Information processing method, information processing device, and information processing program

Abstract

An information processing device (1) acquires input data including a current location and remaining power of an electric moving body, calculates geographic feature values between the current location and multiple points on the basis of map information, calculates a predicted amount of power consumption of the electric moving body from the current location to each point by inputting the geographic feature values to a trained model obtained by machine-learning the relationship between the geographic feature value between two points and the amount of power consumption required to move between the two points, generates travelable range information indicating the travelable range of the electric moving body from the current location on the basis of the remaining power and the predicted amount of power consumption for each point, and outputs the travelable range information.

Description

Technical Field

[0001] The present disclosure relates to a technique for presenting the travelable range of an electric moving body.

Background Art

[0002] Patent Document 1 discloses a technique for calculating the travelable distance of a vehicle by evaluating a plurality of links extending outward from the position of the vehicle and displaying the travelable range of the vehicle on a map image based on the calculation result.

[0003] However, in the technique of Patent Document 1, since the travelable range is calculated based on rules, there is a problem that travelable range information accurately indicating the travelable range cannot be generated.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

[0005] The present disclosure has been made to solve such problems, and provides a technique for generating travelable range information that more accurately indicates the travelable range.

[0006] An information processing method in one aspect of the present disclosure is an information processing method in an information processing device that displays the drivable range of an electric mobile device on a display device, wherein the processor of the information processing device acquires input data including the current location and remaining power of the electric mobile device, calculates geographical feature quantities between the current location and a plurality of points based on map information, inputs the geographical feature quantities into a trained model obtained by machine learning the relationship between the geographical feature quantities between two points and the amount of power consumed for movement between the two points, calculates the predicted power consumption of the electric mobile device from the current location to each point, generates drivable range information indicating the drivable range of the electric mobile device from the current location based on the predicted power consumption of each point and the remaining power, and outputs the drivable range information to the display device.

[0007] According to this disclosure, it is possible to generate drivable range information that more accurately indicates the drivable range. [Brief explanation of the drawing]

[0008] [Figure 1] This is a block diagram showing an example of the configuration of an information processing device in Embodiment 1 of the present disclosure. [Figure 2] A flowchart illustrating an example of processing in the information processing device according to Embodiment 1. [Figure 3] This flowchart shows a detailed example of the process in step S4 shown in Figure 2 in Embodiment 1. [Figure 4] This is a diagram showing the map image of the first example in Embodiment 1. [Figure 5] This figure shows a map image of the second example in Embodiment 1. [Figure 6] This flowchart shows a detailed example of the process in step S4 shown in Figure 2 of Embodiment 2. [Figure 7] This figure shows the map image of the first example in Embodiment 2. [Figure 8] This figure shows a map image of the second example in Embodiment 2. [Figure 9]This figure shows a map image of the second example in Embodiment 2. [Figure 10] This flowchart shows an example of the processing performed by the information processing device in Embodiment 3 when performing machine learning. [Modes for carrying out the invention]

[0009] (Knowledge forming the basis of this disclosure) Because electric vehicles and other electrically powered vehicles have shorter driving ranges than gasoline-powered vehicles, it is necessary to display the driving range from the current location on a map image. On the other hand, the performance of electric vehicles changes over time, and consequently, their driving range also changes over time.

[0010] However, conventional technologies, including Patent Document 1, have the problem that they cannot flexibly respond to changes in the performance of electric mobile devices because the drivable range is calculated on a rule basis, and therefore cannot generate drivable range information that accurately indicates the drivable range.

[0011] This disclosure was made to address these issues.

[0012] (1) The above information processing method, in an information processing device that displays the drivable range of an electric mobile body on a display device, wherein the processor of the information processing device acquires input data including the current location and remaining power of the electric mobile body, calculates geographical features between the current location and a plurality of points based on map information, inputs the geographical features into a trained model obtained by machine learning the relationship between the geographical features between two points and the amount of power consumed for movement between the two points, calculates the predicted power consumption of the electric mobile body from the current location to each point, generates drivable range information indicating the drivable range of the electric mobile body from the current location based on the predicted power consumption of each point and the remaining power, and outputs the drivable range information to the display device.

[0013] According to this configuration, a learned model obtained by machine learning the relationship between the geographical feature quantity between two points and the power consumption required for moving between the two points is used. Therefore, the running history data of the electric moving body can be acquired at any time, and the learned model can be updated at any time using the acquired running history data. As a result, the learned model can be updated to conform to the performance of the latest electric moving body, and the predicted power consumption from the current location to each point can be accurately calculated. As a result, the travelable range information indicating the travelable range accurately can be generated.

[0014] Furthermore, according to this configuration, since the geographical feature quantity between the current value and a plurality of points is calculated by the information processing device, the labor of calculating the geographical feature quantity by the display device is saved, and the processing cost of the display device can be reduced. In addition, since the electric moving body does not need to provide the running history data including the geographical feature quantity, the running history data can be easily provided. As a result, an environment in which a larger amount of running history data is provided is prepared, and a learned model conforming to the performance of the latest electric moving body can be easily generated.

[0015] (2) In the information processing method described in (1) above, the geographical feature quantity may be the distance between the two points.

[0016] According to this configuration, since the distance between two points is adopted as the geographical feature quantity, the predicted power consumption can be calculated more accurately.

[0017] (3) In the information processing method described in (1) or (2) above, the geographical feature quantity may further include at least one of the elevation difference and the number of intersections between the two points.

[0018] According to this configuration, since at least one of the elevation difference and the number of intersections between two points is further used as the geographical feature quantity, the predicted power consumption can be calculated more accurately.

[0019] (4) In the information processing method according to any one of (1) to (3) above, the learned model is further a model obtained by machine learning the relationship between at least one of user identification information, vehicle information regarding the electric moving body, battery information regarding the battery mounted on the electric moving body, and date and time information indicating the travel date and time, and the power consumption, and the input data may further include at least one of the user identification information, the vehicle information, the battery information, and the date and time information.

[0020] According to this configuration, input data further including at least one of user identification information, vehicle information, battery information, and date and time information is input into the learned model that is further machine-learned using the relationship between at least one of user identification information, vehicle information, battery information, and date and time information, and the power consumption. Therefore, the predicted power consumption corresponding to at least one of user identification information, vehicle information, battery information, and date and time information can be calculated.

[0021] (5) In the information processing method according to any one of (1) to (4) above, in the generation, the travelability at the plurality of points is determined, and the travelable range information may include a map image in which the plurality of points are displayed in a display mode corresponding to the travelability.

[0022] According to this configuration, since a map image in which a plurality of points are displayed in a display mode corresponding to the travelable range is displayed, the user can easily confirm the travelable range.

[0023] (6) In the information processing method according to (5) above, the map image may display a point where the remaining battery power is greater than or equal to the predicted power consumption in a display mode indicating that it is travelable, and display a point where the remaining battery power is less than the predicted power consumption in a display mode indicating that it is not travelable.

[0024] With this configuration, locations where the remaining power is greater than or equal to the predicted power consumption are displayed in a way that indicates that driving is possible, and locations where the remaining power is less than the predicted power consumption are displayed in a way that indicates that driving is not possible. Thus, the map image can display driving possibility in a binary manner.

[0025] (7) In the information processing method described in (5) or (6) above, the map image may display a route from the current location to a drivable point.

[0026] This configuration allows the user to be shown a route to a point where they can travel.

[0027] (8) In the information processing method described in any of (5) to (7) above, the drivability may have a larger value as the ratio of the remaining power to the predicted power consumption increases.

[0028] With this configuration, the map image can display the drivability of each point as a continuous value, because the proportion of remaining power in the predicted power consumption increases with increasing value.

[0029] (9) In the information processing method described in any of (1) to (4) above, the drivable range information may include a map image showing the boundary of the drivable range.

[0030] With this configuration, the map image displays the boundaries of the drivable area, allowing users to easily understand the drivable range.

[0031] (10) In the information processing method described in (9) above, the trained model further calculates a confidence interval for the predicted power consumption, and the boundary may have at least one of its width and density changed according to the confidence interval.

[0032] With this configuration, at least one of the width and density is changed according to the confidence interval, allowing the user to understand the accuracy of the prediction of the drivable range.

[0033] (11) In the information processing method described in (9) or (10) above, the boundary includes an outer contour line and an inner contour line, wherein the outer contour line is a line connecting points where the upper limit of the confidence interval of the predicted power consumption that guarantees a first level of confidence is the remaining power, and the inner contour line is a line connecting points where the upper limit of the confidence interval of the predicted power consumption that guarantees a second level of confidence that is higher than the first level is the remaining power.

[0034] With this configuration, the boundary is displayed using an outer contour line generated by connecting points where the upper limit of the confidence interval for predicted power consumption, which guarantees a first level of confidence, is equal to the remaining power, and an inner contour line generated by connecting points where the upper limit of the confidence interval for predicted power consumption, which guarantees a second level of confidence (higher than the first level of confidence), is equal to the remaining power. Thus, a boundary with a width corresponding to the level of confidence can be represented.

[0035] (12) In the information processing method described in any of (1) to (11) above, the electric mobile device may further acquire driving history data including a departure point and an arrival point, and the amount of power consumed for the movement between the departure point and the arrival point, calculate geographical feature quantities between the departure point and the arrival point included in the driving history data based on the map information, and update the trained model using the geographical feature quantities and the amount of power consumed included in the driving history data.

[0036] With this configuration, geographical features are calculated from travel history data including the departure and arrival points and the amount of electricity consumed during the journey between the departure and arrival points. The trained model is then updated using the relationship between the calculated geographical features and the amount of electricity consumed, allowing the latest performance of the electric vehicle to be reflected in the trained model.

[0037] (13) In the information processing method described in any of (1) to (12) above, the input data includes the display range of the map image displayed on the display device, and the plurality of points may be points within the display range.

[0038] With this configuration, since the multiple locations are locations within the map image displayed on the display device, it is possible to prevent the calculation of geographical features for locations not displayed within the map image.

[0039] (14) An information processing device in another aspect of the present disclosure is an information processing device for displaying the drivable range of an electric mobile body on a display device, comprising: an acquisition unit for acquiring input data including the current location and remaining power of the electric mobile body; a calculation unit for calculating geographical features between the current location and a plurality of points based on map information; a prediction unit for calculating the predicted power consumption of the electric mobile body from the current location to each point by inputting the geographical features into a trained model obtained by machine learning the relationship between geographical features between two points and the amount of power consumed for movement between the two points; a generation unit for generating drivable range information indicating the drivable range of the electric mobile body from the current location based on the predicted power consumption and remaining power of each point; and an output unit for outputting the drivable range information.

[0040] This configuration provides an information processing device that can achieve the same effects as the information processing method described above.

[0041] (15) An information processing program in yet another aspect of the present disclosure is an information processing program that causes a computer to execute an information processing method for displaying the drivable range of an electric mobile body on a display device, the program causing the computer to execute a process that includes: acquiring input data including the current location and remaining power of the electric mobile body; calculating geographical feature quantities between the current location and a plurality of points based on map information; inputting the geographical feature quantities into a trained model obtained by machine learning the relationship between the geographical feature quantities between two points and the amount of power consumed for movement between the two points; calculating the predicted power consumption of the electric mobile body from the current location to each point; generating drivable range information indicating the drivable range of the electric mobile body from the current location based on the predicted power consumption and the remaining power of each point; and outputting the drivable range information.

[0042] This configuration provides an information processing device that can achieve the same effects as the information processing method described above.

[0043] (16) A method for manufacturing a trained model in yet another aspect of the present disclosure involves a computer acquiring travel history data from an electric mobile device, including a starting point and an arrival point, and the amount of electricity consumed during travel between the starting point and the arrival point; calculating geographical features between the starting point and the arrival point based on map information; and generating a trained model by machine learning the relationship between the geographical features and the amount of electricity consumed.

[0044] With this configuration, geographical features are calculated from travel history data including the departure and arrival points and the amount of electricity consumed during the journey between the departure and arrival points. The trained model is then updated using the relationship between the calculated geographical features and the amount of electricity consumed, allowing the latest performance of the electric vehicle to be reflected in the trained model.

[0045] This disclosure can also be implemented as an information update system operated by such an information processing program. It goes without saying that such a computer program can be distributed via a computer-readable non-temporary recording medium such as a CD-ROM or via a communication network such as the Internet. Furthermore, a learning device and a learning program for implementing the manufacturing method of the trained model of this disclosure may also be provided. It goes without saying that the learning program can also be distributed via a computer-readable non-temporary recording medium or via a communication network such as the Internet.

[0046] The embodiments described below are all specific examples of this disclosure. The numerical values, shapes, components, steps, and order of steps shown in the following embodiments are examples only and are not intended to limit this disclosure. Furthermore, among the components in the following embodiments, those not described in the independent claim representing the highest-level concept will be described as optional components. In addition, the contents of each embodiment can be combined.

[0047] (Embodiment 1) Figure 1 is a block diagram showing an example of the configuration of an information processing device 1 in Embodiment 1 of the present disclosure. The information processing device 1 is communicably connected to a user terminal 2 (an example of a display device) and a mobile device 3 via a network NT. The network NT is a wide-area communication network, including, for example, the Internet and a mobile phone network. In Figure 1, one user terminal 2 and one mobile device 3 are shown, but there may be multiple such terminals. The user terminal 2 and mobile device 3 are each uniquely identified by a communication address.

[0048] User terminal 2 is a portable information processing device such as a tablet computer or a smartphone. User terminal 2 is carried by the user riding in the mobile vehicle 3. Mobile vehicle 3 is, for example, an electric mobile vehicle. An electric mobile vehicle is a mobile vehicle that runs using electricity as its power source. An electric mobile vehicle includes a motor, a battery that supplies power to the motor, an inverter that controls the motor, and a display that shows map images, etc. Examples of electric mobile vehicles include electric cars, electric motorcycles, electric bicycles, electric kick scooters, etc. In the following description, the mobile vehicle will be assumed to be an electric car. The battery is a rechargeable secondary battery such as a lithium-ion battery or a nickel-metal hydride battery.

[0049] User terminal 2 sends input data to information processing device 1 to query the information processing device for the drivable range from its current location. Upon receiving the input data, information processing device 1 generates drivable range information indicating the drivable range from the current location and sends it to user terminal 2.

[0050] Mobile unit 3 is powered on and ready to operate. When mobile unit 3 finishes its operation and is powered off, it transmits the operation history data to the information processing device 1. The information processing device 1 updates the trained model for calculating predicted power consumption using the received operation history data. Alternatively, the operation history data may be transmitted to the information processing device 1 via the user terminal 2. In this case, the user terminal 2 and mobile unit 3 are connected via short-range wireless communication.

[0051] Next, the details of the configuration of the information processing device 1 will be described. The information processing device 1 includes a communication unit 11, an inference processor 12, a memory 13, and a learning processor 14. The communication unit 11 is a communication circuit that connects the information processing device 1 to the network NT. The communication unit 11 receives input data from the user terminal 2 and transmits drivable range information to the user terminal 2. Furthermore, the communication unit 11 receives driving history data from the mobile body 3.

[0052] The inference processor 12 is comprised of, for example, a central processing unit and includes a first acquisition unit 121, a first calculation unit 122, a prediction unit 123, a generation unit 124, and an output unit 125. The first acquisition unit 121 to the output unit 125 are realized by the central processing unit executing a predetermined information processing program. However, this is just one example, and the first acquisition unit 121 to the output unit 125 may also be realized by a dedicated electrical circuit such as an ASIC.

[0053] The first acquisition unit 121 acquires input data transmitted from the user terminal 2 using the communication unit 11. The input data includes the current location, remaining power, and display range of the mobile device 3. The current location is location information including latitude and longitude. The current location may also include the altitude of the mobile device 3. The remaining power is the current remaining power of the battery of the mobile device 3. The display range is information indicating the range of the map image displayed on the user terminal 2. The display range includes, for example, the current location and the location information of the top-left (or top-right) vertex and the bottom-right (or bottom-left) vertex of the map image displayed on the user terminal 2. The location information includes latitude and longitude. The display range is changed based on the amount of user operation such as pinch-out or pinch-in on the map image displayed on the user terminal 2.

[0054] The input data may also include additional information. This additional information includes at least one of the following: user identification information, vehicle information relating to the mobile unit 3, battery information relating to the battery installed in the mobile unit 3, and date and time information.

[0055] User identification information includes, for example, user identifiers (user ID), weight, gender, age, and height, and other information indicating the user's characteristics. Vehicle information includes, for example, vehicle type, identifier of mobile unit 3 (mobile unit ID), and weight of mobile unit 3, and other information indicating the vehicle's characteristics. Battery information includes, for example, battery identifiers (battery ID), battery type, and battery capacity, and other information indicating the battery's characteristics. Date and time information indicates the date and time of use of mobile unit 3, i.e., the date and time when the input data was entered. Date and time information includes, for example, year, month, day, day of the week, and time zone. The time zone may be the time the input data was transmitted, or it may be information indicating a time zone that includes the time the input data was transmitted, such as AM and PM. The information indicating the time zone may also be information indicating a time zone when the 24 hours are divided into predetermined time segments, such as 1 o'clock hour and 2 o'clock hour.

[0056] The first calculation unit 122 calculates geographical feature quantities between the current location and multiple points based on map information stored in the map information database 131, and inputs the calculated feature quantities to the prediction unit 123. Map information is information that represents a map using multiple nodes and links connecting each node. Nodes are location information that indicates characteristic locations on a road. Characteristic locations include, for example, road intersections and road endpoints. Links correspond to roads.

[0057] Multiple locations are predetermined locations located within the map area indicated by the display range included in the input data, and for example, major nodes among the nodes included in the map information fall into this category. Geographic features between the current location and multiple locations refer to three geographic features between P0 and P1, between P0 and P2, and between P0 and P3, if the multiple locations are, for example, locations P1, P2, and P3, with the current location being P0. In this case, the three geographic features are input to the prediction unit 123.

[0058] The first calculation unit 122 may calculate the distance of the shortest path between the current location and multiple points as a geographic feature using, for example, a known pathfinding algorithm. A known pathfinding algorithm is, for example, Dijkstra's algorithm. The shortest path is, for example, the path with the minimum cost. However, this is just an example, and the first calculation unit 122 may calculate the geographic feature between the current location and multiple points using an external map API (Application Programming Interface), etc. Furthermore, the first calculation unit 122 may calculate the geographic feature between the current location and multiple points by referring to a geographic feature database that stores the geographic feature between itself and other reachable nodes for each of a predetermined number of major nodes. Details of the algorithm for calculating geographic features are published in the following document: "Takuya Akiba, Yoichi Iwata, Yuichi Yoshida, Efficiency Improvement of Graph Shortest Path Query Processing by Direct Calculation of 2-Hop Labels, March 2014, DEIM Forum 2014, Internet (URL: https: / / db-event.jpn.org / deim2014 / final / proceedings / A8-3.pdf)".

[0059] Geographic features may further include at least one of elevation difference and the number of intersections. Elevation difference may be the elevation gain along the route connecting the current location and each point, or it may be cumulative elevation gain. Elevation gain indicates the elevation difference from the highest point on the route to the current location. Cumulative elevation indicates the sum of the elevation differences of uphill or downhill along the route.

[0060] The prediction unit 123 inputs geographical features into the trained model 133 to calculate the predicted power consumption of the mobile unit 3 from its current location to each point, and inputs the calculated predicted power consumption for each point into the generation unit 124. The trained model 133 is a model obtained by machine learning the relationship between geographical features between two points and the power consumption required for movement between the two points. Details of the trained model 133 will be described later. The two points are the starting point and destination point that the mobile unit 3 has traveled to in the past.

[0061] For example, if three geographical features are input, such as between P0 and P1, between P0 and P2, and between P0 and P3, the prediction unit 123 sequentially inputs each of these geographical features into the trained model 133. From the trained model 133, it sequentially obtains the predicted power consumption for each location P1, P2, and P3, and then associates the location ID with the obtained predicted power consumption and inputs it to the generation unit 124.

[0062] The generation unit 124 generates drivable range information indicating the drivable range of the mobile body 3 from its current location, based on the predicted power consumption at each location calculated by the prediction unit 123 and the remaining power acquired by the first acquisition unit 121.

[0063] The output unit 125 transmits the drivable range information generated by the generation unit 124 to the user terminal 2 using the communication unit 11. The drivable range information is information indicating the points where the mobile device 3 can travel within the display range of the user terminal 2, and includes, for example, a map image G1 as shown in Figure 4.

[0064] Memory 13 consists of a rewritable non-volatile storage device such as a hard disk drive or a solid-state drive. Memory 13 stores a map information database 131, a relational information database 132, and a trained model 133. The map information database 131 stores the map information described above. The relational information database 132 stores relational information used for machine learning and updating of the trained model 133. The relational information is information that associates geographical features with power consumption. The trained model 133 is a machine learning model used by the prediction unit 123 when calculating predicted power consumption.

[0065] The learning processor 14 is composed of, for example, a central processing unit and includes a second acquisition unit 141, a second calculation unit 142, and an update unit 143. The second acquisition unit 141 to the update unit 143 are realized by the central processing unit executing a predetermined learning program. However, this is just an example, and the second acquisition unit 141 to the update unit 143 may be composed of dedicated electrical circuits such as ASICs.

[0066] The second acquisition unit 141 acquires the travel history data transmitted from the mobile unit 3 using the communication unit 11. The travel history data is data that shows the travel history of the mobile unit 3 during one trip. One trip refers to the travel of the mobile unit 3 from when the power is turned on until the power is turned off again. The travel history data includes, for example, the departure point and destination point, and the amount of battery power consumed by the mobile unit 3 during the travel between the departure point and destination point. The departure point is the location information of the mobile unit 3 when the power of the mobile unit 3 is turned on. The destination point is the location information of the mobile unit 3 when the power of the mobile unit 3 is turned off. The location information includes latitude and longitude. Furthermore, the location information may include the difference in altitude between the departure point and destination point.

[0067] The driving history data may also include the additional information described above, similar to the input data. However, the date and time information included in the additional information of the driving history data indicates the departure and arrival times. The departure and arrival times include the departure time from the originating location and the arrival time from the destination location.

[0068] The second calculation unit 142 calculates geographical feature quantities between the departure point and destination point included in the driving history data acquired by the second acquisition unit 141, based on the map information stored in the map information database 131. The second calculation unit 142 then generates relational information by associating the calculated geographical feature quantities with the power consumption amounts included in the driving history data, and stores this information in the relational information database 132. The relational information database 132 stores, for example, one piece of relational information corresponding to one piece of driving history data. If the driving history data includes additional information, the relational information will also include the additional information in addition to the geographical feature quantities and power consumption amounts.

[0069] The update unit 143 updates the trained model 133 using the relevant information. The update unit 143 may update the trained model 133 at predetermined intervals, or it may update the trained model 133 each time a predetermined number of driving history data are acquired. The predetermined interval is, for example, one day, one week, one month, etc. The predetermined number is, for example, one, ten, one hundred, one thousand, etc. Details of updating the trained model 133 will be described later in Embodiment 3.

[0070] The trained model 133 is a machine learning model obtained by machine learning the relationship between geographical features included in relational information and power consumption. The trained model 133 can be composed of any machine learning model that performs supervised learning, for example. Examples of machine learning models include decision tree-based machine learning models such as random forests, neural networks, or linear regression models.

[0071] If the input data includes additional information, the trained model 133 is further trained using the additional information. In this case, the trained model 133 receives additional information in addition to the geographical features calculated by the first calculation unit 122, and outputs a predicted power consumption corresponding to the input geographical features and additional information.

[0072] The above describes the configuration of the information processing device 1. Next, the processing of the information processing device 1 will be explained. Figure 2 is a flowchart showing an example of the processing of the information processing device 1 in Embodiment 1.

[0073] In step S1, the first acquisition unit 121 acquires input data transmitted from the user terminal 2 using the communication unit 11. In step S2, the first calculation unit 122 identifies multiple points included within the map area indicated by the display range from the current location and display range included in the input data, and calculates the geographical feature quantities between each identified point and the current location by referring to the map information database 131.

[0074] In step S3, the prediction unit 123 calculates the predicted power consumption corresponding to the geographical feature amount of each location by inputting the geographical feature amount of each location calculated by the first calculation unit 122 into the learned model 133. When the learned model 133 is machine-learned using additional information in addition to the geographical feature amount, the prediction unit 123 may calculate the predicted power consumption of each location by inputting the additional information included in the input data in addition to the geographical feature amount calculated by the first calculation unit 122 into the learned model 133.

[0075] In step S4, the generation unit 124 generates the travelable range information based on the predicted power consumption of each location calculated in step S3 and the remaining power amount included in the input data. Details of the processing in step S4 will be described later.

[0076] In step S5, the output unit 125 transmits the travelable range information to the user terminal 2 using the communication unit 11.

[0077] FIG. 3 is a flowchart showing an example of the detailed processing of step S4 shown in FIG. 2 in Embodiment 1. In step S21, the generation unit 124 calculates the travelability of each location where the predicted power consumption is calculated based on the predicted power consumption and the remaining power amount. Assuming that the predicted power consumption is W1 and the remaining power amount is W0, the generation unit 124 determines a location where the remaining power amount W0 is greater than or equal to the predicted power consumption W1 (W0≧W1) as a travelable location, and determines a location where the remaining power amount W0 is less than the predicted power consumption W1 (W0<W1) as a non-travelable location. The generation unit 124 may assign a flag of "1" to the travelable location and a flag of "0" to the non-travelable location.

[0078] In step S22, the generation unit determines the display mode of each location. For example, the generation unit 124 determines the display mode of the travelable location as the first display mode indicating that it is travelable, and determines the display mode of the non-travelable location as the second display mode indicating that there is no travelability.

[0079] In step S23, the generation unit 124 generates a map image in which drivable locations are displayed in a first display mode and drivable locations are displayed in a second display mode, and generates drivable range information including the generated map image. Here, the generated map image is a map image having the display range included in the input data.

[0080] Figure 4 shows a map image G1 of the first example in Embodiment 1. Map image G1 includes a plurality of points 401. Point 401 is a point for which geographical features have been calculated by the first calculation unit 122. Point 401s indicates the current location. Point 401a indicates a drivable area. Point 401b indicates an undrivable area. Point 401a is displayed, for example, as a circular symbol with a first density (first display mode), and point 401b is displayed, for example, as a circular symbol with a second density (second display mode). The second density is lighter than the first density. This allows the user to easily grasp the drivable area. Here, the first and second display modes can be any display mode as long as they allow for the distinguishing display of drivable and undrivable areas. For example, the first display mode may be a circle, and the second display mode may be a shape other than a circle (e.g., a triangle, a square, etc.).

[0081] Figure 5 shows a map image G2 of the second example in Embodiment 1. In map image G2, each point 401 is displayed with a density corresponding to the drivability. In this case, in step S21, the generation unit 124 should calculate the drivability such that it has a larger value as the ratio of the remaining power amount W0 to the predicted power consumption amount W1 (W0 / W1) increases. Specifically, the generation unit 124 should calculate the ratio (W0 / W1), normalize the calculated ratio so that the minimum value is 0 and the maximum value is 1, and calculate the normalized value as the drivability. Then, the generation unit 124 should determine the display mode of each point with a density corresponding to the drivability and display each point on the map image with the determined density. As a result, the drivability of each point 401 is represented by a continuous value.

[0082] For example, location 401c is displayed with a higher concentration than location 401d because it is more likely to be traversable than location 401d. Location 401d is displayed with a higher concentration than location 401e because it is more likely to be traversable than location 401e.

[0083] As described above, the information processing device 1 in this embodiment uses a trained model 133 obtained by machine learning the relationship between geographical features between two points and the amount of power consumed during travel between the two points. Therefore, it is possible to acquire travel history data of the mobile device 3 at any time and update the trained model at any time using the acquired travel history data. As a result, the trained model 133 can be updated to match the performance of the latest mobile device 3, and the predicted power consumption from the current location to each point can be accurately calculated. Consequently, it is possible to generate travel range information that accurately indicates the travelable range.

[0084] Furthermore, with this configuration, since the geographical features of the current value and multiple locations are calculated by the information processing device 1, the effort of calculating geographical features on the user terminal 2 is eliminated, and the processing cost of the user terminal 2 can be reduced. In addition, the mobile device 3 does not need to provide driving history data that includes geographical features, so driving history data can be easily provided. As a result, an environment is created in which a larger amount of driving history data is provided, and trained models that match the performance of the latest electric mobile devices can be easily generated.

[0085] (Embodiment 2) Embodiment 1 displayed a map image in which each point was displayed in a manner corresponding to the drivability. Embodiment 2 generates a map image that displays the boundary of the drivable range. In this embodiment, the same reference numerals are used to denote the same components as in Embodiment 1. The difference between Embodiment 2 and Embodiment 1 lies in the details of the process in step S4 of Figure 2.

[0086] Figure 6 is a flowchart showing a detailed example of the processing of step S4 shown in Figure 2 in Embodiment 2. In Embodiment 2, the trained model 133 is a generalized linear model. Therefore, the trained model 133 outputs the predicted power consumption, the confidence level of the predicted power consumption, and the confidence interval of the predicted power consumption that guarantees the confidence level. Accordingly, the generation unit 124 receives the predicted power consumption, confidence level, and confidence interval for each point from the prediction unit 123. The confidence level includes a first confidence level and a second confidence level that is higher than the first confidence level.

[0087] A confidence interval is a range within which the probability that the amount of power consumption falls within that range is equal to the confidence level. It includes a first confidence interval when the confidence level is first confidence level and a second confidence interval when the confidence level is second confidence level. The first confidence level is, for example, 50%. However, this is just an example, and the first confidence level can be any appropriate value such as 50%, 40%, etc. The second confidence level is higher than the first confidence level, for example, 95%. However, this is just an example, and the second confidence level can be any appropriate value higher than the first confidence level, such as 80%, 90%, etc.

[0088] For example, when the remaining power is 60Wh, if the lower limit of the first confidence interval for a certain point is 0Wh and the upper limit is 60Wh, the probability of reaching that point is equal to the first confidence level. If the remaining power is 80Wh, it can be said that the point can be reached with a confidence level of at least the first confidence level. The second confidence level is higher than the first confidence level, so the upper limit of the second confidence interval is greater than the upper limit of the first confidence interval. In general, the power consumption increases at points farther from the starting point, so the upper limits of the first and second confidence intervals increase as you move further away from the starting point. Therefore, a point where the upper limit of the second confidence interval is equal to a certain power consumption is closer to the starting point than a point where the upper limit of the first confidence interval is equal to the same power consumption. The higher the accuracy of the trained model 133, the closer the first and second confidence intervals become, so these two points are closer together.

[0089] In step S31, the prediction unit 123 calculates the upper limit of the first confidence interval for the predicted power consumption for each of the multiple locations where the predicted power consumption has been calculated.

[0090] In step S32, the generation unit 124 extracts points where the upper limit of the first confidence interval is equal to the remaining power. For example, if the remaining power is 60Wh, points where the upper limit of the first confidence interval is 60Wh are extracted.

[0091] In step S33, the generation unit 124 generates an outer contour line of the boundary by connecting the points extracted in step S32. This outer contour line indicates the range where the probability of reaching it with the remaining power is at least the first confidence level.

[0092] In step S34, the prediction unit 123 calculates the upper limit of the second confidence interval for each of the multiple locations where the predicted power consumption has been calculated.

[0093] In step S35, the generation unit 124 extracts points where the upper limit of the second confidence interval is equal to the remaining energy. For example, if the remaining energy is 100Wh, points where the upper limit of the second confidence interval is 100Wh are extracted.

[0094] In step S36, the generation unit 124 generates an inner contour line of the boundary by connecting the points extracted in step S35. This outer contour line indicates the range where the probability of reaching it with the remaining power is at least the second confidence level.

[0095] In step S37, the generation unit 124 generates drivable area information, which includes a map image showing the outer boundary line and the inner boundary line.

[0096] Figure 7 shows the map image G3 of the first example in Embodiment 2. The boundary object 700, which indicates the boundary, is a donut-shaped image surrounded by an outer contour line 701 and an inner contour line 702. In map image G3, only the main points 401 are plotted, and the points connecting the outer contour line 701 and the inner contour line 702 are omitted from the illustration.

[0097] The outer contour line 701 connects points where the remaining power is at the upper limit of the first confidence interval, and the inner contour line 702 connects points where the remaining power is at the upper limit of the second confidence interval. Therefore, the width between the outer and inner contour lines widens in areas where the prediction accuracy of the trained model 133 is low. Conversely, the width between the outer and inner contour lines narrows in areas where the prediction accuracy of the trained model 133 is high. Thus, the boundary object 700 allows the user to recognize areas with high and low prediction accuracy for predicted power consumption based on the width between the outer contour line 701 and the inner contour line 702.

[0098] Furthermore, the generation unit 124 generates boundary objects 700 such that the density is higher closer to the inner contour line 702 and lower closer to the outer contour line 701. As a result, the density of the boundary objects 700 changes abruptly where the width between the outer contour line 701 and the inner contour line 702 is narrow, and changes more smoothly where the width is wider. This makes it easier for the user to recognize areas where the predicted power consumption accuracy is high and low. In the example in Figure 7, the generation unit 124 generates map image G3 such that the density inside the inner contour line 702 is displayed at a higher density than that of the boundary objects 700.

[0099] Figure 8 shows map image G4 of the second example in Embodiment 2. Boundary objects 700 are generated in map image G4 by the same process as in map image G3. However, in map image G4, the difference between the point where the remaining power is at the upper limit of the first confidence interval and the point where the remaining power is at the upper limit of the second confidence interval was generally smaller compared to map image G3. Therefore, the width of the boundary objects 700 in map image G4 is narrower than the width of the boundary objects 700 in map image G3.

[0100] Figure 9 shows map image G5 of the second example in Embodiment 2. Boundary objects 700 are generated in map image G5 by the same process as in map images G3 and G4. In map image G5, the difference between the point where the remaining power is at the upper limit of the first confidence interval and the point where the remaining power is at the upper limit of the second confidence interval was larger than in map image G4, but this difference was generally uniform. Therefore, the width of the boundary objects 700 in map image G5 is constant.

[0101] Thus, according to Embodiment 2, the width and density of the boundary object 700 are changed according to the accuracy of the trained model 133, so that the user can understand the accuracy of the prediction of the drivable range.

[0102] (Embodiment 3) The information processing device 1 in Embodiment 3 updates the trained model 133. Note that in Embodiment 3, the same reference numerals are used for components identical to those in Embodiment 1, and their descriptions are omitted. Figure 10 is a flowchart showing an example of the processing performed by the information processing device 1 in Embodiment 3 when performing machine learning. This flowchart is executed, for example, each time driving history data is acquired.

[0103] In step S51, the second acquisition unit 141 acquires the driving history data transmitted by the mobile body 3 using the communication unit 11.

[0104] In step S52, the second calculation unit 142 calculates the geographical feature quantities of the departure point and arrival point included in the driving history data based on the map information stored in the map information database 131.

[0105] In step S53, the second calculation unit 142 generates relational information by associating the geographical feature quantities calculated in step S52 with the power consumption amounts included in the driving history data. If the driving history data includes additional information, the relational information will include the additional information in addition to the geographical feature quantities and power consumption amounts.

[0106] In step S54, the second calculation unit 142 stores the relationship information in the relationship information database 132.

[0107] In step S55, the update unit 143 determines whether the number of driving history data acquired by the second acquisition unit 141 has increased by a predetermined number since the last update of the trained model 133. If it is determined that the number of driving history data has increased by a predetermined number (YES in step S55), the update unit 143 updates the trained model 133 using the increased relational information (step S56). In this case, the update unit 143 updates the trained model 133 by machine learning the trained model 133 so that when geographical features included in the relational information are input, the power consumption amount corresponding to that geographical feature is output. If additional information is included in the relational information, the update unit 143 should machine learn the trained model 133 so that when geographical features and additional information are input, the corresponding power consumption amount is output.

[0108] If it is determined that the number of driving history data points has not increased by a predetermined number (NO in step S55), the process ends.

[0109] Thus, according to the information processing device 1 in Embodiment 3, the trained model 133 is updated using relational information, so the latest performance of the mobile device 3 can be reflected in the trained model 133.

[0110] (modified version)

[0111] (1) In the above embodiment, the input data included a display range, but this disclosure is not limited thereto, and the display range does not have to be included in the input data. In this case, the first calculation unit 122 only needs to calculate the geographical features of points within a certain range from the current location.

[0112] (2) In Embodiment 1, map images G1 and G2 may display the route from the current location to a drivable point. In this case, the generation unit 124 may use a known route search algorithm to search for the shortest route from the current location to the drivable point and display it on map images G1 and G2. [Industrial applicability]

[0113] According to this disclosure, the drivable range can be accurately displayed, making it useful for application to electric mobile devices.

Claims

1. An information processing method in an information processing device that displays the drivable range of an electric mobile body on a display device, The processor of the aforementioned information processing device The input data, including the current location and remaining power of the electric mobile unit, is acquired. Based on map information, calculate geographical feature quantities between the current location and multiple points, By inputting the geographical features into a trained model obtained by machine learning the relationship between geographical features between two points and the amount of power consumed during travel between those two points, the predicted power consumption of the electric mobile device from the current location to each point is calculated. Based on the predicted power consumption and remaining power consumption at each location, drivable range information is generated indicating the drivable range of the electric mobile body from the current location. The drivable range information is output to the display device, The drivable range information includes a map image showing the boundaries of the drivable range. The trained model further calculates a confidence interval for the predicted power consumption, The boundary is modified in such a way that at least one of its width and density is changed according to the confidence interval. Information processing methods.

2. The aforementioned geographical feature is the distance between the two points. The information processing method according to claim 1.

3. The aforementioned geographical feature further includes at least one of the elevation difference between the two points and the number of intersections, The information processing method according to claim 1 or 2.

4. The aforementioned trained model is a model obtained by machine learning the relationship between user identification information, vehicle information relating to the electric mobile device, battery information relating to the battery installed in the electric mobile device, and date and time information indicating the date and time of operation, and the amount of power consumed. The input data further includes at least one of the user identification information, the vehicle information, the battery information, and the date and time information. The information processing method according to claim 1.

5. In the above generation, the feasibility of driving at the multiple locations is determined, The drivable range information includes a map image in which the plurality of points are displayed in a display manner corresponding to the drivability. The information processing method according to claim 1.

6. The map image displays locations where the remaining power is equal to or greater than the predicted power consumption in a display mode indicating that travel is possible, and locations where the remaining power is less than the predicted power consumption in a display mode indicating that travel is impossible. The information processing method according to claim 5.

7. The aforementioned map image displays the route from the current location to a point that is drivable. The information processing method according to claim 5.

8. The aforementioned drivability has a larger value as the ratio of the remaining power to the predicted power consumption increases. The information processing method according to any one of claims 5 to 7.

9. The boundary includes an outer contour line and an inner contour line, The outer contour line is a line connecting points where the upper limit of the confidence interval of the predicted power consumption that guarantees the first level of confidence is the remaining power. The inner contour line is a line connecting points where the upper limit of the confidence interval for the predicted power consumption, which guarantees a second level of confidence higher than the first level of confidence, is equal to the remaining power consumption. The information processing method according to claim 1.

10. Furthermore, travel history data including the departure point and arrival point, and the amount of electricity consumed during the movement between the departure point and arrival point, is acquired from the electric mobile device. Based on the aforementioned map information, the geographical feature quantities between the departure point and the arrival point included in the driving history data are calculated. The trained model is updated using the geographical features and power consumption included in the aforementioned driving history data. The information processing method according to claim 1.

11. The input data includes the display range of the map image displayed on the display device, The aforementioned multiple locations are locations within the indicated range. The information processing method according to claim 1.

12. An information processing device that displays the range of travel of an electric mobile body on a display device, An acquisition unit that acquires input data including the current location and remaining power of the electric mobile unit, A calculation unit that calculates geographical feature quantities between the current location and multiple points based on map information, A prediction unit calculates the predicted power consumption of the electric mobile device from the current location to each point by inputting the geographical features into a trained model obtained by machine learning the relationship between geographical features between two points and the amount of power consumed for movement between the two points. A generation unit generates drivable range information indicating the drivable range of the electric mobile body from the current location, based on the predicted power consumption and remaining power consumption at each location. It comprises an output unit that outputs the aforementioned drivable range information, The drivable range information includes a map image showing the boundaries of the drivable range. The trained model further calculates a confidence interval for the predicted power consumption, The boundary is modified in such a way that at least one of its width and density is changed according to the confidence interval. Information processing device.

13. An information processing program that causes a computer to execute an information processing method for displaying the drivable range of an electric mobile device on a display device, To the aforementioned computer, The input data, including the current location and remaining power of the electric mobile unit, is acquired. Based on map information, calculate geographical feature quantities between the current location and multiple points, By inputting the geographical features into a trained model obtained by machine learning the relationship between geographical features between two points and the amount of power consumed during travel between those two points, the predicted power consumption of the electric mobile device from the current location to each point is calculated. Based on the predicted power consumption and remaining power consumption at each location, drivable range information is generated indicating the drivable range of the electric mobile body from the current location. The process is executed to output the aforementioned drivable range information, The drivable range information includes a map image showing the boundaries of the drivable range. The trained model further calculates a confidence interval for the predicted power consumption, The boundary is modified in such a way that at least one of its width and density is changed according to the confidence interval. Information processing program.

Images