JP7621233B2 - Movement/stay estimation model and method incorporating area characteristics, and movement trajectory prediction device and program - Google Patents

Description

The present invention relates to a technology that uses sensing data to estimate the behavior of an object (e.g., a person) as it moves or stays there.

In recent years, in the field of transportation policy, the Activity Based Model (ABM) has been attracting attention. It estimates human activities and predicts the resulting traffic behavior of those activities. Specifically, ABM is a traffic behavior prediction model that simulates the activities of individual people based on human attributes such as place of residence, age, sex, and occupation, as well as traffic conditions and regional characteristics, and then compiles and predicts the overall traffic volume that may result from the simulations.

For example, the transportation demand forecasting technology disclosed in Non-Patent Document 1 uses this type of ABM to forecast not only the transportation demand that already exists, but also the transportation demand that will arise for the first time when new facilities and transportation systems are built.

Here, ABM training requires learning data that includes activity details, and this activity detail is obtained from survey results such as the Ministry of Land, Infrastructure, Transport and Tourism's Person Trip (PT) Survey, which collects detailed transportation data (trip data) such as the starting point, destination, purpose, means of transportation, and time of activity through questionnaire surveys.

Therefore, such questionnaire surveys are essential to constructing an ABM. For example, the reality is that an ABM cannot be constructed using only location measurement information from communication devices such as smartphones, which are expected to be used for traffic demand forecasting.

In response to this, Non-Patent Document 2 discloses a location prediction technology that applies BERT (Bidirectional Encoder Representations from Transformers), a well-known natural language processing model. In this technology, BERT is pre-trained with a large amount of movement trajectory data to extract a multidimensional vector containing the meaning of the user's movement trajectory, and this multidimensional vector is used to predict the user's future location. Here, the above pre-training uses a task of estimating masked positions and times in the time series of the user's positions, instead of filling in the blanks of words in natural language processing.

By the way, the above BERT is explained in detail in Non-Patent Document 3.

Ryo Kanamori, Takayuki Morikawa, Toshiyuki Yamamoto, and Tomio Miwa, "Development of an integrated transportation demand forecasting model for formulating a comprehensive transportation strategy," Journal of the Japan Society of Civil Engineers, Vol. 65, No. 4, pp. 503-518, 2009. Lin, Y., Wan, H., Guo, S., & Lin, Y. “Pre-training Context and Time Aware Location Embeddings from Spatial-Temporal Trajectories for User Next Location Prediction”, Proceedings of the AAAI Conference on Artificial Intelligence , 35(5), pp. 4241-4248. 2021, <https://ojs.aaai.org/index.php/AAAI/article/view/16548> Jacob Devlin, Ming-Wei Chang, Kenton Lee, Kristina Toutanova, “BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding”, NAACL-HLT (1) 2019: 4171-4186, 2019, <https://aclanthology .org/N19-1423.pdf＞

However, conventional location prediction technology that does not rely on the results of questionnaire surveys, such as that disclosed in Non-Patent Document 2, is difficult to apply to traffic demand prediction for traffic policies, for example.

Specifically, the location prediction model that applies BERT in Non-Patent Document 2 is constructed by learning only the relationship between locations in a time series, for example, that if you have traveled to location A and then B, it is likely that you will travel to location C at the next point in time. In other words, the basic premise of location prediction is that the conditions within the prediction target area, such as the distribution of transportation means and the locations of various facilities, remain the same as they are now. Therefore, it is impossible to predict the locations of traffic congestion when the conditions within the area, which are important in traffic demand prediction, change, for example, the location of traffic congestion that would occur if a hospital at location C was relocated to location D.

In addition, the location prediction model that applies BERT in Non-Patent Document 2 only uses time-series information on the user's location as input, so it is possible to predict the destination to which the user will head in the future, but it is not possible to predict the route the user will take to reach the destination. As a result, it is impossible to predict information about individual routes, such as the usage volume of each road and each railway line, which is important in traffic demand forecasting.

The present invention aims to provide a movement/stay estimation model and method that can estimate information related to the movement or stay of an estimation target, which also enables route prediction, while taking into account conditions within the area. It also aims to provide a movement trajectory prediction device and program that can predict the movement trajectory of the prediction target using the estimation results.

According to the present invention, there is provided a movement/stay estimation model for causing a computer to function to estimate information related to a subsequent movement or stay of an object in an area, the movement/stay estimation model comprising:
an area feature position encoding unit which receives as input area feature information generated from information relating to things or events related to or that may affect the movement or stay of the subject in the area, and information relating to a time series of the movement or stay of the subject, and generates area feature embedding information in which the area feature information is embedded, and position embedding information in which information relating to a position in the time series is embedded, or generates area feature position embedding information in which information combining the area feature information and the information relating to a position in the time series is embedded;
an area movement stay encoding unit which receives as input the area feature embedding information and the position embedding information, or receives as input information combining the area feature embedding information and the position embedding information, or receives as input the area feature position embedding information, and generates area movement stay embedding information in which information related to the movement or stay of the target in the area is embedded;
A movement and stay estimation model is provided that causes a computer to function as a movement and stay estimation unit that receives the area movement and stay embedded information as an input and generates information related to the subsequent movement or stay.

Here, it is also preferable that the above-mentioned area characteristic information according to the present invention is generated from area things and events information including at least one of information related to the number, size or content of transportation or means of travel in each of the multiple sections constituting the area, information related to the number, size or content of public facilities or public events, information related to the number, size or content of commercial/service providing facilities or commercial/service providing events, information related to the number, size or content of residential facilities, and information related to the number, size or content of business facilities.

In addition, as one embodiment of the movement and stay estimation model according to the present invention, the movement and stay estimation model causes a computer to function as a movement and stay time point encoding unit that receives information related to a time series of the movement or stay of the target and generates time-embedded information in which information related to a time point in the time series is embedded;
It is also preferable that the area movement stay encoding unit inputs the area feature embedding information, the location embedding information, and the time point embedding information, or inputs a combination of the area feature embedding information, the location embedding information, and the time point embedding information, or inputs the area feature location embedding information and the time point embedding information, or inputs a combination of the area feature location embedding information and the time point embedding information.

Furthermore, in the movement and stay estimation model according to the present invention, the information on the time series of the movement or stay of the target includes information on the location and time of the movement or stay in a certain period of time, and information on the location and time of the movement or stay in an estimation period following the certain period of time,
It is also preferable that the movement/stay estimation unit generates, as the information related to the subsequent movement or stay, information related to the likelihood of information related to the position and time of the movement or stay during the estimation period.

Furthermore, as another embodiment of the movement and stay estimation model according to the present invention, the area feature location encoding unit has an area feature encoding unit that receives the area feature information as an input and generates area feature embedding information in which the area feature information is embedded, and a movement and stay location encoding unit that receives information related to a time series of the movement or stay as an input and generates the location embedding information,
It is also preferable that the area change stay encoding unit receives as input the area feature embedding information and the location embedding information, or receives as input a combination of the area feature embedding information and the location embedding information.

In an embodiment including the area feature encoding unit, the area feature information is area graph information in which information relating to things or events that may affect or are related to the movement or stay of the subject in a target area among the multiple areas constituting the area is used as a node, and a matrix expressing a setting status of links between the target area and other surrounding areas is used as a graph adjacency matrix,
It is also preferable that the area feature encoding unit includes a graph convolution network (GCN) that receives the area graph information as an input.

It is also preferable that the graph adjacency matrix has matrix elements that reflect the presence or absence of a specified means of traffic, movement, distribution, or communication between the target area and other surrounding areas, or the traffic volume, movement volume, distribution volume, or communication volume per unit time.

Furthermore, as yet another embodiment of the movement and stay estimation model according to the present invention, it is also preferable that the area characteristic position encoding unit generates, from the received area characteristic information, a category value that represents a preset category and expresses the characteristics of a target area among multiple areas that make up the area, and generates area characteristic position embedding information in which information combining the category value and information related to the position in the time series is embedded.

Furthermore, it is also preferable that the movement and stay estimation model according to the present invention is a model trained by a training process in which training data includes information relating to a time series of the movement or stay of the subject having a masked location and correct area characteristic information, and the training process is provided for training purposes together with the movement and stay estimation unit downstream of the area movement and stay encoding unit, and the loss is the sum of the error from the correct answer in the output from a masked location estimation unit for solving the problem of guessing data for a masked location and the error from the correct answer in the output from the movement and stay estimation unit.

According to the present invention, there is also provided an area feature generating means for generating the area feature information from information related to things or events that may affect or are related to the movement or stay of the subject;
a movement/stay time series information generating means for generating information related to a time series of the movement or stay of the target from the acquired terminal location information;
The above-mentioned travel and stay estimation model and
A movement and stay prediction means for determining information related to the subsequent movement or stay using the movement and stay estimation model ;
and a movement trajectory prediction means for predicting information relating to the movement trajectory of the object based on the information relating to the subsequent movement or stay determined sequentially.

According to the present invention, there is further provided an area feature generating means for generating the area feature information from information related to things or events that may affect or are related to the movement or stay of the subject;
a movement/stay time series information generating means for generating information related to a time series of the movement or stay of the target from the acquired terminal location information;
a movement and stay prediction means for providing the area feature information and information related to the time series of the movement or stay of the target as input information to the movement and stay estimation model described above, and determining the information generated by the movement and stay estimation unit as information related to the subsequent movement or stay;
a movement trajectory prediction means for predicting information related to the movement trajectory of the target based on the information related to the subsequent movement or stay determined in sequence;
There is provided a trajectory prediction device having the following .

According to the present invention, there is further provided a computer-implemented movement/stay estimation method for estimating information related to a subsequent movement or stay of an object in an area, the method comprising:
a step of taking as input area feature information generated from information relating to things or events that may affect or are related to the movement or stay of the subject in the area, and information relating to a time series of the movement or stay of the subject, and generating area feature embedded information in which the area feature information is embedded, and location embedded information in which information relating to the location in the time series is embedded, or generating area feature location embedded information in which information combining the area feature information and the information relating to the location in the time series is embedded;
a step of receiving the area feature embedding information and the location embedding information, or receiving information combining the area feature embedding information and the location embedding information, or receiving the area feature location embedding information, and generating area movement/stay embedded information in which information related to the movement or stay of the target in the area is embedded;
A movement and stay estimation method is provided, the movement and stay estimation method including a step of receiving the area movement and stay embedding information and generating information related to the subsequent movement or stay.

The movement/stay estimation model and method of the present invention can estimate information related to the movement or stay of an estimation target, which information also enables route prediction, while taking into account conditions within the area. Furthermore, the movement trajectory prediction device and program of the present invention can use the estimation results to predict the movement trajectory of the estimation target.

FIG. 1 is a schematic diagram showing an embodiment of a movement/stay estimation model according to the present invention and an embodiment of a movement trajectory prediction device according to the present invention. FIG. 2 is a schematic diagram for explaining one embodiment of movement stay time-series information (input to a movement stay position encoding unit) according to the present invention. 5 is a schematic diagram for explaining an area feature encoding process in the area feature encoding unit according to the present invention; FIG. FIG. 2 is a schematic diagram for explaining an embodiment of a training process in a movement and stay estimation model according to the present invention. FIG. 11 is a schematic diagram for explaining another embodiment of the movement and stay estimation model according to the present invention.

The following describes in detail an embodiment of the present invention with reference to the drawings.

[Movement and Stay Estimation Model]
FIG. 1 is a schematic diagram showing an embodiment of a movement/stay estimation model according to the present invention and an embodiment of a movement trajectory prediction device according to the present invention.

The movement and stay estimation model 1 of this embodiment shown in FIG. 1 is a model that estimates information (future movement and stay information) related to future movement or stay of an estimation target (in this embodiment, a user (a mobile terminal owned by the user)) in a survey target area (survey target area), that is, route information of one future trip in this embodiment. Specifically,
(A) an area feature position encoding unit 11 which receives as input (a) "area feature information" generated from "area thing and event information" relating to things or events that may affect or are related to the movement or stay of a user (or a mobile terminal carried by the user) in a survey target area, and (b) "movement and stay time series information" which is information relating to the time series of the movement or stay of an estimated target user, and generates (c) "area feature embedded information" emb_ar in which the "area feature information" is embedded, and (d) "position embedded information" emb_p in which information relating to the position in the time series of the movement or stay of the user is embedded;
(B) an area movement stay encoding unit 14 that receives the generated “area feature embedding information” emb_ar and “location embedding information” emb_p as input, or receives information combining these embedding information as input, and generates “area movement stay embedding information” z in which information related to the movement or stay of the estimated target user in the survey target area is embedded;
(C) This is a machine learning model that causes a computer to function as a movement and stay estimation unit 15 that uses the generated “area movement and stay embedded information” z as input and generates route information (later movement and stay information) for one future trip for the user to be estimated.

Here, the "area thing and event information" that is the basis for generating the "area characteristic information" in (A) above is, for example, the following in each of the multiple zones (meshes in this embodiment) that make up the survey target area:
(A1) Information on the number, scale, or content of transportation or means of transportation (for example, the number of train stations, the number of bus stops, the number of interchanges on expressways, whether express trains stop at existing train stations or not, the number of stops per day at existing bus stops, and the length of roads with a certain width or more, etc.);
(a2) Information concerning the number, scale, or content of public facilities or public events (e.g., the number of schools and the number of students, the number of hospitals and the number of beds, etc.);
(a3) Information on the number, size, or content of commercial or service-providing facilities or commercial or service-providing events (for example, the number of commercial facilities by size, etc.);
(A4) Information on the number, size or content of residential facilities (e.g., the number and number of apartment buildings, the number of general houses, etc.); and
(A5) Information on the number, size or content of business facilities (for example, the number of business establishments by size, etc.)
The information includes at least one of the above, and preferably information relating to each of (A1) to (A5).

Incidentally, such "area thing and event information" may be information that has been compiled in advance, for example, by collecting and aggregating relevant information for each region (mesh) from map data or timetable data that includes the survey target area. Furthermore, "area feature information" may be a feature quantity that reflects such "area thing and event information," for example, a multidimensional feature vector.

In this way, the movement/stay estimation model 1 is capable of estimating route information for one future trip (future movement/stay information) by taking into account the "area characteristic information" that is the in-area condition that may affect or is related to the user's movement/stay, as well as the "movement/stay time series information" of the estimation target user. In addition, as will be described in detail later, the movement trajectory prediction device 9 equipped with this movement/stay estimation model 1 also makes it possible to predict the route of the estimation target user.

As a result, the movement/stay estimation model 1 (the movement trajectory prediction device 9 equipped with it) can also be applied to traffic demand prediction for traffic policies, for example. For example, the movement/stay estimation model 1 takes into account not only "movement/stay time series information" including information on the relationship between movement/stay locations in a time series, but also "area characteristic information" which is the conditions within the area that may affect or are related to the user's movement/stay, so it is possible to predict the location of traffic congestion when the conditions within the area change, which is important in traffic demand prediction, for example, the location of traffic congestion that would occur if a hospital at point C was relocated to point D.

Furthermore, the movement and stay estimation model 1 (the movement trajectory prediction device 9 equipped with it) can predict the movement trajectory of the estimation target user. In other words, it is possible to predict not only the destination to which the estimation target user will head in the future, but also the route the estimation target user will take to reach the destination. As a result, it is also possible to predict information about individual routes, such as the usage volume of each road and each railway line, which is important in traffic demand prediction.

[Model configuration and movement/stay estimation method]
The configuration of the moving stay estimation model 1 of this embodiment will be described in more detail below. Also in Fig. 1, the moving stay estimation model 1 includes an area feature position encoding unit 11 including an area feature encoding unit 111 and a moving stay position encoding unit 112, a moving stay time encoding unit 12, an integration unit 13, an area moving stay encoding unit 14, and a moving stay estimation unit 15 as functional components (embodied by a program installed in a computer). In other words, it is a model that causes a computer to function as these functional components.

<Area feature position encoding section>
Each of the above-mentioned functional components will be specifically described below. Similarly, in FIG. 1, the area characteristic position encoding unit 11 of this embodiment includes an area characteristic encoding unit 111 and a moving stay position encoding unit 112.

(Moving stay position encoding part)
Of these, in this embodiment, the movement and stay position encoding unit 112 receives as input movement and stay time series information generated from the terminal position information of the estimated user, and generates position embedding information emb_p in which information related to the position in the time series of the movement or stay of the estimated user is embedded.

Here, the movement and stay position encoding unit 112 uses an embedding section of BERT (Bidirectional Encoder Representations from Transformers), which is a well-known natural language processing model, specifically, a one-layer fully connected one-hot encoder Ω, as in the position prediction technology disclosed in Non-Patent Document 2. In the embedding section of this BERT, a multidimensional vector in which the meaning of the input sentence (word string) is embedded (representing the meaning) is output, and in the movement and stay position encoding unit 112, movement and stay time series information, for example,
(u1, p1, t1), (u1, p2, t2), (u1, p3, t3), ・・・
When you enter
Ω(p1), Ω(p2), Ω(p3), ・・・
Thus, the position-embedded information emb_p in which the position part in the input movement/stay time-series information is embedded is output. Here, for example, (u1, p1, t1) is data indicating that user 1 (u1) is at position p1 (p1 is, for example, a corresponding mesh code as described later) at time t1.

Figure 2 is a schematic diagram for explaining one embodiment of the movement and stay time series information (input to the movement and stay position encoding unit 112) according to the present invention.

According to the route diagram shown at the top of Figure 2, in a given day, User 1 travels along a route such as Position A (home) → Position B → Position C (α Station) → Position D (β Station) → Position E (work) → Position F → Position G → .... These positions are the results of measuring the position of the mobile terminal carried by User 1 using a base station positioning method or GPS (Global Positioning System). This measured position may also be expressed discretely, and may be displayed, for example, using the mesh code of the mesh that includes the measured position, or the base station identifier (ID) of the base station that includes the mobile terminal under its control.

In any case, for user 1 who took the route shown in FIG. 2 on a certain day, the following terminal location information is obtained: A, A, A, B, C, D, E, E, E, F, G, H, I, I, I, J, ...
Here, each capital alphabet letter separated by a comma in order is a measurement position measured every time a predetermined time (for example, 15 minutes) has elapsed.

Next, a known movement/stay determination process is performed on such terminal location information to extract trips (movements), and the locations are expressed discretely (e.g., in mesh code display) to obtain time-series information on the movement/stay of user 1 on a certain day, e.g., A, A, A, [STAY_END], A, B, C, D, E, [TRIP_END], E, E,...
Here, [STAY_END] indicates that the stay ended at this position (time), and [TRIP_END] indicates that the trip (movement) ended at this position (time).

As the above movement/stay determination process, it is preferable to use the method disclosed in the non-patent document: Nao Kobayashi, Hiroki Ishizuka, Atsuyuki Minamigawa, Shigeki Muramatsu, Tomohiro Ono, "Proposal of a movement/stay state estimation method suitable for mobile phone communication history", Information Processing Society of Japan Transactions on Database, vol. 10, no. 1, pp. 13-23, 2017. Here, a trip (movement) may be extracted as a section between one stay and the next stay.

Below, we will explain the movement and stay time series information as input information when using this movement and stay estimation model 1 to estimate next trip information (subsequent movement and stay information) related to the next possible trip (movement) of the estimation target user. This explains what type of movement and stay time series information can be input when estimating the movement route (next trip information) that user 1 will take after commuting to work based on the movement route (actual trip information) up to the point where user 1 arrives at work.

Specifically, such movement/stay time series information is a set of trip information related to an actual trip (movement) of the estimation target user and a candidate (candidate next trip information) of next trip information which is information related to a trip (movement) that may occur next, for example,
(u1, p1, t1), (u1, p2, t2), ..., (u1, p5, t5), [TRIP_END]
[SEP]
(u1, p6, t6), (u1, p7, t7), ..., (u1, p9, t9), [TRIP_END]
Here, for example, (u1, p1, t1) is data indicating that user 1 (u1) is at position p1 (p1 is, for example, a corresponding mesh code) at time t1. Also, [SEP] (the second line above) is a symbol that separates the actual trip information (the first line above) from the candidate next trip information (the third line above).

The movement/stay estimation model 1 (movement/stay estimation unit 15) of this embodiment inputs movement/stay time series information including such candidate next trip information, and outputs the likelihood that the next trip to occur will be a trip related to this candidate next trip information. The one with the highest likelihood among the multiple input candidate next trip information is determined to be the information related to the estimated (likely to occur next) next trip (later movement/stay information).

Here, the candidate next trip information can be, for example, randomly selected from the user's trip information collected up to now, trip information whose starting point location matches the final location of the trip information before [SEP] (for example, the location at the estimated time) or is separated within a specified range. It is also possible to generate trip information completely randomly to use as the candidate next trip information, but it is also preferable to generate trip information whose trajectory is included within a specified travel distance range from the final location of the trip information before [SEP] (for example, the location at the estimated time) to use as the candidate next trip information.

In addition, it is also preferable to perform at least one of the filtering processes described below on the above-described movement/stay time-series information in order to remove noise, particularly during training of the movement/stay estimation model 1.
(a) If the total time interval between one "stay" and the "stays" occurring consecutively thereto is shorter than a predetermined time, these "stays" are integrated into one "stay."
(b) For positions that are discretely represented (e.g., displayed as a mesh code) and have consecutive identical values, delete the duplicates (e.g., if the values are A, A, A, B, B, B, C, C, C, then change them to A, B, C).
(c) If the result of the deletion in (b) above is that the "movement" remains in the same position (for example, in the same mesh), this "movement" is deleted as noise.

Although one embodiment of the movement/stay time series information input to the movement/stay position encoding unit 112 has been described above, the movement/stay time series information is not limited to this. For example, a time series of "stay" and/or "movement" for a predetermined period, for example,
..., [TRIP_END], ..., [STAY_END], ...[TRIP_END]
[SEP]
・・・, [STAY_END], ・・・[TRIP_END]
It is also possible to use the above as the movement/stay time series information. Here, the candidate information after [SEP] can be set in various forms, for example, only "stay".

(Area feature encoding part)
Returning to FIG. 1, the area feature encoding unit 111
(a) inputting area feature information generated from area object event information in a mesh (target mesh) that includes a time-series position of the movement or stay of the estimated target user among a plurality of pre-defined meshes (districts) that constitute a survey target area;
(b) Generate area feature embedding information emb_ar in which the area feature information of this target mesh is embedded.

Figure 3 is a schematic diagram for explaining the area feature encoding process in the area feature encoding unit 111.

According to FIG. 3(A), first, from map data, timetable data, etc., area object and event information in each mesh that constitutes the survey target area is obtained, such as the above-mentioned (a1) number of transportation or means of transportation, ..., (a2) number of public facilities or public events, ..., (a3) number of commercial/service providing facilities or commercial/service providing events, ..., (a4) number of residential facilities, ..., (a5) number of business facilities, ...
has been compiled.

The area feature information of each mesh is generated from such area object event information. In this embodiment, the area feature information of each mesh is, for example,
([Number of train stations], [Number of trains stopping at each station per day], [Number of schools], [Number of large retail stores], ...)
Here, it is preferable that the numerical values in [.] are normalized to a range of 0 to 1. In any case, in this embodiment, such area feature information for each mesh is stored and managed in the area feature database (DB) 111a.

Here, in this embodiment, the area feature encoding unit 111:
(a) Identifying a mesh code including a travel route (related to actual trip information) from the acquired travel and stay time series information;
(b) Specifying and inputting each identified mesh code to the area characteristics DB 111a to obtain area characteristics information of each target mesh (related to actual trip information);
(c) The area feature information of each obtained target mesh is input to a multi-layer perceptron (MLP, Multi-Layer Perceptron) for compression, and area feature embedding information emb_ar is output.

As another embodiment of the area feature encoding process, as shown in FIG. 3B, the area feature encoding unit 111 is provided with a known graph convolution network (GCN) that receives graph information as input, extracts and outputs the feature amount of the graph,
(a) The area graph information is defined as area feature information, in which the area object event information in the target mesh (area) is defined as a "node" and the matrix expressing the setting status of the link between the target mesh and other surrounding meshes (areas) is defined as a "graph adjacency matrix."
(b) It is also preferable to input area graph information (area feature information) to the above GCN and output area feature embedding information emb_ar as a feature quantity expressing the input graph.

The GCN used here can be the model disclosed in, for example, the non-patent document: Thomas N. Kipf, Max Welling, "Semi-Supervised Classification with Graph Convolutional Networks", Published as a conference paper at ICLR (Poster) 2017, <https://arxiv.org/pdf/1609.02907.pdf>.

In this embodiment, the area graph information (area feature information) for each mesh may also be stored and managed in the area feature DB 111a. The area feature encoding unit 111 specifies and inputs the mesh code of each target mesh (including the actual movement route) identified based on the acquired movement and stay time series information to this area feature DB 111a, acquires the area graph information of each target mesh (related to the actual trip information), inputs the acquired area graph information of each target mesh to the GCN for feature quantification processing, and outputs the area feature embedding information emb_ar.

In addition, the graph adjacency matrix of a target mesh included in the area graph information (area feature information) can be a matrix in which a link is established between the target mesh and a geographically adjacent mesh, and the matrix elements corresponding to the adjacent mesh to which the link is established are set to a predetermined value, for example 1.

As a modification, it is also preferable that this graph adjacency matrix has matrix elements that reflect the presence or absence of a specified means of transportation, movement, distribution, or communication between a given target mesh and other surrounding meshes (adjacent or within a specified range), or the traffic volume, movement volume, distribution volume, or communication volume per unit time.

For example, in a graph adjacency matrix, a matrix element corresponding to another mesh that includes another railway station on a railway line that includes a railway station in the target mesh may be set to a predetermined value, for example, 1. Alternatively, the number (standard value) of "other railway stations on the railway line" included in this other mesh may be set as a matrix element. Furthermore, a matrix element corresponding to another mesh that includes another railway station on a railway line that includes a railway station in the target mesh may be the number (standard value) of trains running between these two stations per unit time (for example, 1 hour or 24 hours).

In addition, in the graph adjacency matrix, for example, the traffic volume (e.g., the number of cars passing) per unit time (e.g., 1 hour or 24 hours) on the road between the target mesh and its adjacent mesh (its standard value) may be set as the matrix element corresponding to this adjacent mesh. In any case, by inputting the graph adjacency matrix as described above into the GCN, the area feature encoding unit 111 can generate area feature embedded information emb_ar in which the geographical relationships between the meshes and the relationships related to their mutual entry and exit and exchange are also embedded.

Furthermore, when using a graph adjacency matrix as described above, the area graph information may contain not only the node information (area thing event information) of the target mesh, but also node information (area thing event information) of other surrounding meshes in which non-zero finite values are set as matrix elements of the graph connectivity matrix. This allows more information that is useful for estimating the movement and stay of the estimation target to be reflected in the area feature embedding information emb_ar.

<Movement stay time encoding section>
Returning to FIG. 1, the movement stay time encoding unit 12 receives the above movement stay time series information generated from the terminal location information as input, and generates time embedded information emb_t in which information relating to a time in the time series of this information is embedded.

In this embodiment, the movement stay time encoding unit 12 uses a temporal embedding module Φ of the location prediction technology disclosed in Non-Patent Document 2. This temporal embedding module Φ uses, for example, the following as the movement stay time series information:
(u1, p1, t1), (u1, p2, t2), (u1, p3, t3), ・・・
When is entered,
Φ(t1), Φ(t2), Φ(t3), ・・・
The time-embedded information emb_p in which the time-point portion of the input movement/stay time series information is embedded is output. Here, for example, Φ(t1) is expressed by the following formula (1) Φ(t1)
= [cos(w1×t1), sin(w1×t1), ..., cos(wd×t1), sin(wd×t1)]
In the above formula (1), w1, ..., wd are learning parameters, and d is half the number of vector dimensions of the generated multidimensional feature vector (here, the time point embedding information emb_t).

By using such a temporal embedding module Φ to generate time embedding information emb_t and to perform the movement/stay estimation process, it is possible to obtain a more accurate movement/stay estimation result that takes into account information that is very important for movement/stay estimation, such as the duration of stay along the movement route and the speed of movement. However, this movement/stay time encoding unit 12 can be omitted, although this advantage cannot be enjoyed.

<Integration Department>
In this embodiment, the integration unit 13 shown in FIG.
(a) area feature embedding information emb_ar output from the area feature encoding unit 111;
(b) location embedding information emb_p output from the movement stay location encoding unit 112;
(c) Generate a result of taking a (vector) sum or a (vector) concatenation of the time embedding information emb_t output from the movement stay time encoding unit 12, or compressing the result of taking a (vector) sum or the result of concatenation using a multilayer perceptron (MLP), and output this embedded information as integrated information to the area movement stay encoding unit 14. Alternatively, the integration unit 13 can output the embedded information (a) to (c) above as is to the area movement stay encoding unit 14.

<Area Movement and Stay Encoding Section>
Also in Figure 1, the area movement stay encoding unit 14 uses the output from the above-mentioned integration unit 13 as input and generates area movement stay embedded information z in which information related to the movement or stay of the estimated target user in the survey target area is embedded.

Specifically, in this embodiment, the area movement stay encoding unit 14 uses transformer encoder layers of the position prediction technology disclosed in Non-Patent Document 2. These transformer encoder layers are applications of the BERT transformer encoder, and each of the N layers in the transformer encoder layers is a layer composed of a multi-head self-attention module and a position-wise fully connected feed-forward network.

For example, if the input output from the integration unit 13 is z' = ( _z'1 , _z'2 , z'3, ...), the output of the kth layer is expressed by the following equation (2), where the processing in the layer is Trans(·): (h(k) ₁ , h(k) ₂ , h(k) ₃ , ...) = Trans(h(k) ₁ , h(k) ₂ , h(k) _{3 ,} ...).
Here (h(0) ₁ , h(0) ₂ , h(0) ₃ ,...)=(z' ₁ , z' ₂ , z' ₃ ,...)
The area change stay embedding information z, which is the final output of the transformer encoder layers (area change stay encoding unit 14), is expressed by the following formula (3): z = ( _z1 , _z2 , _z3 , ...) = (h(N) ₁ , h(N) ₂ , h(N) ₃ , ...)
It is calculated as follows.

Such area movement and stay embedded information z is a representation vector similar to a word distributed representation vector in which the meaning of a sentence, which is a word string in BERT, is embedded, and is a multidimensional representation vector in which information on the activity intention and activity content (which also depends on the conditions within the area) of the movement and stay of the estimated target user in the survey area is embedded.

<Movement and Stay Estimation Unit>
Also in Figure 1, the movement/stay estimation unit 15 receives the area movement/stay embedding information z output from the area movement/stay encoding unit 14 as input, and generates and outputs information related to the future movement or stay of the estimated target user, in this embodiment, route information for one trip at the next time point (future).

Specifically, in this embodiment, the movement/stay estimation unit 15 outputs the likelihood that the candidate next trip information included in the movement/stay time series information input to the movement/stay estimation model 1 is information related to the actual next trip. That is, the movement/stay time series information input to the model 1 is
The information includes (a) trip information relating to an actual trip (movement) of the user to be estimated during a certain period, and (b) candidate next trip information which is information relating to a trip (movement) that may occur during an estimated period following the certain period. The movement/stay estimation unit 15 generates and outputs information relating to the likelihood of the candidate next trip information (b) above based on the area movement/stay embedding information z acquired from the area movement/stay encoding unit 14.

Here, the movement and stay estimation unit 15 is configured to apply Next Sentence Prediction (NSP) in BERT, which determines whether two passed sentences are consecutive or not, and can have a configuration of, for example, a fully-connected feed-forward neural network (FFN).

As described above, the movement and stay estimation unit 15 can generate information (later movement and stay information) related to the likelihood of candidate next trip information based on the area movement and stay embedded information z in which information on the activity intention and activity content (which also depends on the conditions within the area) of the estimation target user during movement and stay is embedded. In other words, it is possible to obtain a more accurate movement and stay estimation result that also takes into account the conditions within the area.

[Training movement and residence estimation model 1]

Figure 4 is a schematic diagram for explaining one embodiment of the training process in the movement and stay estimation model 1.

As shown in FIG. 4, in the training process of the movement and stay estimation model 1 of this embodiment, specifically in pre-training for setting the initial values of the learning parameters, a masked part estimation unit 16 is provided in addition to the movement and stay estimation unit 15 after the area movement and stay encoding unit 14. This masked part estimation unit 16 is configured to apply MLM (Masked Language Model) for solving a problem in BERT in which data in a masked part is assigned (word fill-in-the-blank problem), and can have an FFN configuration like the above-mentioned NSP (movement and stay estimation unit 15).

Here, the training (learning) data is, for example,
(u1, p1, t1), (u1, p2, t2), ..., (u1, p5, t5), [TRIP_END]
[SEP]
(u1, p6, t6), (u1, [MASK], [MASK]), ・・, (u1, p9, t9), [TRIP_END]
For example, if the actual trip information in the first row is followed by the next trip information in the third row (with some masked areas), the training data above will be the correct answer data.

In addition, data is often missing or noise is superimposed on device location information from smartphones, etc., and trip information that is not actually consecutive may be selected as the correct next trip information, creating training data that becomes noise during pre-training. Therefore, if location data cannot be collected continuously for the previous "stay," for example because the measurement time span of positioning exceeds a specified threshold, it is preferable to exclude the corresponding trip information from the training data.

In the pre-training of this embodiment, such a large amount of training data and (correct) area feature information are input to the movement and stay estimation model 1 of the training specification shown in FIG.
(a) The error from the correct answer (1 or 0) in the output of the movement/stay estimation unit 15 (NSP),
(b) The amount obtained by combining (for example, weighted addition of) the error from the correct value in the output of the mask location estimation unit 16 (MLM) is regarded as a loss, and the initial values of the learning parameters in the model 1 are determined according to the backpropagation method. By carrying out such pre-training, the area movement and stay encoding unit 14 (transformer encoder) can output area movement and stay embedding information z that more appropriately reflects information on the activity intention and activity content that also depend on the in-area conditions during movement and stay.

[Another embodiment of the movement and stay estimation model]
FIG. 5 is a schematic diagram for explaining another embodiment of the movement and stay estimation model according to the present invention.

The movement/stay estimation model 2 of the embodiment shown in FIG. 5 is the same as the movement/stay estimation model 1 shown in FIG. 1, and is a model that estimates and outputs the likelihood of subsequent movement/stay information, for example, candidate next trip information, from area feature information in the survey target area and the movement/stay time series information of the estimation target user. In addition, the movement/stay time encoding unit 22, area movement/stay encoding unit 24, and movement/stay estimation unit 25 of this movement/stay estimation model 2 each have the same functional configuration as the functional configuration units of the same name in the movement/stay estimation model 1 (FIG. 1).

In addition, training (pre-training) of the moving and staying estimation model 2 can be performed in the same way as the moving and staying estimation model 1 (Figure 4). Furthermore, the moving and staying time encoding unit 22 of the moving and staying estimation model 2 can be omitted, like the moving and staying time encoding unit 12 of the moving and staying estimation model 1.

However, the area feature position encoding unit 21 of the movement stay estimation model 2 differs from the area feature position encoding unit 11 of the movement stay estimation model 1 in that the area feature processing unit (category value generation unit 211) is provided before the movement stay position encoding unit 213. As a result, instead of outputting the area feature embedding information emb_ar and the position embedding information emb_p, it outputs "area feature position embedding information" emb_ar_p in which information combining the area feature information and the position information in the movement stay time series information is embedded.

In response to this, the integration unit 23 (a) generates information that combines the area characteristic position embedding information emb_ar_p and the time point embedding information emb_t and outputs the combined information, or (b) each of these pieces of embedded information, to the area movement and stay encoding unit 24.

Specifically, in the movement and stay estimation model 2 of this embodiment, the area characteristic position encoding unit 21
(A) a category value generation unit 211 that generates a category value representing a preset category and expressing the characteristics of a target mesh (including the time-series position of the movement or stay of the estimation target user) among multiple meshes that constitute the survey target area based on the received area characteristic information;
(a) a mesh code unit 212 for generating time-series information of a position (mesh code) from the received movement/stay time-series information;
(c) A movement stay position encoding unit 213 that receives the above-mentioned (a) category value and the above-mentioned (b) mesh code time-series information, and generates and outputs area characteristic position embedding information emb_ar_p.

In this embodiment, the category value generation unit 211 (a) above first receives area feature information (related to the movement or stay of the estimated target user), such as the duration of the movement or stay of the estimated target user, the movement speed, the movement direction, the date and time, the home location, the workplace location, etc.

Next, the category value generation unit 211 uses the received area feature information to calculate, for each target mesh (including the time-series positions of the movement or stay of the estimation target user),
(a) Stay time category value (for example, if the stay time in the target mesh is less than 15 minutes, it is "A", if it is between 15 minutes and 60 minutes, it is "B", and if it is 60 minutes or more, it is "C");
(b) Speed category value (for example, if the average travel speed in the target mesh is less than 5 km/h, it is "X", if it is between 5 km/h and 20 km/h, it is "Y", and if it is 20 km/h or more, it is "Z");
(c) Movement direction category value (for example, if the direction of the movement route going out from the target mesh is north (0 degrees or more and less than 45 degrees and 315 degrees or more and less than 360 degrees in the left-handed north reference azimuth), it is "K", if it is east (45 degrees or more and less than 135 degrees), it is "E", if it is south (135 degrees or more and less than 225 degrees), it is "S", if it is west (225 degrees or more and less than 315 degrees), it is "T", if there is no movement route going out, it is "O");
(d) a home location category value (e.g., the mesh code of the mesh that contains the home location);
(e) a work location category value (e.g., the mesh code of the mesh that contains the work location);
(f) Weekday/holiday category value (for example, "W" if the day related to the time series in the target mesh is a weekday, and "H" if it is a holiday), and (g) time period category value (for example, "D" if the time period related to the time series in the target mesh is daytime (from 6:00 to 19:00), and "N" if it is nighttime (from 20:00 to 5:00 the next day)).
At least one, for example all, of

Next, the movement/stay position encoding unit 213 (c) above receives the generated category value and the mesh code time series information (b) above, creates "mesh code time series information with additional information added" from this information, and generates area characteristic position embedding information emb_ar_p from this mesh code time series information.

Here, "mesh code time series information with additional information" refers to <·| as mesh code time series information and |·> as additional information, and for a certain target mesh, for example,
<5000000003|-|AYK>
<5000000003|-|5000000123>-|5000000987>
<5000000003|-|WD>
For example, |AYK> in the first row is additional information indicating that the estimated target user's stay time in this target mesh is less than 15 minutes, the average moving speed is between 5km/h and 20km/h, and the moving direction is north.

In addition, |5000000123> in the second row indicates that the home location is within the mesh with mesh code 5000000123, and |5000000987> indicates that the workplace location is within the mesh with mesh code 5000000987. Furthermore, |WD> in the third row indicates that the day related to the time series in the target mesh is a weekday and that the time period is daytime.

Here, the additional information in the first row (|AYK> in the above) is, for example, a behavior feature B related to a total of 45 (=3×3×5) types of category value sequences, which are a combination of three types, "A", "B", and "C", three types, "X", "Y", and "Z", and five types, "K", "E", "S", "T", and "O". The additional information in the second row (|5000000123>-|5000000987> in the above) is an individual feature P related to two category value sequences, which are a combination of n types of category values and m types of category values, where the number of meshes that can be determined as home and work is n and m, respectively. The additional information in the third row (|WD> in the above) is, for example, a time feature T related to a total of 4 (=2×2) types of category value sequences, which are a combination of two types, "W" and "H", and two types, "D" and "N".

The functional configuration of the area characteristic position encoding unit 21 has been described above. In this embodiment, the category values (which are the source of additional information) generated by the category value generating unit 211 correspond to subwords in BERT. Similarly, the mesh code time series information generated by the mesh code unit 212 corresponds to words (strings of words) in BERT.

In BERT, the WordPiece algorithm is used to split a word (e.g., "doing") into a word (e.g., "do") and a subword (e.g., "##ing", where ## is a symbol that represents a subword), making it easier to understand the meaning of the word and preventing an increase in the number of dimensions of the word representation vector, thereby realizing efficient training and reducing computational costs.

Moreover, the "mesh code time series information with additional information" <·|-|·> in the first and third rows above corresponds to word information in BERT: Word-Subword. The "mesh code time series information with additional information" <·|-|·>-|·> in the second row above corresponds to word information with two subwords: Word-Subword-Subword. Furthermore, the area feature position encoding unit 21 applies the BERT transformer encoder to generate area feature position embedding information emb_ar_p from the "mesh code time series information with additional information" described above.

In this way, the area characteristic position encoding unit 21 of this embodiment can fully utilize the functions of BERT to generate area characteristic position embedding information emb_ar_p that can accurately estimate future movement and stay information (e.g., the likelihood of candidate next trip information).

[Trajectory prediction device, trajectory prediction program and method]
1, a description will be given of a movement trajectory prediction device 9 that is equipped with the movement stay estimation model 1 and predicts the future movement trajectory of a prediction target user. Incidentally, the following description will be given assuming that the device 9 is equipped with the movement stay estimation model 1 shown in FIG. 1, but it is of course possible for the device 9 to be equipped with the movement stay estimation model 2 shown in FIG.

The movement trajectory prediction device 9 of this embodiment shown in the lower part of FIG.
(A) an area feature generating unit 94 that generates area feature information from area object event information acquired from an external device, for example, a map data management server, via an input unit 91 and stored and managed in an area object information storage unit 92;
(B) a movement/stay time-series information generating unit 95 that generates a plurality of pieces of movement/stay time-series information (including mutually different pieces of candidate next trip information) for the prediction target user from terminal location information acquired via an input unit 91 from an external device, for example, a mobile data management server, and stored and managed in terminal location information 93;
(C) a movement stay prediction unit 96 that inputs these multiple pieces of movement stay time series information and area feature information into a movement stay estimation model 1, and determines the candidate next trip information with the highest likelihood (included in the movement stay time series information) output from this model 1 as predicted next trip information (later movement stay information), and that sequentially determines next trip information (later movement stay information) that is continuous in time series by repeatedly performing a process of predicting the above-mentioned next trip information from actual trip information including an initial position of the user to be predicted, and then predicting the next next trip information using this next trip information as the actual trip information;
(D) a movement trajectory prediction unit 97 that predicts movement trajectory information (for example, for a certain day) of the prediction target user based on the next trip information (subsequent movement and stay information) that is sequentially determined;
(E) An output unit 98 having a communication function and transmitting the predicted movement trajectory information to an external information processing device, or having a display and displaying the predicted movement trajectory information.

In this way, the movement trajectory prediction device 9 can predict the movement trajectory of the prediction target user, taking into account not only movement and stay time series information but also area characteristic information (conditions within the area). This makes it possible to predict, for example, the movement trajectories of a single day, including the routes along the way, of multiple users in a certain urban area, taking into account information related to the behavior and activity of the users, and to predict the traffic behavior of the entire urban area, similar to the ABM (activity model).

In other words, the movement trajectory prediction device 9 can perform traffic demand prediction for traffic measures equivalent to ABM without relying on questionnaire surveys. Furthermore, it is also possible to predict the location of traffic congestion when the conditions within the area change, which is important in traffic demand prediction, for example, the location of traffic congestion that would occur if a hospital located at point C was relocated to point D.

The area feature generation unit 94, the movement/stay time series information generation unit 95, the movement/stay prediction unit 96, and the movement trajectory prediction unit 97 in (A) to (D) above are main functional components that implement an embodiment of the movement trajectory prediction method according to the present invention, and can also be considered as functions of a processor/memory that stores an embodiment of the movement trajectory prediction program according to the present invention. In addition, for this reason, the movement trajectory prediction device 9 may be a dedicated device for movement trajectory prediction, but it can also be, for example, a cloud server, a non-cloud server device, a personal computer (PC), a notebook or tablet computer, or a smartphone, etc., equipped with the movement trajectory prediction program according to the present invention.

As described above in detail, according to the present invention, it is possible to estimate subsequent movement and stay information by taking into account area characteristic information, which is an area condition that may affect or be related to the movement and stay of the subject, as well as the movement and stay time series information of the subject to be estimated. This also makes it possible to clarify the route of the subject to be estimated and predict its movement trajectory.

For example, the movement trajectory prediction method according to the present invention can be applied to perform accurate and detailed traffic behavior prediction in urban areas, develop resilient traffic and transportation infrastructure, and implement efficient urban planning to solve the challenges brought about by urbanization. In other words, the present invention can contribute to Goal 9 "Develop resilient infrastructure, promote sustainable industrialization and foster innovation" and Goal 11 "Make cities inclusive, safe, resilient and sustainable" of the United Nations-led Sustainable Development Goals (SDGs).

A person skilled in the art can easily make various changes, modifications, and omissions to the various embodiments of the present invention described above within the scope of the technical ideas and viewpoints of the present invention. The above description is merely an example of an embodiment and does not impose any unnecessary restrictions. The present invention is limited only by the scope of the claims and their equivalents.

1, 2 Movement and stay estimation model 11, 21 Area feature position encoding unit 111 Area feature encoding unit 111a Area feature database (DB)
112, 213 Movement stay position encoding unit 12, 22 Movement stay time encoding unit 13, 23 Integration unit 14, 24 Area movement stay encoding unit 15, 25 Movement stay estimation unit 16 Mask location estimation unit 211 Category value generation unit 212 Mesh code unit 9 Movement trajectory prediction device 91 Input unit 92 Area thing and event information storage unit 93 Terminal position information storage unit 94 Area feature generation unit 95 Movement stay time series information generation unit 96 Movement stay prediction unit 97 Movement trajectory prediction unit 98 Output unit

Claims (12)

A movement and stay estimation model that causes a computer to estimate information related to a future movement or stay of an object in an area, comprising:
an area feature position encoding unit which receives as input area feature information generated from information relating to things or events related to or that may affect the movement or stay of the subject in the area, and information relating to a time series of the movement or stay of the subject, and generates area feature embedding information in which the area feature information is embedded, and position embedding information in which information relating to a position in the time series is embedded, or generates area feature position embedding information in which information combining the area feature information and the information relating to a position in the time series is embedded;
an area movement stay encoding unit which receives as input the area feature embedding information and the position embedding information, or receives as input information combining the area feature embedding information and the position embedding information, or receives as input the area feature position embedding information, and generates area movement stay embedding information in which information related to the movement or stay of the target in the area is embedded;
A movement and stay estimation model characterized by making a computer function as a movement and stay estimation unit that receives the area movement and stay embedding information as an input and generates information related to the subsequent movement or stay. The movement and stay estimation model according to claim 1, characterized in that the area characteristic information is generated from area thing and event information including at least one of information related to the number, size or content of transportation or means of travel in each of the multiple sections that make up the area, information related to the number, size or content of public facilities or public events, information related to the number, size or content of commercial/service facilities or commercial/service events, information related to the number, size or content of residential facilities, and information related to the number, size or content of business facilities. making the computer function as a movement/stay time point encoding unit that receives information related to a time series of the movement or stay of the object as input and generates time-embedded information in which information related to a time point in the time series is embedded;
The movement stay estimation model according to claim 1 or 2, characterized in that the area movement stay encoding unit inputs the area feature embedding information, the position embedding information, and the time embedding information, or inputs information combining the area feature embedding information, the position embedding information, and the time embedding information, or inputs the area feature position embedding information and the time embedding information, or inputs information combining the area feature position embedding information and the time embedding information. The information on the time series of the movement or stay of the object includes information on the location and time of the movement or stay during a certain period of time and information on the location and time of the movement or stay during an estimated period after the certain period of time;
The movement and stay estimation model according to any one of claims 1 to 3, characterized in that the movement and stay estimation unit generates, as the information related to the subsequent movement or stay, information related to the likelihood of information related to the location and time of the movement or stay during the estimation period. the area feature position encoding unit has an area feature encoding unit that receives area feature information as an input and generates area feature embedded information in which the area feature information is embedded, and a movement/stay position encoding unit that receives information related to a time series of the movement or stay and generates the position embedded information,
The movement stay estimation model according to any one of claims 1 to 4, characterized in that the area movement stay encoding unit receives the area feature embedding information and the location embedding information as input, or receives information that combines the area feature embedding information and the location embedding information as input. The area characteristic information is area graph information in which information on things or events that may affect or are related to the movement or stay of the subject in a target area among multiple areas constituting the area is used as a node, and a matrix expressing the setting status of links between the target area and other surrounding areas is used as a graph adjacency matrix,
The movement and stay estimation model according to claim 5 , wherein the area feature encoding unit includes a graph convolution network (GCN) that receives the area graph information as an input. The movement and stay estimation model according to claim 6, characterized in that the graph adjacency matrix has matrix elements that reflect the presence or absence of a specified means of traffic, movement, distribution, or communication between the target area and other surrounding areas, or the traffic volume, movement volume, distribution volume, or communication volume per unit time. The movement and stay estimation model according to any one of claims 1 to 4, characterized in that the area characteristic position encoding unit generates, from the received area characteristic information, a category value that represents a preset category and expresses the characteristics of a target area among multiple areas that make up the area, and generates area characteristic position embedding information in which information combining the category value and information related to the position in the time series is embedded. The moving stay estimation model according to any one of claims 1 to 8, characterized in that the moving stay estimation model is a model trained by a training process in which training data includes information related to a time series of movement or stay of the target having a masked area and correct area feature information, and the training process assumes a loss equal to the sum of an error from a correct answer in an output from a masked area estimation unit for solving a problem of guessing data for a masked area, which is provided for training together with the moving stay estimation unit and is provided downstream of the area moving stay encoding unit, and an error from a correct answer in an output from the moving stay estimation unit. An area feature generating means for generating the area feature information from information on things or events that may affect or are related to the movement or stay of the subject;
a movement/stay time series information generating means for generating information related to a time series of the movement or stay of the target from the acquired terminal location information;
A movement and stay estimation model according to any one of claims 1 to 9 ;
A movement and stay prediction means for determining information related to the subsequent movement or stay using the movement and stay estimation model ;
and a movement trajectory prediction means for predicting information relating to the movement trajectory of the object based on the information relating to the subsequent movement or stay determined sequentially. An area feature generating means for generating the area feature information from information on things or events that may affect or are related to the movement or stay of the subject;
a movement/stay time series information generating means for generating information related to a time series of the movement or stay of the target from the acquired terminal location information;
a movement and stay prediction means for providing the area feature information and information related to a time series of the movement or stay of the target as input information to the movement and stay estimation model according to any one of claims 1 to 9, and determining information generated by the movement and stay estimation unit as information related to the subsequent movement or stay;
a movement trajectory prediction means for predicting information related to the movement trajectory of the target based on the information related to the subsequent movement or stay determined in sequence;
A movement trajectory prediction device comprising : 1. A computer-implemented movement and residence estimation method for estimating information related to a future movement or residence of an object in an area, comprising:
a step of taking as input area feature information generated from information relating to things or events that may affect or are related to the movement or stay of the subject in the area, and information relating to a time series of the movement or stay of the subject, and generating area feature embedded information in which the area feature information is embedded, and location embedded information in which information relating to the location in the time series is embedded, or generating area feature location embedded information in which information combining the area feature information and the information relating to the location in the time series is embedded;
a step of inputting the area feature embedding information and the location embedding information, or inputting information combining the area feature embedding information and the location embedding information, or inputting the area feature location embedding information, and generating area movement/stay embedded information in which information related to the movement or stay of the target in the area is embedded;
A movement and stay estimation method comprising a step of receiving the area movement and stay embedding information and generating information related to the subsequent movement or stay.

Images