provides the migrant with an appropriate movement method. The transportation service cooperation system will be described using drawings.

<System configuration example of transportation service cooperation system>
FIG. 1 is an explanatory diagram showing a system configuration example of a transportation service cooperation system. The traffic service cooperation system 100 includes, for example, a plurality of image sensors 101a and 101b, a plurality of acoustic sensors 102a to 102e, a movement method estimation system 110, and a transportation operator system 120. When the image sensors 101a and 101b are not distinguished, they are simply referred to as the image sensor 101. The number of image sensors 101 is not limited to two, and may be three or more. When the acoustic sensors 102a to 102e are not distinguished, it is simply referred to as the acoustic sensor 102. The number of acoustic sensors 102 is not limited to 5, and may be 2 or more, 4 or less, or 6 or more.

The movement method estimation system 110 is based on the observation data from the image sensors 101a and 101b and the acoustic sensors 102a to 102e arranged at the station, and the environmental information such as the weather around the station provided by the external service 103. Estimate the movement method that 130 will use next.

The migrant 130 may carry a luggage 131 such as a trunk case. The movement method estimation system 110 transmits the estimation result to the transportation operator system 120 of the transportation operator around the station. The transportation operator system 120 receives the estimation result of the movement method from the movement method estimation system 110, and provides the movement method 140 to the movement person.

In Fig. 1, the station yard is estimated, but it is not limited to the station yard, but areas within various facilities such as airports, ports, office buildings, commercial buildings, exhibition halls, parks, museums, art galleries, stadiums, and concert halls. But it may be.

<Example of observing method of migrants>
FIG. 2 is an explanatory diagram showing an example of a method of observing a moving person using a plurality of image sensors 101 and an acoustic sensor 102. The image sensors 101a and 101b and the acoustic sensors 102a and 102e are installed on the ceiling and walls of the station yard so that the observation area is formed toward the area where the migrants 130a and 130b are expected to exist.

In FIG. 2, two image sensors 101 and five acoustic sensors 102 are installed near the ticket gates 201a to 201c of the station in order to provide the following movement method to the migrants leaving the station. The number and location of installation can be set arbitrarily.

<Hardware configuration example of transportation service cooperation system 100>
FIG. 3 is a block diagram showing a hardware configuration example of the transportation service cooperation system 100. The moving method estimation system 110 includes a communication I / F (Interface) 301, a CPU (Central Processing Unit) 302, a storage device 303, and a bus 304.

The communication I / F 301 is communicably connected to the image sensor 101, the acoustic sensor 102, the external service 103, and the transportation company system 120 via a network 320 such as LAN (Local Area Network), WAN (Wide Area Network), and the Internet. Will be done. The bus 304 connects the communication I / F 301, the CPU 302, and the storage device 303 in a communicable manner.

The CPU 302 is a processor that controls the movement method estimation system 110 in an integrated manner. The storage device 303 stores various programs including the movement method estimation program 305, and various data such as observation data 355, attribute environment information DB 356, movement method estimation model DB 357, estimation result 358, and pattern information 359, which are non-temporary or temporary. Recording medium. Examples of the storage device 303 include a ROM (Read Only Memory), a RAM (Random Access Memory), an HDD (Hard Disk Drive), and a flash memory.

When the movement method estimation program 305 is executed by the CPU 302, the CPU 302 functions as an attribute analysis unit 350, a movement method estimation unit 351, an information cooperation unit 352, an external cooperation unit 353, and an estimation model update unit 354.

The attribute analysis unit 350 analyzes the observation data 355 of the image sensor 101 and the acoustic sensor 102, and stores the attribute environment information of the migrant in the attribute environment information DB 356. Here, the observation data 355 is a set of image data acquired from the image sensor 101 and audio data acquired from the acoustic sensor 102.

The attribute environment information is information in which the environment information obtained from the external service 103 is associated with the attribute information that is a feature of the migrant 130. The characteristics of the migrant 130 mean, for example, the difficulty of movement (eg, wheelchair use, cane use), the amount of luggage, the need for foreign language support, and the size of the group with which they are acting.

The external cooperation unit 353 collects the weather information 520 around the area where the migrant 130 is located from the external service 103 that distributes the weather information 520. Here, the weather information 520 means, for example, the weather and the temperature, and is associated with the attribute information as the environmental information.

The moving method estimation unit 351 estimates the moving method 140 to be used next by the moving person 130 based on the attribute environment information of the moving person 130. The information cooperation unit 352 links information with the transportation company system 120. Specifically, for example, the information cooperation unit 352 transmits the movement method 140 estimated by the movement method estimation unit 351 to the transportation operator system 120.

The estimation model update unit 354 receives the usage record 362 of the movement method 140 estimated by the movement method estimation unit 351 from the transportation company system 120, and estimates the movement method based on the estimation result 358 of the movement method 140 and the usage record 362. The movement method estimation model DB357 used in the unit 351 is updated. The movement method estimation model DB 357 is a database that stores a movement method estimation model that represents the probability that the mover 130 selects each movement method 140. By reflecting the usage record 362, the estimation accuracy of the moving method 140 can be improved.

The pattern DB 359 is a database that stores patterns of the migrant 130. The pattern of the migrant 130 is information that defines the type of combination of the attribute information of the migrant 130 and the environmental information.

The transportation operator system 120 has a communication I / F 311 and a CPU 312 and a storage device 313. The communication I / F 311 is communicably connected to the external service 103 and the transportation operator system 120 via the network 320. The bus 314 connects the communication I / F 311, the CPU 312, and the storage device 313 in a communicable manner.

The CPU 312 is a processor that controls the transportation operator system 120 in an integrated manner. The storage device 313 is a non-temporary or temporary recording medium that stores various programs including the performance management program 315 and the vehicle allocation management program 316, and various data such as the estimation result 358, the usage record 362, and the movement method resource 372. .. The storage device 313 includes, for example, a ROM, a RAM, an HDD, and a flash memory.

When the performance management program 315 is executed by the CPU 312, the CPU 312 functions as the performance management unit 361. The performance management unit 361 manages the movement method 140 estimated by the movement method estimation program 305.

When the vehicle allocation management program 316 is executed by the CPU 312, the CPU 312 functions as the vehicle allocation management unit 371. The vehicle allocation management unit 371 arranges the provision of the moving method 140 from the performance management unit 361, collects the usage record 362, and transmits it to the performance management unit 361.

<Attribute analysis processing by attribute analysis unit 350>
FIG. 4 is a flowchart showing an example of the attribute analysis processing procedure by the attribute analysis unit 350. FIG. 5 is an explanatory diagram showing an example of the attribute environment information DB 356 of the mover 130 output by the attribute analysis unit 350. The attribute environment information DB 356 has a mover ID 501, environment information 502, and attribute information 503 as fields.

The mover ID 501 is a field that stores identification information (mover ID) that uniquely identifies the mover as a value. The environment information 502 is a field that stores the weather information 520 from the external service 103 as a value. By considering the weather information 520, the moving method estimation system 110 can estimate the moving method 140 in which the moving person 130 is not adversely affected by the weather information 520. For example, if the bus stop in front of the station has no roof and the taxi stop has a roof and the weather information 520 is rainy, the movement method estimation system 110 uses a movement method in which the taxi is more suitable for the migrant 130 than the bus. Presumed to be.

The attribute information 503 has, as values, a group size 531 of the mover 130 specified by the mover ID (example: the number of the mover 130), a luggage amount 532 (example: the number of luggage), and a movement difficulty 533 (example: number of people). This field stores wheelchairs (whether or not wheelchairs are used) and foreign language support 534 (eg, foreign language support is required or not required).

By considering the group size 531 the movement method estimation system 110 can estimate the movement method 140 suitable for transferring the number of migrants 130. By considering the baggage amount 532, the moving method estimation system 110 can estimate the moving method 140 suitable for the transfer of the moving person 130 and the moving person 130 with the baggage amount 532. By considering the movement difficulty 533, the movement method estimation system 110 can estimate the movement method 140 in consideration of the movement difficulty 533 of the mover 130. By considering the foreign language correspondence 534, the movement method estimation system 110 can estimate the movement method 140 in which the foreign language correspondence 534 of the migrant 130 is possible or unnecessary.

[S401]
Returning to FIG. 4, the attribute analysis unit 350 acquires the observation data from the image sensor, that is, the video data (step S401). The video data is used, for example, to identify the attribute information 503 in image recognition.

[S402]
Next, the attribute analysis unit 350 detects the target of the moving method 140 from the observation data acquired in step S401 (step S402). Specifically, for example, the attribute analysis unit 350 extracts an area in which the mover 130 is photographed on the frame image of the video data acquired in step S401.

More specifically, the attribute analysis unit 350 extracts the difference between the frame image of the video data acquired in step S401 and the frame image of the video data observed by the same image sensor 101 in the time zone when there is no migrant 130 in advance. By doing so, the area where the moving person 130 is photographed is extracted. In addition, instead of using the previously observed video data, for the video data acquired in step S401, frame images with different observation times are used to determine the image shift width and direction when the correlation between the frame images is maximized. By calculating, the presence / absence of the moving person 130 and the moving direction thereof may be estimated.

When the mover 130 is detected as the target of providing the move method 140, the attribute analysis unit 350 newly issues the mover ID 501 to generate an entry in the attribute environment information DB. Then, the attribute analysis unit 530 stores the group scale 531 in the entry for the mover ID 501.

[S403]
Next, the attribute analysis unit 350 determines the movement-restricted person (step S403). Specifically, for example, the attribute analysis unit 350 moves with difficulty because the mover 130 is using a wheelchair or has a large baggage based on the feature amount of the area extracted in step S402. Determine if you are a constrained person. Here, as the feature amount, the number of pixels and the shape of the extracted region occupied in the frame image of the video data, and the aspect ratio of the region are used.

The moving method estimation system 110 uses an image sensor 101 to photograph a moving person 130 in a wheelchair or a moving person 130 with a large luggage in advance. The attribute analysis unit 350 calculates the number of pixels, the shape, and the aspect ratio of the shooting area of the moving person 130 in the frame image, and determines the movement-restricted person by comparing with the values. As a result, the movement difficulty 531 and the luggage amount 532 are set for the mover 130.

[S404]
Next, the attribute analysis unit 350 estimates the position of the mover 130 (step S404). Specifically, for example, the attribute analysis unit 350 calculates the position information of the mover 130 extracted in step S402 based on the information on the installation position of the image sensor 101, the shooting direction, and the specifications of the image sensor 101. Here, the installation position of the image sensor 101 and the position of the migrant 130 are set to positions on the coordinate axes determined in the station yard.

[S405]
Next, the attribute analysis unit 350 acquires acoustic data which is observation data of the acoustic sensor 102 (step S405). Specifically, for example, the attribute analysis unit 350 acquires an acoustic signal which is observation data of the acoustic sensor 102 installed around the image sensor 101 in which the moving person 130 is photographed.

[S406]
Next, the attribute analysis unit 350 executes sound source separation (step S406). Specifically, for example, the attribute analysis unit 350 extracts an acoustic signal from the position of the mover 130 estimated in step S404. More specifically, the attribute analysis unit 350 adjusts the phase of the acoustic signal observed by the plurality of acoustic sensors 102, and adds the phase of the acoustic signal arriving from the estimation direction of the position of the mover 130 as the same phase. , Maximize the gain of the acoustic sensor 102 in the estimation direction of the position of the mover 130. As a result, the attribute analysis unit 350 emphasizes the acoustic signal from the position of the mover 130 and selectively extracts it.

In addition, the attribute analysis unit 350 uses the difference between the noise level estimated from the acoustic signal observed by the same acoustic sensor 102 in advance and the level of the acoustic signal acquired in step S405 during the time when the migrant 130 is absent, and the acoustic of the migrant 130. It is also possible to extract a signal or estimate the acoustic signal of the mover 130 from the time variation of the frequency spectrum obtained by a frequency analysis method such as short-time Fourier transform.

[S407]
Next, the attribute analysis unit 350 executes the characteristic sound analysis (step S407). Specifically, for example, the attribute analysis unit 350 assumes that the acoustic signal extracted in step S406 is a human voice, analyzes it by a voice recognition method, and estimates which language it is. Here, as a speech recognition method, in which language, the frequency spectrum of the time series is used as the feature amount of the acoustic signal, and the feature amount of the standard word prepared in advance and the feature amount of the observed acoustic signal are correlated. Various methods can be used, such as estimating whether it is a word. As a result, a foreign language correspondence 534 is set for the migrant 130.

When the correlation with the word is low, the attribute analysis unit 350 analyzes the frequency component included in the acoustic signal by a frequency analysis method such as a short-time Fourier transform. In the striking sound like making a white cane, paying attention to the fact that the range of the frequency component included is large, it is determined whether or not the sound like making a white cane can be heard from the estimated position direction of the moving person 130. As a result, a difficulty in moving 533 (for example, using a white cane) is set for the moving person 130.

[S408]
Next, the attribute analysis unit 350 executes the movement 130 tracking (step S408). Specifically, for example, the attribute analysis unit 350 tracks the movement of the migrant 130 detected in step S402. More specifically, the attribute analysis unit 350 analyzes the correlation between the region detected as the mover 130 in step S402 and the region on the continuous frame image, and the mover 130 moves to the region where the correlation is maximum. Judge that it has moved. The attribute analysis unit 350 tracks the movement of the mover 130 by repeating this process on a continuous frame image.

[S409]
Next, the attribute analysis unit 350 executes group determination (step S409). Specifically, for example, when the attribute analysis unit 350 detects a plurality of migrants 130 in step S402, it determines whether the group is a group in which the plurality of migrants 130 are acting together. More specifically, for example, the attribute analysis unit 350 executes the process of step S408 for all the migrants 130 detected in step S402.

The attribute analysis unit 350 determines that the migrant 130 is acting as a group when the migrant 130 having the same luggage amount 532 and foreign language support 534 is moving in the same direction at a preset fixed time. To do. As a result, the migrants 130 are merged and the group scale 531 is stored.

[S410]
Next, the attribute analysis unit 350 creates attribute data (step S410). Specifically, for example, the attribute analysis unit 350 creates a mover ID 501 that uniquely identifies the mover 130 detected in step S402, and summarizes the data set in steps S403, S407, and S409 for the mover ID 501. The value of the attribute information 503 is created and used as the attribute data.

[S411]
Next, the attribute analysis unit 350 stores the attribute data created in step S410 in the attribute information 503 of the attribute environment information DB 356. Further, the external cooperation unit 353 acquires the weather information 520 around the station by using a Web API (Application Programming Interface) provided by the external service 103 that distributes the weather information 520. As the weather information 520, the weather information and the temperature information as to whether or not it is raining are acquired. It is also possible to use the sensible temperature index instead of the air temperature. The attribute analysis unit 350 associates this information with the attribute data as the environment information 502 and stores it in the attribute environment information DB 356.

<Definition processing of the movement method estimation model by the movement method estimation unit 351>
FIG. 6 is a flowchart showing an example of a processing procedure for defining a moving method estimation model used for estimating the moving method 140. The process of FIG. 6 is executed when the moving method estimation unit 351 is used for the first time, and when the attribute information 503 and the environment information 502 of the moving person 130 used for estimating the moving method 140 are changed. It is executed each time the attribute analysis unit 350 detects the mover 130.

[S601]
First, the movement method estimation unit 351 assumes a movement method estimation model P (step S601). The movement method estimation model DB357 stores the probability Pi that the mover 130 selects each movement method 140. For example, the probability Pi that the mover 130 selects the movement method i is defined as the probability that the utility function Ui of the movement method i is larger than the utility function Uj of all other movement methods j (i, j = 1,). 2, ..., N, i ≠ j, N is the number of selectable movement methods 140). Here, the utility function Ui is defined by the linear sum of the explanatory variables Xhi, which are factors that determine the utility of the moving method i (coefficients Wh, h = 1, 2, ..., M, M are factors that determine the utility. Number of). Here, the movement method estimation model P will be described.

[Movement method estimation model P]
FIG. 7 is an explanatory diagram showing an example of the movement method estimation model P. In the movement method estimation model P, an explanatory variable Xhi is defined for each movement method 140. i indicates a moving method 140. The moving method i defines, for example, a minibus 711 (i = 1), a small taxi 712 (i = 2), a large taxi 713 (i = 3), a ride share 714 (i = 4), and the like. The explanatory variable Xhi defines, for example, convenience 701, number of passengers 702, loadability of luggage 703, ease of getting on and off 704, and ease of use by foreigners 705 for each movement method i.

Convenience 701 is an explanatory variable X1i indicating a numerical value that relatively evaluates how useful each moving method 140 is in a situation where the difficulty of moving increases due to the influence of the natural environment such as rain. The larger the number, the higher the convenience. The number of passengers 702 and the loadability of luggage 703 are explanatory variables X2i and X3i indicating the number of passengers and the amount of luggage that can be loaded, which are defined as the specifications of each movement method 140, respectively. The larger the number, the larger the number of people and the load capacity. As the definition of the amount of luggage, the capacity of the trunk for loading the luggage and the number of luggage such as the maximum number of luggage that can be loaded in the case of a suitcase are used.

The ease of getting on and off 704 is an explanatory variable X4i indicating a numerical value indicating the ease of use when used by a migrant 130 who has difficulty walking, based on the structural difference of each moving method 140. .. The larger the number, the higher the difficulty. The ease of use for foreigners 705 is an explanatory variable X5i indicating a numerical value relatively evaluated from the status of efforts for supporting foreign users for each movement method 140. The larger the number, the easier it is (that is, the higher the degree of correspondence with foreign languages). The movement method estimation model P is defined for each pattern of the mover 130. The pattern of the migrant 130 is defined as a combination pattern of the attribute information 503 of the migrant 130 and the environment information 502.

[Pattern DB]
FIG. 8 is an explanatory diagram showing an example of the stored contents of the pattern DB. The pattern DB 359 is a database that stores patterns of the migrant 130. The pattern of the migrant 130 is information that defines the type of combination of the attribute information of the migrant 130 and the environmental information. As shown in FIG. 8, the group size 531 is {large, small}, the luggage amount 532 is {large, small}, the difficulty of movement is {yes, no}, the foreign language support 534 is {necessary, unnecessary}, and the weather information. When 520 is defined as a binary value that can be taken by each item of attribute information 503 and environment information 502 in the form of {good, bad}, the pattern 801 of the mover 130 including the environment information 502 becomes 32 patterns. Become.

Regarding the possible values of each item, the group size 531 is "large" when there are 5 or more people, the luggage amount 532 is "large" when you have one or more suitcases, and the mobility difficulty 533 is a wheelchair or Defined as "Yes" when using a cane, "Required" for foreigners, and "Good" when the weather is sunny or cloudy and the temperature is 20 to 25 degrees Celsius. To do.

[S602]
Returning to FIG. 6, the information cooperation unit 352 acquires the usage record 362 from the transportation company system 120 (step S602). FIG. 9 shows an example of the usage record 362.

[Usage record 362]
FIG. 9 is an explanatory diagram showing an example of the usage record 362. The usage record 362 has a performance management ID 901, a usage date 902, a movement method 903, and a user attribute 904 as fields. The performance management ID 901 is a field that stores identification information that uniquely identifies the usage record 362 as a value in order to manage the usage record 362.

The usage date 902 is a field for storing the date and time using the moving method 903 as a value. The movement method 903 is a field for storing the method of moving to the usage date 902 as a value. The user attribute 904 is a field that stores, as values, the group size 941 of the migrant 130 who used the movement method 903 on the usage day 902, the amount of luggage 942, the difficulty of movement 943, and the foreign language correspondence 944.

[S603]
Returning to FIG. 6, the external cooperation unit 353 acquires the external environment data (step S603). Specifically, for example, the external cooperation unit 353 uses the external service 103 and provides the weather information 520 of the usage day 902 (the day when the transportation company around the station provides the transportation method 140) included in the usage record 362. get.

[S604]
The movement method estimation unit 351 estimates the parameters of the movement method estimation model (step S604). Specifically, for example, the movement method estimation unit 351 selects one pattern 801 of the mover 130 from the pattern DB, and uses the entry of the usage record 362 corresponding to the selection pattern in the acquired usage record 362. Then, the coefficient Wh (h = 1, 2, ..., M, M is the number of factors that determine the utility) that maximizes the likelihood function defined based on the movement method estimation model P is calculated.

The likelihood function is defined by, for example, the following equation (1).

_{The moving method estimation unit 351 obtains the value of uki on} the right side of the above equation (1) from the usage record 362. For example, when the mover k uses the move method i, uki = 1, and when the mover k does not use the move method i, uki = 0. In the above formula (1), Pi is the probability that the migrant 130 selects the moving method i, and is defined by, for example, the following formula (2).

In the above equation (2), Ui is a utility function of the moving method i, and Ui'is a utility function that does not include the error εi, and is defined by, for example, the following equations (3) and (4).

Regarding the calculation of the coefficient (weight) Wh, the moving method estimation unit 351 substitutes the values of Pi (that is, Xhi) and uki into the likelihood function L, and obtains the coefficient Wh that maximizes the likelihood function L.

Further, the movement method estimation unit 351 determines the value of the explanatory variable Xhi using the movement method estimation model P shown in FIG. 7. For example, when the migrant 130 uses the minibus (i = 1), Xh1 is as follows.
Xh1 = {1 (convenience), 20 (number of passengers), 10 (loadability of luggage), 1 (easiness of getting on and off), ...}

Here, a corresponding example of the pattern 801 is shown. For example, in the case of the entry in the first line of the usage record 362 shown in FIG. 9, since the group size 941 is "5 people", the group size 531 is "large" and the luggage amount 942 is "5 suitcases". Therefore, since the cargo amount 532 is "large" and the movement difficulty 943 is "not applicable", the movement difficulty 533 is "none" and the foreign language support 944 is "unnecessary", so the foreign language support 534 is "not applicable". It becomes "unnecessary".

Further, assuming that the external cooperation unit 353 acquires "weather: sunny, temperature: 23 degrees" as the weather information 520 of "2018 (year) 12 (Monday) 03 (Sunday)" on the usage day 902, the pattern DB 359 is used. The weather information 520 is "good". Therefore, the entry in the first line of the usage record 362 shown in FIG. 9 corresponds to "Pattern 1" of the pattern 801 of the pattern DB 359.

[S605]
The movement method estimation unit 351 calculates the selection probability of the movement method i (step S605). Specifically, for example, the movement method estimation unit 351 substitutes the value of the coefficient Wh and the explanatory variable Xhi calculated in step S604 into the equation (4), and the movement method for the pattern 801 of the mover 130 selected in step S604. The selection probability Pi of i is calculated by the equation (2).

[S606]
The movement method estimation unit 351 calculates the selection probabilities P1 to PN for all the movement methods 1 to N (step S606: No, S605). Specifically, for example, the movement method estimation unit 351 repeats step S605 for all movement methods 1 to N, and selects all movement methods 1 to N for the pattern 801 of the mover 130 selected in step S604. Calculate the probabilities P1 to PN.

[S607]
The moving method estimation unit 351 calculates the selection probabilities P1 to PN of all the moving methods 1 to N for all the patterns (patterns 1 to 32) of the moving person 130 (steps S607: No, S604, S605, S606). Specifically, for example, the movement method estimation unit 351 repeats steps S604 to S606 for all patterns 801 of the mover 130.

FIG. 10 is an explanatory diagram showing an example of calculating the selection probabilities of all the moving methods 140 with respect to all the patterns calculated by the moving method estimating unit 351. As illustrated in FIG. 10, the movement method estimation unit 351 calculates the selection probabilities 1000 of all the movement methods 140 for each pattern of the mover 130.

[S608]
Returning to FIG. 6, the moving method estimation unit 351 stores in the storage device 303 the selection probabilities 1000 of all the moving methods 140 for all the patterns of the moving person 130 created in step S607.

<Estimation processing of movement method estimation by movement method estimation unit 351>
FIG. 11 is a flowchart showing an example of a movement method estimation processing procedure by the movement method estimation unit 351.

[S1101]
The movement method estimation unit 351 acquires the attribute data of the mover 130 (step S1101). The attribute data of the mover 130 is the value of the attribute information 503 corresponding to the mover ID 501 stored in FIG.

[S1102]
The movement method estimation unit 351 acquires the current weather information 520 around the station collected by the external cooperation unit 353 from the storage device 303.

[S1103]
The moving method estimation unit 351 determines the pattern of the moving person 130 (step S1103). Specifically, for example, the movement method estimation unit 351 determines which pattern 801 the mover 130 detected in step S402 corresponds to based on the attribute information 503 and the environment information 502.

As shown in FIG. 8, the attribute information 503 and the environment information 502 are binary, and the pattern 801 of the migrant 130 is defined. The movement method estimation unit 351 has detected the attribute information 503 (531 to 534) of the migrant 130 and the observed weather for the attribute information 503 and the environment information 502 based on the definition of the binary value shown in the explanation of step S601. It is determined which value the information 520 has.

[S1104]
The moving method estimation unit 351 acquires the selection probability of the moving method 140 (step S1104). Specifically, for example, the moving method estimation unit 351 acquires the selection probability 1000 of the moving method 140 corresponding to the detected pattern 801 of the moving person 130 from the storage device 303.

[S1105]
The movement method estimation unit 351 repeatedly executes step S1104 for all movement methods 140, and acquires the selection probabilities of all movement methods 140 for the pattern 801 of the target mover 130 (step S1105: No, S1104).

[S1106]
The movement method estimation unit 351 compares the selection probabilities 1000 of each movement method 140 acquired in step S1105, and selects the movement method 140 having the maximum selection probability 1000. The moving method estimation unit 351 determines the selected moving method 140 as the estimation result of the moving method 140 suitable for the target moving person 130.

[S1107]
The movement method estimation unit 351 stores the estimation result of the movement method 140 in step S1106 in the storage device 303 (step S1107).

The information cooperation unit 352 transmits the estimated movement method 140 (estimation result) to the transportation company system 120 by using the Web API or the like provided by the transportation company system 120. The information cooperation unit 352 may add the position information of the migrant 130 to the estimation result to be transmitted. As the position information of the migrant 130, the position information of the image sensor 101 acquired by the attribute analysis unit 350 is used.

As a result, it is possible to know from which direction the migrant 130 comes out, and to narrow down the movement methods that can be provided by the transportation operator system 120. Further, the information cooperation unit 352 provides a Web API or the like to the transportation company system 120, receives the usage record 362 of the moving method 140, and stores it in the storage device 303.

In the transportation company system 120, the performance management unit 361 stores the estimated result of the received movement method 140 in the storage device 313, and uses the Web API or the like provided by the vehicle allocation management unit 371 to obtain the movement method 140. Send the estimation result. In addition, the performance management unit 361 transmits the usage record 362 using the Web API or the like provided by the information cooperation unit 352.

The vehicle allocation management unit 371 actually arranges the estimated movement method 140. At this time, if the selection probability of the moving method 140 is lower than a certain threshold value, a determination process such as not arranging may be performed. When the vehicle is actually dispatched, the vehicle allocation management unit 371 collects the attributes of the migrant 130 used for the movement method 140 and the environmental information at that time as the usage record 362, and the Web API provided by the record management unit 361. It is transmitted to the performance management unit 361 by using the above. Here, the usage record 362 can be data in which the attributes of the migrant 130 used and the environmental information at the time of use are combined.

As described above, the moving method 140 suitable for the moving person 130 is estimated by using the attribute information of the moving person 130 and the environmental information around the moving person 130 obtained from the observation data of the plurality of image sensors and the acoustic sensors. As a result, it is possible to support the movement of the migrant 130 by coordinating a plurality of transportation means, and it is possible to support the transportation business operator to provide a service that effectively utilizes the limited resources.

The present invention is not limited to the above-described embodiment, and includes various modifications and equivalent configurations within the scope of the attached claims. For example, the above-described examples have been described in detail in order to explain the present invention in an easy-to-understand manner, and the present invention is not necessarily limited to those having all the described configurations. Further, a part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Further, the configuration of another embodiment may be added to the configuration of one embodiment. In addition, other configurations may be added, deleted, or replaced with respect to a part of the configurations of each embodiment.

Further, each of the above-described configurations, functions, processing units, processing means, etc. may be realized by hardware by designing a part or all of them by, for example, an integrated circuit, and the processor realizes each function. It may be realized by software by interpreting and executing the program to be executed.

Information such as programs, tables, and files that realize each function is recorded in a memory, hard disk, storage device such as SSD (Solid State Drive), or IC (Integrated Circuit) card, SD card, DVD (Digital Versaille Disc). It can be stored in a medium.

In addition, the control lines and information lines indicate those that are considered necessary for explanation, and do not necessarily indicate all the control lines and information lines necessary for implementation. In practice, it can be considered that almost all configurations are interconnected.

P Movement method estimation model Pi Selection probability Ui Utility function Xhi Explanatory variable 100 Transportation service cooperation system 103 External service 110 Movement method estimation system 120 Transportation operator system 130 Mobility 131 Luggage 140 Movement method 305 Movement method estimation program 315 Performance management program 316 Vehicle allocation management program 350 Attribute analysis unit 351 Movement method estimation unit 352 Information cooperation unit 353 External cooperation unit 354 Estimation model update unit 356 Attribute environment information DB
357 Movement method estimation model DB
359 pattern DB
361 Performance Management Department 362 Usage Performance 371 Vehicle Allocation Management Department 372 Travel Method Resource 501 Traveler ID

Claims (12)

It has a processor that executes a program and a storage device that stores the program.
The processor
Analytical processing that analyzes the attributes of migrants based on the observation data of migrants in the area,
Based on the attributes of the mover analyzed by the analysis process, a specific move method to be used when the mover moves from the inside of the region to the outside of the region is estimated from a plurality of movement methods. Estimate processing and
Output processing that outputs the estimation result by the estimation processing and
An estimation system characterized by running. The estimation system according to claim 1.
In the analysis process, the processor analyzes the size of the group in which the mover acts together as the attribute based on the video data obtained by capturing the image of the mover, which is the observation data.
An estimation system characterized by that. The estimation system according to claim 1.
In the analysis process, the processor analyzes the amount of luggage possessed by the mover as the attribute based on the video data obtained by capturing the image of the mover, which is the observation data.
An estimation system characterized by that. The estimation system according to claim 1.
In the analysis process, the processor analyzes the necessity of foreign language correspondence for the mover as the attribute based on the observation data, the acoustic data about the mover.
An estimation system characterized by that. The estimation system according to claim 1.
In the analysis process, the processor analyzes the movement difficulty of the mover as the attribute based on at least one of the observation data, the video data obtained by capturing the mover, and the acoustic data related to the mover. ,
An estimation system characterized by that. The estimation system according to claim 1.
In the estimation process, the processor calculates a selection probability for each movement method based on the utility of the movement method, and estimates the specific movement method based on the selection probability for each movement method.
An estimation system characterized by that. The estimation system according to claim 6.
In the estimation process, the processor estimates the method of maximum selection probability to the particular movement method.
An estimation system characterized by that. The estimation system according to claim 6.
The processor
An update process is executed to update the utility of the movement method based on the usage record of the movement method.
In the estimation process, the processor calculates a selection probability for each of the movement methods based on the utility of the movement method after the update by the update process.
An estimation system characterized by that. The estimation system according to claim 1.
In the analysis process, the processor analyzes the position of the mover based on the position of the sensor that outputs the observation data.
In the output process, the processor outputs information indicating the position of the mover.
An estimation system characterized by that. The estimation system according to claim 1.
In the estimation process, the processor estimates the specific movement method based on the weather information around the area.
An estimation system characterized by that. The estimation system according to claim 10.
The weather information includes at least one of weather or temperature.
An estimation system characterized by that. An estimation method executed by an estimation system having a processor that executes a program and a storage device that stores the program.
The processor
An analysis process that analyzes the attributes of the migrants based on the observation data of the migrants in the area, and
Based on the attributes of the mover analyzed by the analysis process, a specific move method to be used when the mover moves from the inside of the region to the outside of the region is estimated from a plurality of movement methods. Estimate processing and
Output processing that outputs the estimation result by the estimation processing and
An estimation method characterized by performing.

Images