JP6692740B2 - Selection device, selection method, and selection program - Google Patents

Description

  The present invention relates to a selection device, a selection method, and a selection program.

  BACKGROUND ART Conventionally, there is known a car navigation (hereinafter, also referred to as “guidance”) technique for guiding a vehicle in which a user is riding to a destination by using a portable terminal device such as a smartphone. As an example of such a technique, the search cost is corrected according to the result of comparison between the route calculated using the data recorded in a predetermined recording medium and the actual traveling trajectory data, and based on the corrected search cost. A technique for calculating a route to a destination is known.

Japanese Patent No. 3596704

  However, in the above-mentioned conventional technique, it may not always be said that an appropriate route is calculated.

  For example, in the above-mentioned conventional technique, the navigation device installed in a predetermined vehicle only corrects the search cost according to the result of comparison between the traveling locus data of the vehicle and the calculated route. Therefore, in the conventional technique, even if a road that can reach the destination in a shorter time is created, there is a possibility that the road cannot be selected as a route if the user does not know the road.

  The present application has been made in view of the above, and an object thereof is to improve the calculation accuracy of a route in the guidance processing.

  The selection device according to the present application, the acquisition unit that acquires the movement information indicating the movement route from a predetermined terminal device, as a learning data of the model used to provide the movement route, the movement route having a higher evaluation than the reference movement route. It has a selection part which selects the movement information which shows or the movement information which shows a movement course whose evaluation is lower than a reference movement course from the acquired movement information.

  According to the aspect of the embodiment, it is possible to improve the calculation accuracy of the route in the guidance process.

FIG. 1 is a diagram for explaining an example of the operational effects of the information processing apparatus according to the embodiment. FIG. 2 is a diagram illustrating an example of a functional configuration of the information processing apparatus according to the embodiment. FIG. 3 is a diagram showing an example of information registered in the movement information database according to the embodiment. FIG. 4 is a flowchart showing an example of the flow of a selection process executed by the information distribution device according to the embodiment. FIG. 5 is a flowchart showing an example of the flow of processing in which the information distribution device according to the embodiment learns a model using learning data. FIG. 6 is a flowchart showing an example of the flow of processing in which the information distribution device according to the embodiment sets a reward. FIG. 7 is a flowchart showing an example of the flow of guidance processing executed by the information distribution device according to the embodiment. FIG. 8 is a diagram illustrating an example of the hardware configuration.

  Hereinafter, modes (hereinafter, referred to as “embodiments”) for implementing a selection device, a selection method, and a selection program according to the present application will be described in detail with reference to the drawings. Note that the selection device, the selection method, and the selection program according to the present application are not limited by this embodiment. Also, in each of the following embodiments, the same parts and processes are designated by the same reference numerals, and duplicated description will be omitted.

  Further, in the following description, as an example of a selection process executed by the information distribution device 10 which is an example of a selection device, a guidance process of providing a route for driving a vehicle to a destination to a user who drives the vehicle will be described. An example of the information distribution device to be executed will be described, but the embodiment is not limited to this. For example, the information distribution device 10 executes the following selection process when performing guidance processing for providing a route to a walking user or a user who moves by using an arbitrary moving means such as a bicycle or a motorcycle. May be. That is, the information distribution device 10 can apply the following selection processing in arbitrary guidance processing regardless of the means of transportation of the user.

[1. Concept of selection processing]
First, the concept of the selection process executed by the information distribution device 10 will be described with reference to FIG. FIG. 1 is a diagram for explaining an example of the operational effects of the information processing apparatus according to the embodiment. For example, the information distribution device 10 is realized by a single or a plurality of information processing devices such as a server device and a cloud system, and a network N (for example, see FIG. 2) such as a mobile communication network or a wireless LAN (Local Area Network). It is an information processing device capable of communicating with the terminal devices 101 to 103 (hereinafter, may be generically referred to as “terminal device 100”) used by the users U01 to U03 via the device. Although the three terminal devices 100 are described in the example illustrated in FIG. 1, the information distribution device 10 may be capable of communicating with any number of terminal devices 100.

  The terminal device 100 is, for example, a mobile terminal such as a smartphone, a tablet terminal or a PDA (Personal Digital Assistant), or an information processing device such as a notebook PC (Personal Computer). Note that, for example, when the terminal device 100 is an in-vehicle device installed and used in a vehicle, it is realized by, for example, a terminal of a car navigation system. For example, the terminal device 101 is a terminal device used by the user U01, the terminal device 102 is a terminal device used by the user U02, and the terminal device 103 is a terminal device used by the user U03. Shall be

  Here, when guiding the users U01 to U03 to a predetermined destination, the terminal device 100 executes a predetermined application such as a navigation system to perform the following guidance processing. First, the terminal device 100 identifies the current position by using various positioning techniques using GPS (Global Positioning System) and geomagnetism. Then, the terminal device 100 requests the guidance information indicating the route to the destination input by the user, with the identified current position as the departure place. Then, when the terminal device 100 receives the guide information from the information distribution device 10, the terminal device 100 uses the received guide information to execute the guide process for guiding the user to the destination.

  For example, the terminal device 100 includes a route to a destination where the vehicle can be used, information on a highway included in the route, traffic congestion information on the route, facilities serving as a guide mark, map information displayed on the screen, and guidance. It accepts guidance information that includes data such as voices and images such as maps that are sometimes output. In addition, the terminal device 100 measures the current position of the terminal device 100 using various satellite positioning systems. Then, the terminal device 100 displays an image such as a map included in the guidance information on a liquid crystal screen, electroluminescence, an LED (Light Emitting Diode) screen or the like (hereinafter, simply referred to as “screen”), and specifies the image. The current location is displayed on the map each time.

  In addition, the terminal device 100 displays a left turn, a right turn, a change in the lane to be used, an estimated time of arrival at the destination, etc., or an audio output from the speaker of the terminal device 100 or the vehicle, etc., depending on the specified present location. May be. It should be noted that the guidance processing executed by the terminal device 100 is applicable to various publicly known guidance processing technologies such as a car navigation system and a pedestrian navigation system, and detailed description thereof will be omitted.

  When the information distribution device 10 receives the guide information request from the terminal device 100, the information distribution device 10 sets the current position of the user as a departure point and provides it based on various elements from the route to the destination input by the user. Select a route. Then, the information distribution device 10 generates guide information indicating the specified route and distributes the generated guide information to the terminal device 100.

  Here, when specifying the route from the departure place to the destination, the route to be provided is selected according to information such as whether or not there is traffic congestion and the time required from the departure place to the destination. Technology can be considered. However, with such a technique, even if a road that can reach the destination in a shorter time is created, there is a possibility that the road cannot be selected as a route if the user does not know the road.

[2. Regarding selection processing executed by the information processing apparatus according to the embodiment]
Therefore, the information distribution device 10 executes the following selection process. For example, the information distribution device 10 acquires movement information indicating a movement route of a predetermined terminal device 100. Then, the information distribution device 10 uses, as the learning data of the model used to provide the movement route, movement information indicating a movement route having a higher evaluation than the reference movement route or a movement having a lower evaluation than the reference movement route. The movement information indicating the route is selected from the acquired movement information.

  For example, of the terminal devices 100, the terminal device 100 (for example, the terminal devices 101 and 102) that operates as a probe specifies the current position at predetermined time intervals, and identifies the specified current position and the time when the current position is specified. Correlate and store. Then, the terminal device 100 transmits the current position and time as movement information to the information distribution device 10 at a predetermined timing.

  In such a case, the information distribution device 10 identifies the travel route of the user and the time required to travel the travel route from the travel information. In addition, the information distribution device 10 determines a route from the same starting point (start point) to the destination (end point) as the moving route, the route being specified by the same process as the route selecting process in the above-described guidance process. Use as a reference route. The reference route may or may not have been actually provided to the terminal device 100 that serves as a probe in the guidance process. For example, the information distribution device 10 estimates a departure place and a destination of the terminal device 100 based on the movement route of the terminal device 100 that serves as a probe, and guides a reference route from the estimated departure place to the destination. It may be generated by a process similar to that at the time of executing.

  Then, the information distribution device 10 identifies the travel route having a higher evaluation than the reference route and the travel route having a lower evaluation than the reference route by comparing the travel route identified from the travel information with the reference route. For example, the information distribution device 10 identifies a route that reaches the destination in a shorter time than the reference route as a moving route with a high evaluation, and a route that reaches the destination in a longer time than the reference route has a low evaluation. Identified as a travel route. Then, the information distribution device 10 selects the identified travel route as the learning data of the model for selecting the route provided in the guidance process. The information distribution device 10 also sets a predetermined reward for the user of the terminal device 100 who has acquired the travel route selected as the learning data.

  Hereinafter, with reference to the drawings, an example of the functional configuration and operational effects of the information distribution device 10 that realizes the above-described selection processing will be described. In the following description, an example will be described in which the route indicated by the guidance information provided to the terminal devices 101 and 102 serving as probes is used as the reference route.

[2-1. Example of functional configuration]
FIG. 2 is a diagram illustrating an example of a functional configuration of the information processing apparatus according to the embodiment. As shown in FIG. 2, the information distribution device 10 includes a communication unit 20, a storage unit 30, and a control unit 40. The communication unit 20 is realized by, for example, a NIC (Network Interface Card) or the like. Then, the communication unit 20 is connected to the network N by wire or wirelessly, and transmits and receives questions and responses to and from the terminal device 100.

  The storage unit 30 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory (Flash Memory), or a storage device such as a hard disk or an optical disk. The storage unit 30 also includes a movement information database 31, a guidance information database 32, a learning data database 33, and a model database 34, which are various data for executing selection processing and guidance processing. Hereinafter, an example of information registered in the movement information database 31, the guidance information database 32, the learning data database 33, and the model database 34 will be described.

  The movement information acquired from the terminal device 100 is registered in the movement information database 31. For example, FIG. 3 is a diagram showing an example of information registered in the movement information database according to the embodiment. In the example shown in FIG. 3, information such as “movement information ID (Identifier)”, “acquisition source terminal ID”, “movement information”, and “date and time information” is registered in the movement information database 31. In addition to the information shown in FIG. 3, any information may be registered in the movement information database 31 as long as the information can be acquired by the terminal device 100 or information on the movement of the terminal device 100.

  Here, the "movement information ID" is an identifier of movement information. Further, the “acquisition source terminal ID” is an identifier of the terminal device 100 from which the movement information is acquired. Note that the “acquiring source terminal ID” may be an identifier of a user who uses the terminal device, that is, an identifier indicating a user who has moved on the movement route indicated by the movement information. Further, the “movement information” is information indicating the route along which the terminal device 100 has traveled, and is specifically position information indicating the position of the terminal device 100 acquired by the terminal device 100 at a predetermined time interval. The "date and time information" is information indicating the date and time when the associated position information was acquired.

  For example, in the example shown in FIG. 3, the movement information database 31 includes movement information ID “movement information # 1”, acquisition source terminal ID “terminal device # 1”, movement information “position # 1-1 to position # 1-”. 2 to position # 1 to 3 to position # 1 to 4 ..., and date and time information “2016/11/11/11: 10 to 2016/11/11/11: 11 to 2016/11/11/11 : 12 ~ 2016/11/11/11: 13 ~ ... "are registered in association with each other. In such information, the movement information indicated by the movement information ID “movement information # 1” is movement information indicating the movement path of the terminal device 100 indicated by the acquisition source terminal ID “terminal device # 1”, and the terminal device 100 The acquired position information is movement information “position # 1-1 to position # 1-2 to position # 1-3 to position # 1-4 to ...”, and the date and time when each position information is acquired is Date and time information "2016/11/11/11: 10-2016 / 11/11/11: 11-2016 / 11/11/11: 12-2016 / 11/11/11: 13 -..." Indicates.

  More conceptually, such information is obtained by the terminal device 100 indicated by the acquisition source terminal ID “terminal device # 1” from the date / time information “2016/11/11/11: 10” to the date / time information “2016/11 / 11/11: 13 ", the position indicated by the position information" position # 1-1 "to the position indicated by the position information" position # 1-2 "and the position information" position # 1-3 " The position is traced to indicate that the position information “position # 1-4” has been reached.

  Returning to FIG. 2, the description will be continued. In the guidance information database 32, various data used when performing guidance processing are registered. For example, the guidance information database 32 has registered therein route information indicating routes through which vehicles can travel, such as roads such as general roads and highways. More specifically, the guidance information database 32 is data indicating a roadway from a certain point to a certain point, which is data that forms a graph structure, that is, data that includes a set of polylines, and is output in the guidance process. Map data, voice data, etc. are stored. As the data registered in the guidance information database 32, various kinds of data used for publicly known guidance processing can be adopted.

  In the learning data database 33, the movement information selected as learning data is registered. For example, in the learning data database 33, the movement information selected as learning data among the information registered in the movement information database 31 shown in FIG. 3 is registered.

  In the model database 34, a selection model for selecting a route provided in the guidance process is registered. For example, in the model database 34, a model is registered that, when a route from a certain starting point to a certain destination is input, a numerical value that is an index of whether or not the route is the optimum route is output. For example, the model registered in the model database 34 is data indicating an input route, and when data of a node group and an edge group forming a polyline is input, the model traces the route indicated by the input data. Based on the time it takes to reach the destination, it outputs a numerical value that is an index of whether or not the route is the optimum route.

  Note that the type of technology that realizes such a model is not limited, and for example, a DNN (Deep Neural Network) including a plurality of node groups configured in multiple stages and connection paths (weights) connecting the node groups is used. ) It can be realized by other intelligent information processing technology, machine learning technology such as SVM (Support Vector Machine), statistical technology such as decision tree, multivariate analysis, and any other information processing technology. That is, the information distribution device 10 is used to select a route having the shortest required time from routes (hereinafter collectively referred to as “route candidates”) from a certain starting point to a certain destination in the guidance process. Any model may be adopted as long as it is a model.

  The control unit 40 is, for example, a controller and includes a CPU (Central Processing Unit), an MPU (Micro Processing Unit), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), and the like. Various programs stored in the internal storage device are implemented by executing a storage area such as a RAM as a work area. In the example illustrated in FIG. 2, the control unit 40 includes the acquisition unit 41, the selection unit 42, the learning unit 43, the setting unit 44, and the providing unit 45 (hereinafter, may be collectively referred to as the processing units 41 to 45). .).

  The connection relationship between the processing units 41 to 45 included in the control unit 40 is not limited to the connection relationship illustrated in FIG. 2 and may be another connection relationship. Further, the processing units 41 to 45 realize and execute the functions and actions (for example, FIG. 1) of the selection process and the learning process as described below, but these are functional units arranged for explanation. Therefore, it does not matter whether the hardware elements or software modules actually match. That is, the information distribution device 10 may realize and execute the guidance process in arbitrary function units as long as the functions and actions of the following selection process and learning process can be realized and executed.

[2-2. Example of Operation and Effect of Processing Executed by Information Distribution Device]
The contents of the selection process executed and realized by the processing units 41 to 45 will be described below with reference to the flowcharts shown in FIGS. FIG. 4 is a flowchart showing an example of the flow of a selection process executed by the information distribution device according to the embodiment. FIG. 5 is a flowchart showing an example of the flow of processing in which the information distribution device according to the embodiment learns a model using learning data. FIG. 6 is a flowchart showing an example of the flow of processing in which the information distribution device according to the embodiment sets a reward. FIG. 7 is a flowchart showing an example of the flow of guidance processing executed by the information distribution device according to the embodiment.

[2-3. Example of action and effect in selection processing]
First, an example of the flow of selection processing will be described with reference to FIG. For example, the acquisition unit 41 acquires movement information indicating a movement route of a predetermined terminal device 100 (step S101). More specifically, the acquisition unit 41 acquires, as the movement information, movement information indicating the movement route and the movement time of the movement route.

  For example, when the providing unit 45 receives from the terminal device 101 a guide information request that the departure point is the point A and the destination is the point B, the providing unit 45 displays the guide information indicating the reference route which is the route from the point A to the point B. It is provided to the terminal device 101 (step S1 in FIG. 1). In such a case, the acquisition unit 41 acquires the movement information from the terminal device 101 (step S2 in FIG. 1). For example, when the user U01 who uses the terminal device 101 moves via the route A from the point A to the point B instead of the reference route, the acquisition unit 41 acquires the movement information indicating the route A.

  Similarly, when the providing unit 45 receives a guide information request from the terminal device 102 that the departure point is the point A and the destination is the point B, the providing unit 45 displays the guide information indicating the reference route which is the route from the point A to the point B. It is provided to the terminal device 102 (step S3 in FIG. 1). In such a case, the acquisition unit 41 acquires the movement information from the terminal device 102 (step S4 in FIG. 1). For example, when the user U02 who uses the terminal device 102 moves via the route B from the point A to the point B instead of the reference route, the acquisition unit 41 acquires movement information indicating the route B. Then, the acquisition unit 41 registers the movement information acquired from the terminal devices 101 and 102 in the movement information database 31.

  The selection unit 42 reads the movement information from the movement information database 31, and specifies the movement mode and the movement route from the read movement information (step S102). For example, the selection unit 42 uses the movement information and the date and time information included in the movement information acquired from the terminal device 101, the route A from the point A to the point B, and the route A to the point B via the route A. Calculate the time required to reach it. In addition, the selection unit 42 specifies the driving mode such as the moving speed, the acceleration / deceleration, and the moving direction of the user U01 on the route A as the moving mode based on the moving information and the date and time information.

  Subsequently, the selection unit 42 identifies the reference route corresponding to the movement route identified by the acquisition unit 41 (step S103). For example, the selection unit 42 identifies the reference route notified in step S1 or step S3 shown in FIG. The selection unit 42 identifies the route traveled by the users U01 and U02 from the travel information, and the route that is the same as the identified route from the starting point and the destination and that is provided in the guidance process is used as a reference. You may specify as a route.

  Here, the selection unit 42 determines whether or not the movement route indicated by the movement information satisfies a predetermined condition (step S104, step S5 in FIG. 1). Then, when it is determined that the movement mode satisfies the predetermined condition (step S104: Yes), the selection unit 42 sets the movement information as a selection target. On the other hand, when the selection unit 42 determines that the movement mode does not satisfy the predetermined condition (step S104: No), the movement information is excluded from the selection targets (step S108), and the process ends.

  For example, when the required time when using the route A is shorter than the required time when using the reference route estimated in the guidance process, the route A is a better route than the reference route (that is, in the guidance process. It is also considered that the route should be provided preferentially). However, when the user U01 travels at a speed exceeding the legal speed when using the route A, the route A cannot be said to be a better route than the reference route.

  Therefore, when the moving speed when the terminal device 101, which is the acquisition source of the movement information, moves on the route A is within a predetermined range, the selection unit 42 uses the movement information indicating the route A as the learning data of the model. Select it. For example, the selection unit 42 refers to the guidance information database 32 and specifies the legal speed of the route A indicated by the movement information. Then, if the moving speed of the user U01 falls within the specified legal speed range, the moving information is selected as a selection target, and if the moving speed deviates from the legal speed, the moving information is excluded from the selection target.

  The selection unit 42 may use other conditions to determine whether or not the movement information is to be selected. For example, even if the user U01 is moving on the route A within the legal speed range, if the speed of movement of the surrounding vehicles is slower than U01, or faster than the user U01. If the user U01 has moved, the user U01 may not be said to have properly moved on the route A. Therefore, the selection unit 42 refers to the movement information acquired from the terminal device 100 other than the terminal device 101, and specifies the movement speed on the route A. When the difference between the moving speed of the terminal device 101 on the route A and the moving speed of the terminal devices 100 other than the terminal device 101 on the route A exceeds a predetermined threshold, the selecting unit 42 moves the moving route of the terminal device 101. The movement information indicating the route A may be excluded from the selection target. For example, when the difference between the moving speed of the terminal device 101 on the route A and the average moving speed of the terminal devices 100 other than the terminal device 101 is 15 km / h or more, the selection unit 42 acquires from the terminal device 101. The movement information on the route A may be excluded from the selection target.

  Further, for example, even when the user U01 is not making a temporary stop at a railroad crossing or an intersection, there is a possibility that it cannot be said that the user U01 has properly moved on the route A. Therefore, the selection unit 42 determines, based on the movement information, whether the movement mode of the terminal device 101 on the route A satisfies a predetermined condition on the route A, and when it determines that the predetermined condition is satisfied, The movement information on the route A acquired from the terminal device 101 may be selected. For example, the selection unit 42 refers to the guidance information and specifies a temporary stop location, a slow running section, and the like on the route A. Then, the selection unit 42 determines whether the user U01 has made a temporary stop at a temporary stop location based on the position information, the date information, and the like included in the movement information acquired from the terminal device 101, or performs the slow drive operation in the slow drive section (for example, speed per hour). It is determined whether various traffic rules on the route A have been adhered to, such as whether the vehicle has passed 30 km or less). Then, when the user U01 does not comply with the traffic rules, the selection unit 42 excludes the movement information on the route A of the terminal device 101 from the selection target.

  Further, for example, even when the user U01 is performing dangerous driving such as sudden acceleration / deceleration, meandering driving, frequent lane change, etc., it cannot be said that the user U01 has properly moved on the route A, and learning It may be inappropriate as data. Therefore, the selection unit 42 determines whether the movement mode of the terminal device 101 itself on the route A satisfies a predetermined condition based on the movement information, and when it determines that the predetermined movement condition is satisfied, the movement information is determined. May be selected. For example, the selection unit 42 determines, based on the movement information, whether dangerous driving is being performed such as not performing sudden acceleration / deceleration, or not performing meandering driving, and not performing dangerous driving. If it is determined that the movement information of the terminal device 101 on the route A may be selected.

  Further, for example, even if the user U01 complies with the legal speed on the route A, there is a possibility that it cannot be said that the user U01 has properly moved on the route A when the vehicle turns a curve at a high speed. Therefore, when the movement information indicates a movement route that draws a curve, the selection unit 42 determines whether or not the movement speed of the terminal device 101 on the movement route exceeds a predetermined threshold value, and when it is determined that the movement speed exceeds the predetermined threshold, The movement information may be excluded from the selection target. For example, the selection unit 42 refers to the guidance information and determines whether the route A includes a curve. Further, when the route A includes a curve, the selection unit 42 determines whether or not the terminal device 101 has moved the curve at a speed higher than a predetermined speed, based on the movement information. Here, the selection unit 42 may determine whether or not the terminal device 101 has moved at a speed faster than the speed corresponding to the curvature of the curve. Then, when the terminal device 101 moves along the curve at a speed faster than a predetermined speed, the selection unit 42 may exclude the movement information on the route A of the terminal device 101 from the selection target.

  In addition, for example, even if the user U01 complies with the legal speed on the route A, if the route A is a road set as a school zone or if a school or a residential area exists around the route A, guidance is provided. Providing route A in the process may not be appropriate. Therefore, the selection unit 42 may select the movement information to be selected based on the attribute of the area including the route A indicated by the movement information. For example, the selection unit 42 refers to the guidance information and designates a predetermined area such as a school or a residential area around the route A, or specifies the area around the route A as a predetermined section such as a first-class residential section. In such a case, the movement information on the route A of the terminal device 101 may be excluded from the selection target.

  In addition to the process described above, the selection unit 42 may exclude the user U01 indicated by the movement information from the selection target when the driving manner (that is, the movement manner) of the user U01 satisfies a predetermined condition. In addition, not only for the route A but also for the route B, the selection unit 42 similarly determines whether or not the movement information satisfies a predetermined condition.

  In addition, the selection unit 42 determines whether or not the evaluation of the selected moving route is higher than that of the reference route (step S105). When the selection unit 42 determines that the movement route is evaluated higher than the reference route (step S105: Yes), the movement information is registered as a positive example (step S106) in the learning data database 33, and the process ends. To do. On the other hand, when the selection unit 42 determines that the evaluation of the travel route is lower than that of the reference route (step S105: No), the travel information is registered as a negative example in the learning data database 33 (step S107), and the process ends. To do. That is, the selection unit 42 moves the movement information indicating the movement route having a higher evaluation than the reference route, which is the reference movement route, or the movement information indicating the movement route having a lower evaluation than the reference route, to the guidance process or the like. It is selected as the learning data of the model used to provide the route (step S6 in FIG. 1).

  For example, the selection unit 42 acquires movement information indicating a movement route having a shorter movement time than the reference route as movement information indicating a movement route having a higher evaluation than the reference route, and selects a movement route having a lower evaluation than the reference route. As the movement information to be shown, the movement information showing the movement route having a longer movement time than the reference route is selected. For example, the selecting unit 42 may select routes A and B from the same starting point to the same destination as the reference route, which are different from the reference route in the movement information indicating the routes A and B different from the reference route. The movement information indicating a movement route having a high evaluation or a movement route having a lower evaluation than the reference route is selected.

  More specifically, the selection unit 42 sets the movement route calculated as the movement route provided when the terminal devices 101 and 102 move from the point A to the point B as the reference route. That is, the selection unit 42 uses the route from the point A to the point B calculated in the guidance process and provided in steps S1 and S3 in FIG. 1 as the reference route.

  Further, the selection unit 42 estimates the required time from the point A to the point B using the reference route. Then, when the required time when using the route A is shorter than the estimated required time of the reference route, the selecting unit 42 selects the movement information of the route A acquired from the terminal device 101 as a positive example. .. On the other hand, for example, when the required time when the route B is used is longer than the estimated required time of the reference route, the selection unit 42 uses the movement information of the route B acquired from the terminal device 102 as a negative example. select.

  If the difference between the required time using the route A or the route B and the required time of the reference route is within a few seconds, even if the movement information indicating the route A or the route B is used as the learning data, much learning is required. It seems that the accuracy does not improve. Therefore, the selection unit 42, when the required time when the route A is used is shorter than the estimated required time of the reference route by a predetermined threshold value or more, the movement information of the route A acquired from the terminal device 101. May be selected as a positive example. In addition, the selection unit 42, when the time required for using the route B is longer than the estimated time required for the reference route by a predetermined threshold value or more, the movement information of the route B acquired from the terminal device 102. May be selected as a negative example.

  Further, for example, the selection unit 42 may select the movement information to be the learning data based on the movement route mode indicated by the movement information and the reference route mode. For example, the selection unit 42 compares the number of right and left turns on the movement route indicated by the movement information with the number of right and left turns on the reference route, and the difference between the numbers is within a predetermined range and the movement indicated by the movement information is indicated. If the travel time is shorter on the route, the travel information may be selected as a positive example. Further, when the number of right / left turns on the movement route indicated by the movement information is smaller than the number of right / left turns on the reference route, the selection unit 42 may select the movement information as a positive example. In addition, for example, the selection unit 42 may select the movement information as a positive example when the movement route indicated by the movement information includes only a portion that makes a left turn and there is no portion that makes a right turn. That is, the selection unit 42 may estimate movement information indicating a movement route that is better than the reference route or a movement route that is worse than the reference route, in accordance with the manner of turning right and left, the number of turns, and the like.

[2-4. Example of action and effect in the process of learning the model]
Next, an example of the flow of processing for learning a model using learning data will be described with reference to FIG. For example, the learning unit 43 sets a starting point and a destination (step S201) and reads the corresponding positive example and negative example from the learning data database 33 (step S202). Then, the learning unit 43 learns the model so that the route from the departure place to the destination does not become a positive example and a negative example (step S203), and the process ends.

  That is, the learning unit 43 uses the movement information selected by the selection unit 42 to learn the data of the model used for the guidance process (step S7 in FIG. 1). For example, the learning unit 43 reads the movement information registered as the learning data, and sets the plurality of positions indicated by the movement information as nodes, and generates the data of the polyline in which the nodes corresponding to the respective positions are connected in the order in which the terminal device 100 passed. To do. Then, the learning unit 43 outputs a score having a higher value, which is similar to the polyline of the movement information selected as a positive example, when the polyline indicating the route from the set starting point to the destination is input to the model, The model is trained so that a score with a lower value is output so as to be similar to the polyline of the movement information selected as a negative example. For example, the learning unit 43 may perform model learning using an arbitrary model generation technique such as back propagation. Then, the learning unit 43 registers the learned model in the model database 34.

[2-5. Example of action and effect in process of setting reward]
Next, an example of the flow of processing for setting a reward will be described with reference to FIG. The setting unit 44 determines whether the movement information is selected as the learning data (step S301). When the movement information is selected as the learning data (step S301: Yes), the setting unit 44 evaluates the user of the terminal device 100 that is the acquisition source of the movement information selected as the learning data. The reward according to the degree is set (step S302), and the process ends. Further, when the movement information is not selected as the learning data (step S301: No), the setting unit 44 ends the process as it is.

  For example, when the movement information acquired from the terminal device 101 and the movement information indicating the route A is selected as a positive example, the setting unit 44 uses the time required when the route A is used and the reference route. The difference from the required time at the time of the visit is calculated, and a reward corresponding to the calculated difference in the required time is set for the user U01 of the terminal device 101. More specifically, the setting unit 44 sets a larger amount of reward as the calculated difference in required time is larger. Similarly, when the movement information acquired from the terminal device 102 and the movement information indicating the route B is selected as a negative example, the setting unit 44 sets the required time and the reference route when the route B is used. The difference from the required time when used is calculated, and the reward of the amount according to the calculated difference in the required time is set to the user U02 of the terminal device 102.

  If the required time when using the route A is faster than the required time when using the reference route, it is highly possible that the dangerous driving was performed. Therefore, the setting unit 44 may set the reward only when the difference between the required time when the route A is used and the required time when the reference route is used is within a predetermined range.

  When the terminal device 100 uses various mobile communication systems to transmit the movement information to the information distribution device 10, there is a possibility that the movement information may not be properly acquired from the terminal device 100 due to the session holding relationship. Further, even if the terminal device 101 moves along the route A, it is difficult to specify whether the movement is caused by the driving of the user U01 or by the driving of another user. ..

  Therefore, the setting unit 44 may set a predetermined reward for the user who has performed authentication (that is, login) regarding a predetermined service via the terminal device 100. For example, when the movement information acquired from the terminal device 101 is selected as the learning data, the setting unit 44 is a user who has logged in to the navigation service or the like via the terminal device 101 when the movement information is acquired. Specify. That is, the setting unit 44 identifies the user who has logged in to the service provided by the information distribution device 10. Then, the setting unit 44 may set a predetermined reward for the identified user.

  The reward set by the setting unit 44 may be a financial reward or points in various point services. The reward set by the setting unit 44 may be a gift of various souvenirs or the like. The reward set by the setting unit 44 in this way is provided to the user by an external server device or the like.

[2-6. Example of action and effect in guidance processing]
Next, an example of the flow of processing for executing the guidance processing will be described with reference to FIG. 7. Prior to the processing shown in FIG. 7, the providing unit 45 accepts a guide information request from the terminal device 103 used by the user U03 (step S8 in FIG. 1). For example, the user U03 requests distribution of guidance information with the point A as the starting point and the point B as the destination. In such a case, the providing unit 45 uses the model learned in step S7 in FIG. 1 to generate a route candidate from the point A, which is the starting point, to the point B, which is the destination (step S7). S401).

  Further, the providing unit 45 inputs the generated polyline indicating each candidate into the model and provides the route with the highest score (step S402). For example, when the learning shown in step S7 in FIG. 1 is performed, the model outputs a score higher than the reference route for a route that is the same as the route A or close to the route A, and is the same as the route B or the route B. A score that is lower than that of the reference route will be output for the closer route. As a result, the providing unit 45 provides the guide information indicating the route A to the terminal device 103, for example (step S9 in FIG. 1). That is, the providing unit 45 has learned the user U03 who uses the terminal device 103 different from the terminal devices 101 and 102, which are probes, based on the movement information acquired by the terminal devices 101 and 102, which are probes. The path A selected using the model is provided as a new reference path.

  Here, the providing unit 45 determines whether or not the route different from the reference route is moving (step S403), and when the providing unit 45 is moving on the same route as the reference route ( Step S403: No), and the process is terminated. On the other hand, when the providing unit 45 is moving on a route different from the reference route (step S403: Yes), the providing unit 45 acquires movement information by the terminal device 100 moving on the reference route (step S404). Further, the providing unit 45 estimates the position when the terminal device 100 that is the target of the guidance process is traveling on the reference route (step S405). Then, the providing unit 45 distributes the information indicating the terminal device 100 that is the target of the guidance process and the information indicating the estimated position, displays the information on the terminal device 100 (step S406), and ends the process. That is, the providing unit 45 provides the information regarding the moving route on which the terminal device 100 is moving to the terminal device 100 moving on the moving route different from the reference route and reaching the same destination as the reference route. , Provides information related to the reference route.

  For example, the providing unit 45 estimates, based on the position of the terminal device 100 moving on the reference route, a position when the terminal device 100 moving on a moving route different from the reference route moves on the reference route, and estimates the position. The information regarding the determined position is provided as the information related to the reference route. Hereinafter, it is assumed that the user U03 starts moving to the point B via the route B instead of the route A which is the new reference route. In such a case, the providing unit 45 is a terminal device other than the terminal device 103, and based on the position information acquired from the moving terminal device 100 that is a new reference route, the route A, the occurrence of congestion and the route. Estimate various kinds of information regarding the movement through the route A, such as the average movement speed in the A. Further, the providing unit 45 estimates the current position when the user U03 moves using the route A based on the various estimated information. Then, the providing unit 45 distributes the information indicating the estimated current position to the terminal device 103, and displays the position indicated by the distributed information on the map. By performing such processing, the providing unit 45 can indicate to the user U03 whether or not the proposed route was good.

  As described above, the information distribution device 10 does not apply the user's personal preference of the movement mode to the guidance process as in the conventional technique, but rather uses the result actually used by each user for each route as a machine. It is adopted as learning data for learning, and the source of the learning data is provided with a reward for the successful experience of moving on a route better than the reference route. Therefore, the information distribution device 10 can efficiently collect the learning data of the model used for the guidance process, and thus can improve the route selection accuracy in the guidance process.

[3. Modification example)
The information distribution device 10 according to the above-described embodiment may be implemented in various different forms other than the above-described embodiment. Therefore, in the following, another embodiment of the information distribution device 10 will be described.

[3-1. About user evaluation]
For example, the information distribution device 10 may inquire of the user whether to use the movement information as learning data. Specifically, the information distribution device 10 uses the terminal device 100 that is the acquisition source of the movement information indicating a movement route having a higher evaluation than the reference route or the movement information indicating a movement route having a lower evaluation than the reference route. If the person is inquired about permission to use the movement information and the permission to use the movement information is accepted, the movement information may be selected as the learning data of the model.

  Further, the information distribution device 10 may further select, as model learning data, route information indicating a route highly evaluated by the user or route information indicating a low evaluation by the user. For example, when the guidance process for the user U01 is completed, the information distribution device 10 outputs a voice such as “How was the route this time?” And accepts the guidance evaluation. In such a case, for example, the terminal device 101 displays a slider, a five-level evaluation input screen, or the like, and accepts the evaluation from the user U01. Then, when the received evaluation exceeds a predetermined threshold value, the information distribution device 10 may perform model learning using the route information indicating the route provided in the guidance process as the learning data of the positive example. In addition, when the received evaluation is lower than a predetermined threshold, the information distribution device 10 may use the route information indicating the route provided in the guidance process as the negative learning data.

  The information distribution device 10 may estimate the evaluation of the guidance from the content of the user's statement, the tone of the voice, and the like. For example, the information distribution device 10 provides the information when the user U01 makes a sigh or a moody voice while guiding the user U01 from the point A to the point B, or when the user makes a word such as “worst”. The route information indicating the established route may be used as the negative learning data. That is, the information distribution device 10 may reflect any information in learning as long as the information indicates the user's voluntary evaluation of the provided route.

[3-2. About overlapping routes]
In addition, when the movement route indicated by the movement information repeatedly passes through the same section from the predetermined departure point to the predetermined destination, the information distribution device 10 determines that the movement information becomes the reference from the movement route. May be a travel route with a low evaluation. That is, the information distribution device 10 determines whether or not the user is lost on the route indicated by the movement information, and if the user is lost, it is estimated that the route indicated by the movement information is a route that is difficult to understand, and the movement is performed. Information may be a negative example.

  Further, the information distribution device 10 identifies the position of the intersection existing on the route, and determines whether or not the terminal device 100 has stopped for a predetermined time or longer at the identified position. That is, the information distribution device 10 determines whether or not the user hesitates in the traveling direction. Then, when the terminal device 100 has stopped at the intersection for a predetermined time or longer, the information distribution device 10 may use the movement information as a negative example. In addition, the information distribution device 10 may determine whether or not a U-turn has been performed based on the movement information, and if it has been performed, the movement information may be a negative example.

[3-3. Reference route]
In the example described above, the information distribution device 10 uses the route provided in the guidance process for guiding the user U01 from the point A to the point B as the reference route. However, the embodiment is not limited to this. For example, when the user U01 follows the route A and moves from the point A to the point B without performing the guidance process, the information distribution device 10 acquires the movement information indicating the route A from the terminal device 101. In such a case, the information distribution device 10 estimates the departure place and the destination where the user U01 has moved from the acquired movement information.

  For example, the information distribution device 10 estimates the position of the user U01's home, workplace, various POIs (Point Of Interest), etc., based on the history of the movement information that the terminal device 101 of the user U01 has acquired so far. However, it may be estimated that the movement having the estimated position as the destination is performed on the route indicated by the movement information. Then, the information distribution device 10 may use a model to select a route to be provided when guiding from the estimated starting point to the estimated destination and use the selected route as the reference route.

  That is, the information distribution device 10 is a route selected by the model used for the guidance process at the time of executing the selection process, from the same departure point to the same destination as the movement route indicated by the movement information. If so, any route may be set as the reference route. Further, such a reference route may be changed according to the attribute of the user. For example, the information distribution device 10 estimates, based on the travel route, whether the user likes the toll road, the open road, or the back road, and sets the reference route according to the estimation result. You may.

[3-4. About route]
Here, the information distribution device 10 may set an arbitrary section as a “moving route” or a “reference route”. That is, the information distribution device 10 may set an arbitrary position as the point A or the point B. For example, the information distribution device 10 may set a route from a point where the user starts moving to a point where the user ends the movement as one moving route, and a plurality of routes (hereinafter referred to as “routes”) obtained by dividing the route moved by the user into predetermined units. , "Section"). That is, the information distribution device 10 may select the learning data for each arbitrary section. In addition, the information distribution device 10 may use, as one moving route, a route having the current position specified by the terminal device 100 at a certain timing as a start point and the current position specified by the terminal device 100 as an end point.

  Further, the information distribution device 10 may set the range of the travel route based on the unit of the route used in the guidance process. For example, in the case where the guidance information database 32 includes data in which a branch point that is a branch of movement such as an intersection, an entrance of a highway or an interchange is a node, and a road or a lane connecting the branch points is a link. There is. In such a case, the information distribution device 10 may set a route from a branch point corresponding to a certain node to a branch point corresponding to another node as a “moving route” or a “reference route”. That is, the information distribution device 10 may set an arbitrary section as a movement route if the route that the user has traveled from a predetermined departure place to a predetermined destination is set as the movement route.

[3-5. Learning data]
In the above-described example, the information distribution device 10 uses the polyline corresponding to the movement route indicated by the movement information as the learning data, and learns the model so as to learn various characteristics of the polyline. However, the embodiment is not limited to this. That is, the information distribution device 10 may use the movement information itself as the learning data, or may use any information generated from the movement information as the learning data. That is, the information distribution device 10 may use any information as the learning data as long as the information indicates a route that is better or worse than the reference route. Further, the information distribution device 10 may cause the model to learn an arbitrary feature of the learning data by an arbitrary learning method.

[3-6. Other]
The information distribution device 10 may ask the user for consent to collect movement information in the guidance process, for example. For example, the information distribution device 10 distributes a message such as "Give rewards such as points, so please cooperate with the collection of movement information.", And collects movement information only when the user agrees. Good. Further, when the travel time when the route indicated by the travel information is used is shorter than the travel time when the reference route is used, the information distribution device 10 unconditionally selects the travel information as the learning data. Good.

  Further, of the processes described in the above embodiment, all or part of the processes described as being automatically performed may be manually performed, and conversely, the processes described as being manually performed. All or part of the above can be automatically performed by a known method. In addition, the processing procedures, specific names, information including various data and parameters shown in the above-mentioned documents and drawings can be arbitrarily changed unless otherwise specified. For example, the various kinds of information shown in each drawing are not limited to the illustrated information.

  Further, each component of each device shown in the drawings is functionally conceptual and does not necessarily have to be physically configured as shown. That is, the specific form of distribution / integration of each device is not limited to that shown in the figure, and all or a part of the device may be functionally or physically distributed / arranged in arbitrary units according to various loads or usage conditions. It can be integrated and configured.

  Further, the respective embodiments described above can be appropriately combined within a range in which the processing content is not inconsistent.

[3-7. program〕
The information distribution device 10 according to the above-described embodiment is realized by, for example, a computer 1000 having a configuration shown in FIG. FIG. 8 is a diagram illustrating an example of the hardware configuration. The computer 1000 is connected to an output device 1010 and an input device 1020, and an arithmetic device 1030, a primary storage device 1040, a secondary storage device 1050, an output IF (Interface) 1060, an input IF 1070, and a network IF 1080 are connected by a bus 1090. Have.

  The arithmetic unit 1030 operates based on a program stored in the primary storage device 1040 or the secondary storage device 1050, a program read from the input device 1020, or the like, and executes various processes. The primary storage device 1040 is a memory device such as a RAM that temporarily stores data used by the arithmetic device 1030 for various calculations. The secondary storage device 1050 is a storage device in which data used by the arithmetic device 1030 for various calculations and various databases are registered, and is a ROM (Read Only Memory), HDD (Hard Disk Drive), flash memory, or the like. It is realized by.

  The output IF 1060 is an interface for transmitting information to be output to an output device 1010 that outputs various kinds of information such as a monitor and a printer. For example, a USB (Universal Serial Bus) or a DVI (Digital Visual Interface), It is realized by a connector of a standard such as HDMI (registered trademark) (High Definition Multimedia Interface). The input IF 1070 is an interface for receiving information from various input devices 1020 such as a mouse, a keyboard, and a scanner, and is realized by, for example, USB.

  The input device 1020 is, for example, an optical recording medium such as a CD (Compact Disc), a DVD (Digital Versatile Disc), a PD (Phase change rewritable Disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), or a tape. It may be a device that reads information from a medium, a magnetic recording medium, a semiconductor memory, or the like. The input device 1020 may also be an external storage medium such as a USB memory.

  The network IF 1080 receives data from another device via the network N and sends the data to the arithmetic device 1030, and also transmits data generated by the arithmetic device 1030 via the network N to another device.

  The arithmetic device 1030 controls the output device 1010 and the input device 1020 via the output IF 1060 and the input IF 1070. For example, the arithmetic device 1030 loads a program from the input device 1020 or the secondary storage device 1050 onto the primary storage device 1040, and executes the loaded program.

  For example, when the computer 1000 functions as the information distribution device 10, the arithmetic device 1030 of the computer 1000 implements the function of the control unit 40 by executing the program loaded on the primary storage device 1040.

[4. effect〕
As described above, the information distribution device 10 acquires the movement information indicating the movement route of the terminal device 100. Then, the information distribution device 10 uses, as the learning data of the model used to provide the movement route, movement information indicating a movement route having a higher evaluation than the reference route or movement information indicating a movement route having a lower evaluation than the reference route. , Select from the acquired movement information. Therefore, the information distribution device 10 can use the history when the terminal device 100 actually moves on the route as the learning data, so that the selection accuracy when the model selects the moving route can be improved.

  Further, the information distribution device 10 acquires, as the movement information, movement information indicating the movement route and the movement time of the movement route. Then, the information distribution device 10 acquires movement information indicating a movement route having a shorter travel time than the reference route, as movement information indicating a movement route having a higher evaluation than the reference route, and a movement route having a lower evaluation than the reference route. The movement information indicating the movement route having a longer movement time than the reference route is selected as the movement information indicating the. Therefore, the information distribution device 10 can provide a route having a shorter travel time in the guidance process.

  In addition, the information distribution device 10 is a moving route from the same starting point to the same destination as the reference route, and the movement information indicating the moving route different from the reference route is more evaluated than the reference route. The movement information indicating a high movement path or a movement path having a lower evaluation than the reference path is selected. Therefore, the information distribution device 10 can select appropriate learning data for each route.

  Further, the information distribution device 10 uses the movement route provided in the guidance information or the like as the reference route when the terminal device 100 moves from the predetermined departure place to the predetermined destination. For this reason, the information distribution device 10 can use, as learning data, movement information on a route that is better or worse than the reference route that is provided in the guidance process when performing the selection process, and thus the route in the guidance process can be used. The selection accuracy of can be improved appropriately.

  Further, the information distribution device 10 acquires the moving speed when the terminal device 100 that is the acquisition source of the moving information moves on the moving route indicated by the moving information, and the moving speed is within a predetermined range as the model learning data. The movement information that becomes is the selection target. For example, the information distribution device 10 selects, as the learning data of the model, movement information whose movement speed is within the legal movement speed range on the movement route. Further, the information distribution device 10 acquires movement information indicating the movement routes of the plurality of terminal devices 100, the movement speed of the predetermined terminal device 100 on the predetermined movement route, and the other terminal devices 100 on the predetermined movement route. When the difference from the moving speed of No. 1 exceeds a predetermined threshold, the moving information indicating the moving route of the predetermined terminal device 100, which is the moving information indicating the predetermined moving route, is excluded from the selection target.

  In addition, the information distribution device 10 determines whether the movement mode of the predetermined terminal device 100 on the predetermined movement route satisfies the predetermined condition on the predetermined movement route based on the movement information, and determines the predetermined condition. If it is determined that the condition is satisfied, the movement information is selected. In addition, the information distribution device 10 determines whether the movement mode of the predetermined terminal device 100 satisfies a predetermined condition based on the movement information, and when it determines that the predetermined movement condition is satisfied, the movement is determined. Information is selected. Further, when the movement information indicates a movement route that draws a curve, the information distribution device 10 determines whether the movement speed of the predetermined terminal device 100 on the movement route exceeds a predetermined threshold value, and determines that the movement speed exceeds the predetermined threshold value. In that case, the movement information is excluded from the selection target. Further, the information distribution device 10 selects the movement information to be the learning data of the model based on the attribute of the area including the movement route indicated by the movement information.

  By the various processes described above, the information distribution device 10 can use the movement information obtained by the appropriate driving as the learning data, so that the learning of the model used for the guidance process can be further optimized.

  In addition, when the movement information collected from a predetermined terminal device 100 is used as model learning data, the information distribution device 10 sets a predetermined reward for the user associated with the predetermined terminal device 100. To do. For example, the information distribution device 10 sets a predetermined reward for the user who has been authenticated regarding a predetermined service via the predetermined terminal device 100. The information distribution device 10 also sets a reward according to the degree of evaluation. Therefore, the information distribution device 10 can provide the user with a motivation to find a better route, and thus can efficiently obtain better learning data.

  In addition, the information distribution device 10 is a moving route different from the reference route, and the moving route of the terminal device 100 to the terminal device 100 moving along the moving route reaching the same destination as the reference route. It provides information about the route and information related to the reference route. For example, the information distribution device 10 estimates the position when the terminal device 100 moving on a moving route different from the reference route moves on the reference route based on the position of the terminal device 100 moving on the reference route, Information about the estimated position is provided as information related to the reference route. Therefore, the information distribution device 10 can provide the user with the information when the reference route was selected.

  Further, the information distribution device 10 further selects route information indicating a moving route highly evaluated by the user or route information indicating a moving route low evaluated by the user as the model learning data. Therefore, the information distribution device 10 can apply the subjective evaluation of the user to the guidance process.

  In addition, when the movement route indicated by the movement information repeatedly passes through the same section from the predetermined departure place to the predetermined destination, the information distribution device 10 evaluates the movement information more than the reference route. Is a low travel route. Therefore, the information distribution device 10 excludes a route that is easily lost from the learning data, and thus can provide a more appropriate route in the guidance process.

  As described above, some of the embodiments of the present application have been described in detail based on the drawings, but these are examples, and various modifications based on the knowledge of those skilled in the art, including the modes described in the section of the disclosure of the invention, It is possible to implement the present invention in other forms with improvements.

  Further, the above-mentioned "section (module, unit)" can be read as "means" or "circuit". For example, the selection unit can be read as a selection unit or a selection circuit.

10 information distribution device 20 communication unit 30 storage unit 31 movement information database 32 guidance information database 33 learning data database 34 model database 40 control unit 41 acquisition unit 42 selection unit 43 learning unit 44 setting unit 45 providing unit 100 terminal device

Claims (19)

An acquisition unit that acquires movement information indicating a movement route of a predetermined terminal device,
It is the learning data of the model used to provide the travel route, and as the travel route whose evaluation is lower than the reference travel route , the same section is repeatedly passed from the predetermined departure place to the predetermined destination. A selecting device for selecting movement information indicating a movement route from the acquired movement information. The acquisition unit acquires, as the movement information, movement information indicating the movement route and the movement time of the movement route,
The selection unit obtains, as movement information indicating a movement route having a higher evaluation than the reference movement route, movement information indicating a movement route having a shorter movement time than the reference movement route, and moves the reference movement. The selection device according to claim 1, wherein, as the movement information indicating a movement route having a lower evaluation than that of the route, movement information indicating a movement route having a longer movement time than a reference movement route is selected. The selection unit is a movement route from the same starting point to the same destination as the reference movement route, and the movement information indicating a movement route different from the reference movement route is the reference. The selection device according to claim 1 or 2, characterized in that movement information indicating a movement route having a higher evaluation than a movement route having a lower rating or a movement route having a lower rating than the reference movement route is selected. The selection unit sets the movement route calculated as a movement route provided when the predetermined terminal device moves from a predetermined departure place to a predetermined destination as the reference movement route. The selection device according to claim 1. The selection unit acquires a moving speed when the terminal device that is the acquisition source of the moving information moves on the moving route indicated by the moving information, and as the learning data of the model, the moving speed is within a predetermined range. The selection device according to any one of claims 1 to 4, wherein: The selection device according to claim 5, wherein the selection unit selects, as the learning data of the model, movement information in which the movement speed falls within a legal movement speed range of the movement route. The acquisition unit acquires movement information indicating movement routes of a plurality of terminal devices,
If the difference between the moving speed of the predetermined terminal device on the predetermined moving route and the moving speed of the other terminal device on the predetermined moving route exceeds a predetermined threshold, the selecting unit selects the predetermined terminal device. The selection device according to claim 5 or 6, wherein movement information indicating a movement route, which is movement information indicating the predetermined movement route, is excluded from selection targets. Based on the movement information, the selection unit determines whether the movement mode of the predetermined terminal device on a predetermined movement route satisfies a predetermined condition on the predetermined movement route, and satisfies the predetermined condition. If it is determined that the movement information is present, the movement information is set as a selection target. The selection device according to claim 1. Based on the movement information, the selection unit determines whether the movement mode of the predetermined terminal device satisfies a predetermined condition, and when it determines that the predetermined condition is satisfied, the movement information is set to The selection device according to claim 1, wherein the selection device is a selection target. If the movement information indicates a movement route that draws a curve, the selection unit determines whether or not the movement speed of the predetermined terminal device on the movement route exceeds a predetermined threshold value, and if it is determined that the movement speed exceeds the predetermined threshold value, The selection device according to claim 1, wherein the movement information is excluded from selection targets. The selection unit selects movement information to be learning data of the model based on an attribute of a region including a movement route indicated by the movement information. The selection device described. When the movement information collected from the predetermined terminal device is used as the learning data of the model, it has a setting unit for setting a predetermined reward for the user associated with the predetermined terminal device. The selection device according to any one of claims 1 to 11. The selection device according to claim 12, wherein the setting unit sets the predetermined reward for a user who has been authenticated regarding a predetermined service via the predetermined terminal device. The selection device according to claim 12 or 13, wherein the setting unit sets a reward according to the degree of the evaluation. Regarding a moving route which is different from the reference moving route and which is moving on a moving route reaching the same destination as the reference moving route, the moving route on which the terminal device is moving The selection device according to claim 1, further comprising: a providing unit that provides information and information related to the reference movement route. The providing unit, based on the position of a terminal device moving along the reference moving route, moves a terminal device moving along a moving route different from the reference moving route along the reference moving route. The selection device according to claim 15, wherein the position at the time of estimation is estimated, and information regarding the estimated position is provided as information related to the reference movement route. The selection unit further selects route information indicating a moving route highly evaluated by the user or route information indicating a moving route low evaluated by the user as learning data of the model. The selection device according to any one of claims 1 to 16. An acquisition method performed by the selection device,
An acquisition step of acquiring movement information indicating a movement route from a predetermined terminal device;
It is the learning data of the model used to provide the travel route, and as the travel route whose evaluation is lower than the reference travel route , the same section is repeatedly passed from the predetermined departure place to the predetermined destination. A selection step of selecting movement information indicating a movement route from the acquired movement information. An acquisition procedure for acquiring movement information indicating a movement route from a predetermined terminal device,
It is the learning data of the model used to provide the travel route, and as the travel route whose evaluation is lower than the reference travel route , the same section is repeatedly passed from the predetermined departure place to the predetermined destination. A selection program that causes a computer to execute a selection procedure for selecting movement information indicating a movement route from the acquired movement information.

Images