Information processing system and information processing method

Abstract

To contribute to improving an occupancy rate of a commercial vehicle.SOLUTION: An information processing device comprises a display control unit that controls the display of a map including a search result for a recommended route from among a plurality of routes, on the basis of scores for each predetermined unit in a route that each of the plurality of route passes through based on passenger demand forecast data for a commercial vehicle. The present technology can be applied, for example, to an information processing device that displays a prediction result for taxi passenger demand.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] This technology is System and information processing method In particular, information processing that can contribute to improving the occupancy rate of commercial vehicles System and information processing method Regarding. [Background technology]

[0002] In the taxi industry, efforts to predict demand for taxis and implement more effective business operations have become more active (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2017-194863 Summary of the Invention [Problem to be solved by the invention]

[0004] The present technology has been developed in light of such circumstances, and is intended to contribute to improving the occupancy rates of commercial vehicles such as taxis. [Means for solving the problem]

[0005] An information processing device according to one aspect of the present technology includes a display control unit that controls the display of a map including search results for recommended routes from among a plurality of routes based on scores for each predetermined unit along the route that each of the plurality of routes passes through, which is based on passenger demand forecast data for commercial vehicles.

[0006] In one aspect of the present technology, the display of a map including search results for recommended routes from among multiple routes is controlled based on scores for each predetermined unit along the route that each of the multiple routes passes through, which is based on passenger demand forecast data for commercial vehicles.

[0007] The information processing device according to one aspect of the present technology can be realized by causing a computer to execute a program.

[0008] Furthermore, in order to realize the information processing device according to one aspect of the present technology, a program to be executed by a computer can be provided by being transmitted via a transmission medium or by being recorded on a recording medium.

[0009] The information processing device may be an independent device or an internal block constituting one device. [Effects of the Invention]

[0010] According to one aspect of the present technology, it is possible to contribute to improving the occupancy rate of commercial vehicles.

[0011] The effects described here are not necessarily limited to those described herein, and may be any of the effects described in this disclosure. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a block diagram showing a configuration example of an embodiment of a prediction system to which the present technology is applied. [Figure 2] FIG. 10 is a diagram illustrating an example of a demand prediction screen by a demand prediction application. [Figure 3] FIG. 1 is a block diagram illustrating an example configuration of a prediction system. [Figure 4] FIG. 2 is a diagram illustrating an example of vehicle behavior log data. [Figure 5] FIG. 10 is a diagram illustrating an example of generating actual vehicle data. [Figure 6] FIG. 10 is a diagram showing an example of actual vehicle sequence data. [Figure 7] 10 is a flowchart illustrating a process of generating actual vehicle sequence data. [Figure 8] 10 is a flowchart illustrating a learning prediction process. [Figure 9] FIG. 10 is a diagram showing an example of a result of a first clustering. [Figure 10] FIG. 10 is a diagram showing an example of a result of a first clustering. [Figure 11] FIG. 10 is a diagram showing an example of the results of two-stage clustering. [Figure 12] 10 is a flowchart illustrating an unknown area cluster classification process. [Figure 13] FIG. 10 is a diagram showing a first display example of a demand forecast screen. [Figure 14] FIG. 10 is a diagram showing a second display example of the demand forecast screen. [Figure 15] FIG. 10 is a diagram showing a third display example of a demand forecast screen. [Figure 16] FIG. 10 is a diagram illustrating learning of a riding position. [Figure 17] FIG. 10 is a diagram showing a fourth display example of the demand forecast screen. [Figure 18] FIG. 10 is a diagram showing a fifth display example of a demand forecast screen. [Figure 19] FIG. 10 is a diagram showing a sixth display example of a demand forecast screen. [Figure 20] FIG. 10 is a diagram showing an example of a fare prediction screen. [Figure 21] FIG. 10 is a diagram illustrating learning of a riding position. [Figure 22] FIG. 10 is a diagram illustrating learning of a drop-off location. [Figure 23] FIG. 10 is a diagram illustrating learning of a riding position. [Figure 24] FIG. 10 is a diagram showing an example of a demand forecast screen in the case of a wide-area map display. [Figure 25] FIG. 10 is a diagram showing an example of a demand forecast screen in the case of a detailed map display. [Figure 26] FIG. 10 is a diagram showing another example of a detailed map display. [Figure 27] 10 is a flowchart illustrating a voice guidance control process. [Figure 28] FIG. 10 is a diagram illustrating a heat map for extracting areas of expertise. [Figure 29] FIG. 10 is a diagram showing a display example of a recommended route presentation screen. [Figure 30] 10 is a flowchart illustrating a route presentation process. [Figure 31] FIG. 10 is a diagram illustrating how a riding point score is assigned. [Figure 32] FIG. 10 is a diagram showing an example of a demand forecast screen displaying no-boarding areas. [Figure 33] FIG. 10 is a diagram showing an example of a demand forecast screen displaying the location of delivery. [Figure 34] FIG. 10 is a diagram showing an example of a demand forecast screen displaying the time of the last train. [Figure 35] FIG. 10 is a diagram showing an example of a demand forecast screen on which the reverse lookup boarding point display function has been executed. [Figure 36] FIG. 10 is a diagram showing an example of a demand forecast screen on which the reverse lookup boarding point display function has been executed. [Figure 37] FIG. 10 is a diagram showing an example of a demand forecast screen that distinguishes between dispatch, on-demand, and drop-off. [Figure 38] FIG. 10 is a diagram showing an example of a demand forecast screen displaying an average fare and a confidence interval. [Figure 39] FIG. 10 is a diagram showing an example of a demand forecast screen displaying the number of available vehicles in real time. [Figure 40] FIG. 10 is a diagram showing an example of a demand forecast screen that displays the number of available vehicles in real time. [Figure 41] FIG. 10 is a diagram showing an example of an evaluation screen that outputs one day's business evaluation information. [Figure 42] FIG. 10 is a diagram illustrating an example of an operation history screen. [Figure 43] FIG. 10 is a diagram showing an example of a demand forecast screen that displays additional information taking into account distance and direction. [Figure 44] FIG. 10 is a diagram showing an example of a demand forecast screen displaying information according to a travel direction. [Figure 45] 1 is a block diagram illustrating an example of the configuration of an embodiment of a computer to which the present technology is applied. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, modes for carrying out the present technology (hereinafter referred to as embodiments) will be described in the following order. 1. Example of a prediction system configuration 2. Example of a demand forecasting app screen 3. Block Diagram 4.Generating actual vehicle sequence data 5. Learning prediction processing 6. Unknown area cluster classification process 7. Area AR combined display 8. Display of demand direction and frequency 9. Display pinpoint predictions 10. Display of estimated waiting time 11. Display of predicted length 12. Display of predicted riding distance 13. View estimated fare 14. Learn your boarding location 15. Learning drop-off locations 16. Learn your boarding location 17. Audio guide to passenger demand 18. Recommended route presentation processing 19. No-Riding Area Signs 20. Placement indication 21.Train time display 22. Reverse boarding point display 23. Display of demand forecast classification for dispatch / driving / waiting 24. View estimated fare 25. Display of real-time available vehicle numbers 26. Display of daily business evaluation 27. Displaying additional information taking distance and direction into account 28. Displaying information according to the direction of travel 29. Computer Configuration Example

[0014] <1. Example of prediction system configuration> FIG. 1 shows an example of the configuration of an embodiment of a prediction system to which the present technology is applied.

[0015] The prediction system 1 in FIG. 1 is composed of a plurality of taxis 11 and a server (information processing device) 12, and is a system that predicts demand for rides in the operating area of ​​the taxis 11 based on data acquired from the taxis 11.

[0016] The taxi 11 is a commercial vehicle that travels within a predetermined business area and picks up passengers. The taxi 11 is equipped with a fare meter 21, a vehicle management device 22, and a terminal device 23.

[0017] The fare meter 21 accepts the driver's operation of "occupied vehicle" and "empty vehicle." "Occupied vehicle" indicates a state in which the vehicle is traveling with passengers on board, while "empty vehicle" indicates a state in which the vehicle is traveling without passengers on board. When the fare meter 21 is "occupied vehicle," it calculates the fare (fare) according to at least one of the traveled time or distance, and displays it on a predetermined display unit.

[0018] The vehicle management device 22 generates vehicle movement log data that records the locations (routes) traveled by the taxi 11, the status of "occupied" or "vacant" and the like in chronological order at predetermined time intervals, and transmits the data to the server 12 via a predetermined network. The status of "occupied" or "vacant" is obtained from the fare meter 21.

[0019] The terminal device 23 is configured with an information processing device such as a smartphone, a tablet terminal, etc. The terminal device 23 stores an application program (hereinafter also simply referred to as a demand forecast application) that displays a demand forecast for riding on a display using riding demand forecast data transmitted from the server 12.

[0020] The demand prediction application is started and executed on the terminal device 23 by the driver's operation. The demand prediction application receives ride demand prediction data transmitted from the server 12 via a predetermined network, and displays the prediction results of predicting ride demand on a map based on the received ride demand prediction data on a display. A specific display example of the prediction results of predicting ride demand will be described later with reference to FIG. 2 etc.

[0021] The server 12 acquires vehicle movement log data via the network from the plurality of taxis 11. Then, the server 12 generates ride demand forecast data using the acquired large amount of vehicle movement log data, and transmits the data to each of the plurality of taxis 11 via the network.

[0022] The network connecting the server 12, the vehicle management device 22, and the terminal device 23 is configured, for example, by a mobile communication network such as a so-called 3G line or 4G line, the Internet, a public telephone line network, a satellite communication network, or the like.

[0023] The driver of the taxi 11 drives the taxi 11 to acquire passengers while referring to the passenger demand forecast displayed on the display of the terminal device 23 by the demand forecast application.

[0024] <2. Example of a demand forecasting app screen> FIG. 2 shows an example of a demand forecast screen displayed on the terminal device 23 by the demand forecast application.

[0025] The demand forecast screen of FIG. 2 displays a map 41, and also displays a current location mark 61, zoom buttons 62, a demand forecast mesh 63, a setting button 64, and the like superimposed on the map 41.

[0026] In addition, the demand forecast screen has a predicted time setting area 42 in an area different from the display area of ​​the map 41, and the predicted time setting area 42 includes a predicted time display 71 and predicted time change buttons 72A and 72B.

[0027] The current location mark 61 indicates the current location of the taxi 11. The zoom button 62 is operated to increase or decrease the scale of the map 41.

[0028] The demand forecast mesh 63 is configured by arranging a plurality of areas AR in a matrix. An area AR represents one region obtained by dividing the demand forecast mesh 63 into a grid. In the example of FIG. 2, 28 areas AR (4x7) are arranged in a partial region on the map 41, but the areas AR may be superimposed and displayed on the entire region on the map 41.

[0029] Each area AR in the demand forecast mesh 63 is displayed in a color or density according to the degree of ride demand, based on the ride demand forecast data transmitted from the server 12. For example, in Fig. 2, an area AR with a high density represents an area AR with high ride demand, and an area AR with a low density represents an area AR with low ride demand.

[0030] The setting button 64 is operated when making various settings related to the display of the demand forecast screen, such as the selection of items that can be displayed on the demand forecast screen, the display order, etc. Details of each item that can be displayed on the demand forecast screen will be described later.

[0031] The predicted time display 71 in the predicted time setting area 42 displays the corresponding time of the demand forecast displayed in the demand forecast mesh 63. In other words, the demand forecast mesh 63 displays the demand forecast for the time displayed in the predicted time display 71. Tapping the predicted time display 71 resets it to the current time. The predicted time change buttons 72A and 72B are operated to advance or reverse the predicted time in the predicted time display 71 in predetermined increments (for example, 10 minutes).

[0032] As described above, the demand forecasting application of the terminal device 23 receives the ride demand forecast data sent from the server 12, and displays the demand forecast mesh 63 that predicts the ride demand on the map 41 based on the received ride demand forecasting data as the forecast result on the display.

[0033] In the example of Figure 2, each area AR of the demand forecast mesh 63 is displayed in different colors and intensities depending on the degree of passenger demand, but as shown in Figure 13 described below, the predicted number of passengers can also be displayed together.

[0034] <3. Block diagram> Next, the detailed configuration of each device installed in the taxi 11 and the server 12 will be described.

[0035] FIG. 3 is a block diagram showing an example of the configuration of the server 12, the toll meter 21, the vehicle management device 22, and the terminal device 23.

[0036] The fare meter 21 receives an "occupied vehicle" or "empty vehicle" operation from the driver and displays the "occupied vehicle" or "empty vehicle" status and the fare (fare) on a predetermined display unit. The fare meter 21 supplies the "occupied vehicle" or "empty vehicle" status to the vehicle management device 22.

[0037] The vehicle management device 22 includes a position detection unit 101 , a speed detection unit 102 , a control unit 103 , a storage unit 104 , and a communication unit 105 .

[0038] The position detection unit 101 is configured with, for example, a GPS (Global Positioning System) receiver or the like, and receives positioning signals broadcast by positioning satellites to detect the current position of the taxi 11. The position detection unit 101 also includes a gyro sensor, a geomagnetic sensor, etc., and detects the traveling direction of the taxi 11.

[0039] The speed detection unit 102 is composed of a speed sensor, an acceleration sensor, etc., and detects the moving speed of the taxi 11. Note that the speed detection unit 102 may detect the moving speed of the taxi 11 by acquiring a measurement value from a speed sensor that detects the rotation speed of the wheels of the taxi 11.

[0040] The control unit 103 is composed of, for example, a CPU (Central Processing Unit), RAM (Random Access Memory), etc., and reads out an operation control program stored in the storage unit 104 and controls the operation of the entire vehicle management device 22 in accordance with the operation control program. Specifically, the control unit 103 acquires data at regular time intervals from each of the fare meter 21, the position detection unit 101, and the speed detection unit 102, generates vehicle movement log data, and stores the data in the storage unit 104. The control unit 103 also transmits the vehicle movement log data stored in the storage unit 104 to the server 12 via the communication unit 105 at predetermined timings, either periodically or irregularly.

[0041] The storage unit 104 is configured with, for example, a hard disk, a ROM (Read Only Memory), a RAM, and an NVRAM (Non-Volatile RAM), and stores vehicle dynamics log data. The communication unit 105 performs predetermined communication with the server 12 under the control of the control unit 103. The communication unit 105 is configured with a network interface that performs network communication via a predetermined network.

[0042] The server 12 includes a control unit 121 , a storage unit 122 , and a communication unit 123 .

[0043] The control unit 121 is configured with, for example, a CPU, a RAM, etc., and reads out an operation control program stored in the storage unit 122, and controls the operation of the entire server 12 in accordance with the operation control program.

[0044] The control unit 121 functionally includes at least a data generation unit 131, a learning unit 132, and a prediction unit 133, and predicts, by machine learning, the demand for rides for each area AR on the map 41. Any machine learning method can be selected, for example, a k-means method, a self-organizing map (SOM), a neural network, or an HMM (hidden Markov model).

[0045] The data generation unit 131 stores the vehicle movement log data acquired from each of the vehicle management devices 22 of the multiple taxis 11 via the communication unit 123 in the storage unit 122 .

[0046] FIG. 4 shows an example of vehicle movement log data generated by the vehicle management device 22 of the taxi 11 and transmitted to the server 12.

[0047] The vehicle management device 22 generates and accumulates vehicle movement log data at predetermined time intervals (for example, one minute intervals).

[0048] As shown in Figure 4, the items generated as vehicle movement log data include a company ID that identifies the company to which taxi 11 belongs, a radio ID that identifies taxi 11, a crew ID that identifies the driver on board taxi 11, a status time that indicates the time the status was generated, latitude and longitude that are the location information of taxi 11, direction and speed that indicate the driving speed and direction of travel of taxi 11, and a status of "occupied" or "empty."

[0049] The data generation unit 131 generates actual vehicle data, which is data relating to an actual vehicle, from the vehicle movement log data stored in the storage unit 122.

[0050] FIG. 5 shows an example of generating actual vehicle data.

[0051] The actual vehicle data is data extracted from the vehicle movement log data regarding the boarding of taxi 11, and is generated from information on the boarding change point in the vehicle movement log data where the status changes from "empty" to "actual vehicle" and the disembarking change point where the status changes from "actual vehicle" to "empty vehicle."

[0052] The actual vehicle data includes, for example, as shown in FIG. 5, the following items: ID, boarding time, departure point, arrival point, boarding time, boarding distance, and fare.

[0053] The ID is data that combines the company ID, radio ID, and crew ID from the vehicle movement log data.

[0054] The boarding time is calculated and recorded as the time between the status time of "empty vehicle" at the boarding change point and the status time of "occupied vehicle."

[0055] At the starting point, the latitude and longitude between the latitude and longitude of the "empty vehicle" at the boarding change point and the latitude and longitude of the "actual vehicle" are calculated and recorded.

[0056] At the arrival point, the latitude and longitude between the latitude and longitude of the "empty vehicle" at the drop-off change point and the latitude and longitude of the "actual vehicle" are calculated and recorded.

[0057] The boarding time is calculated and recorded as the time (in minutes, for example) from the boarding time to the time between the status time of "empty vehicle" at the disembarkation change point and the status time of "occupied vehicle."

[0058] As the travel distance, the distance from the departure point to the arrival point (unit: km, for example) is calculated and recorded.

[0059] The fare is calculated and recorded based on the travel time and distance in accordance with taxi fare regulations.

[0060] The method for calculating each item of the actual vehicle data is not limited to the above-described method, and other methods may be used. For example, each of the above items may be calculated from the first and last vehicle movement log data whose status is "actual vehicle." Furthermore, information on the fare and travel distance may be obtained from the vehicle management device 22 as part of the vehicle movement log data, rather than being calculated from the positions of the boarding and disembarking change points.

[0061] The data generation unit 131 generates actual vehicle sequence data, which is time-series data representing the number of passengers per predetermined time unit (10 minutes), for each area AR based on a large number of actual vehicle data generated from the vehicle movement log data of the vehicle management device 22 of a large number of taxis 11. For example, the data generation unit 131 generates actual vehicle sequence data, which is time-series data counting the number of passengers per 10 minutes, for each area AR.

[0062] FIG. 6 shows examples of actual vehicle sequence data for three areas AR, namely, area 1223, area 1224, and area 1225, among a plurality of areas AR obtained by dividing the business area of ​​the taxi 11.

[0063] The horizontal axis of the actual vehicle sequence data represents date and time, and the vertical axis represents the number of passengers. The actual vehicle sequence data shown in Figure 6 is data for eight days, but the period for creating the actual vehicle sequence data can be set to any period, such as one week, one month, or one year. For example, setting the period for creating the actual vehicle sequence data to one week makes it possible to capture fluctuations due to the day of the week, while setting it to a longer period, such as several months or one year, makes it possible to capture not only fluctuations due to the day of the week but also seasonal fluctuations such as the New Year holidays, Golden Week, and summer vacation.

[0064] In the example of actual vehicle sequence data for area 1223, for example, the number of actual vehicle data items whose boarding times fall between 10:00 and 10:10 on March 21, 2017 and whose departure points are located within area 1223 is counted as the number of boardings. The count result becomes the actual vehicle sequence data for area 1223 from 10:00 to 10:10 on March 21, 2017. Similar processing is performed for the entire period of the acquired actual vehicle data, and actual vehicle sequence data for area 1223 is generated.

[0065] Returning to Figure 3, the learning unit 132 generates a predictor by learning that predicts passenger demand using a large amount of long-term actual vehicle sequence data generated based on actual vehicle data obtained from the vehicle management devices 22 of a large number of taxis 11.

[0066] The prediction unit 133 predicts passenger demand at a predetermined time or time period using the predictor generated by the learning unit 132. The prediction result of the prediction unit 133 is transmitted to the terminal device 23 as passenger demand prediction data.

[0067] The storage unit 122 stores the vehicle movement log data acquired from each vehicle management device 22 and the actual vehicle sequence data generated from the vehicle movement log data. The storage unit 122 may also store actual vehicle data, which is intermediate data for generating the actual vehicle sequence data from the vehicle movement log data.

[0068] The communication unit 123 performs predetermined communication with the vehicle management device 22 and the terminal device 23 under the control of the control unit 121. The communication unit 123 is configured with a network interface that performs network communication via a predetermined network.

[0069] The terminal device 23 includes a control unit 141 , an operation unit 142 , a display unit 143 , a communication unit 144 , a speaker 145 , and a microphone 146 .

[0070] The control unit 141 is configured with, for example, a CPU, a RAM, etc., and controls the overall operation of the terminal device 23 in accordance with an operation control program stored in a storage unit (not shown). For example, the control unit 141 executes a demand prediction application based on an operation by a driver who is a user. The control unit 141 also functions as a display control unit that controls the display unit 143, and causes the display unit 143 to display the execution result of the demand prediction application, for example, the demand prediction screen of FIG. 2 .

[0071] The operation unit 142 is composed of a plurality of operation buttons provided on the terminal device 23, a touch panel superimposed on the display unit 143, etc., and accepts user operations and supplies an operation signal corresponding to the accepted operation to the control unit 141.

[0072] The display unit 143 is configured by, for example, an LCD (Liquid Crystal Display) or the like, and displays predetermined information such as the demand forecast screen of FIG.

[0073] The communication unit 144 performs predetermined communication with the server 12 under the control of the control unit 141. The communication unit 144 is configured with a network interface that performs network communication via a predetermined network.

[0074] The speaker 145 outputs sounds such as electronic sounds, sound effects, voice messages, etc. The microphone 146 detects voices emitted by the user and collects ambient sounds.

[0075] The server 12, the fare meter 21, the vehicle management device 22, and the terminal device 23 are configured as described above.

[0076] The following describes in detail the processes executed by the server 12, the vehicle management device 22, and the terminal device 23.

[0077] <4. Real vehicle sequence data generation process> First, the actual vehicle sequence data generation process by the server 12 will be described with reference to the flowchart of Fig. 7. This process can be executed at predetermined timing, for example, periodically or irregularly.

[0078] First, in step S1, the data generation unit 131 of the server 12 acquires (receives) vehicle movement log data transmitted via the network from each of the vehicle management devices 22 of the multiple taxis 11. Note that each vehicle management device 22 can transmit the vehicle movement log data to the server 12 individually at any timing, and does not have to transmit the data simultaneously.

[0079] In step S2, the data generation unit 131 generates actual vehicle data from the acquired vehicle movement log data. The actual vehicle data includes data calculated from items of the vehicle movement log data, such as boarding time and departure point, and external data added by the server 12, such as fares. Other external data may include, for example, date-related information related to dates such as days of the week, weekdays, or holidays, event information related to events held in the corresponding area AR on the data acquisition date, and weather information. By adding external data as actual vehicle data, it is possible to learn and predict ride demand for each situation, for example, depending on the day of the week, whether or not there is an event, the weather, etc.

[0080] In step S3, the data generation unit 131 generates actual vehicle sequence data for each area AR based on a large number of actual vehicle data generated from the vehicle management devices 22 of a large number of taxis 11, stores the data in the memory unit 122, and terminates the actual vehicle sequence data generation process.

[0081] <5. Learning prediction processing> Next, a learning and prediction process for learning and predicting passenger demand using the generated actual vehicle sequence data for each area AR will be described with reference to the flowchart in Fig. 8. This process can also be executed at predetermined timings, for example, periodically or irregularly.

[0082] First, in step S21, the learning unit 132 of the server 12 extracts a representative area from among a plurality of area ARs obtained by dividing the business region of the taxi 11. The learning unit 132 selects a predetermined number of area ARs from among the plurality of area ARs to set them as representative areas. The representative areas may be determined randomly, or may be selected by a knowledgeable user based on predetermined criteria, such as an area AR in the city center and an area AR in the suburbs, an area AR close to a station and an area AR far from a station, or an area AR with many stations and an area AR with few stations.

[0083] In step S22, the learning unit 132 performs two-stage clustering using the actual vehicle sequence data of each area AR extracted as a representative area. More specifically, the learning unit 132 performs first clustering to cluster each of the extracted areas AR using a first parameter, and performs second clustering to cluster each of the extracted areas AR using a second parameter.

[0084] For example, the learning unit 132 performs the first clustering using the average and variance of the number of passengers per unit time (e.g., per day) in the area AR as the first parameter, and performs the second clustering using the waveform of the average number of passengers per unit time (e.g., per day) in the area AR as the second parameter. Note that the clustering method may be, for example, the k-means method.

[0085] 9 and 10 show examples of the results of the first clustering in which a plurality of areas AR, which are representative areas, are clustered using the average and variance of the number of passengers as parameters.

[0086] FIG. 9 shows the distribution of a plurality of areas AR extracted as representative areas, with the horizontal axis representing the average and the vertical axis representing the variance.

[0087] 10 is a diagram showing the actual vehicle sequence data for a plurality of areas AR, which are representative areas, for each cluster. The horizontal axis of FIG. 10 represents time (from 0:00 to 24:00), and the vertical axis represents the number of passengers.

[0088] Since actual vehicle sequence data basically exhibits similar characteristics for each time period of the day (morning, noon, night, etc.), clustering is performed using data obtained by dividing the actual vehicle sequence data into basic units (one day).

[0089] In FIGS. 9 and 10, (the actual vehicle sequence data of) a plurality of areas AR extracted as representative areas are classified into six clusters.

[0090] FIG. 11 shows an example of a two-stage clustering result that combines the clustering results of the first clustering and the clustering results of the second clustering.

[0091] In Fig. 11, the horizontal direction (columns) represents the results of the first stage of clustering, and the vertical direction (rows) represents the results of the second stage of clustering. The horizontal and vertical axes of each graph arranged in a matrix are the same as those in Fig. 10.

[0092] In Figure 11, columns 1, 2, 3, and so on represent the clustering results of the first clustering. The vertically arranged area ARs are a group of area ARs with similar average and variance passenger counts (area AR groups). Rows A, B, C, and so on represent the clustering results of the second clustering. The area AR groups resulting from the first clustering were further clustered into area ARs with similar average passenger count waveforms. The numbers in each matrix graph represent the number of area ARs classified into that cluster. For example, the number "468" in the graph for cluster D-2 in row D and column 2 indicates that 468 area ARs out of the representative areas were classified into cluster D-2. The average and variance of the number of passengers per unit time, used as the first parameter, represent the size of the number of passengers per unit time and the magnitude of change in the number of passengers within a unit time, while the waveform of the average number of passengers per unit time, used as the second parameter, represents the trend of change over time in the number of passengers within a unit time.

[0093] The second stage of clustering may be performed individually for each of the first stage clustering results, or may be performed on all of the multiple areas AR extracted as representative areas, separately from the first stage clustering results.

[0094] In this embodiment, for example, the business area of ​​taxi 11 is divided into 4,400 area ARs, and half of the 4,400 area ARs, or 2,200 area ARs, are extracted as representative meshes. Two-stage clustering is then performed on the 2,200 area ARs, which are then classified into 44 clusters.

[0095] Next, in step S23 of FIG. 8, the learning unit 132 adjusts the learning parameters of the predictor, such as the learning rate, for each cluster using the actual vehicle sequence data belonging to the cluster, and then proceeds to step S24.

[0096] In step S24, the learning unit 132 uses the adjusted learning parameters and the actual vehicle sequence data of one or more areas AR belonging to the cluster to learn a predictor that predicts passenger demand for each cluster, and then proceeds to step S25.

[0097] In step S25, the prediction unit 133 predicts the demand for travel at a predetermined time in a predetermined area AR using the predictor generated by the learning unit 132. For example, when predicting the demand for travel in an area AR belonging to cluster C-4, the predictor of cluster C-4 is used to predict the demand for travel at a predetermined time.

[0098] The learning process of steps S21 to S24 and the prediction process of step S25 may be executed as consecutive processes, or the prediction process of step S25 may be executed at a timing different from the processes of steps S21 to S24.

[0099] For example, the process of step S25 is executed following the process of step S24, and the demand for rides at a predetermined time in each area AR constituting the operating region of the taxi 11 is calculated and stored in the storage unit 122. Then, in response to a request from the terminal device 23, the demand forecast data stored in the storage unit 122 is transmitted to the terminal device 23 as ride demand forecast data.

[0100] Alternatively, when terminal device 23 requests predicted data on passenger demand at a specified time for one or more areas AR, processing of step S25 is executed, and the processing result of step S25 is transmitted to terminal device 23 as passenger demand predicted data.

[0101] The demand forecasting application of the terminal device 23 that has received the passenger demand forecast data displays a demand forecasting mesh 63 with different colors and intensities depending on the number of passengers in each area AR, which is the forecast result, as shown in FIG.

[0102] According to the above learning prediction process, of the 4,400 area ARs that make up the business area, 2,200 area ARs are extracted as representative areas, and each of them is classified into a predetermined cluster, and passenger demand can be predicted based on the classification results.

[0103] On the other hand, for the remaining 2,200 area ARs (hereinafter referred to as unknown area ARs) that were not extracted as representative areas, it is unclear at this stage which cluster they will be classified into, and it is therefore not possible to predict passenger demand.

[0104] <6. Unknown area cluster classification processing> Next, the process for predicting passenger demand in the unknown area AR will be described.

[0105] An unknown area cluster classification process for determining the cluster to which the unknown area AR belongs will be described with reference to the flowchart of Fig. 12. This process can be executed at a predetermined timing, for example, periodically or irregularly.

[0106] First, in step S41, the learning unit 132 of the server 12 learns the characteristics (mean, variance, shape) of the actual vehicle sequence data of each cluster classified by the learning prediction process. In other words, the relationship between the actual vehicle sequence data of the 2200 areas AR extracted as representative areas and the clusters is learned by the learning device.

[0107] In step S42, the prediction unit 133 of the server 12 inputs the actual vehicle sequence data of the unknown area AR into a classifier that uses the parameters obtained by the learning in step S41, and determines the cluster of the unknown area AR.

[0108] As described above, the unknown area cluster classification process makes it possible to perform clustering of unknown areas AR other than the representative area using a classifier generated by learning the relationship between the clustering results of the representative area and the actual vehicle sequence data.

[0109] Once the cluster of the unknown area AR can be identified, the predictor of the identified cluster is used to perform the prediction process of step S25 described above, thereby making it possible to predict the demand for riding in the unknown area AR.

[0110] Therefore, by executing both the learning prediction process of FIG. 8 and the unknown area cluster classification process of FIG. 12, it is possible to predict the demand for rides in all 4,400 areas AR that make up the operating area of ​​taxi 11.

[0111] In the learning prediction process of Figure 8, by extracting representative areas, which is the process of step S21, the number of areas AR to be learned, in other words, the amount of data of actual vehicle sequence data, can be reduced, thereby reducing the computational load and the cost and time required for passenger demand prediction.

[0112] Furthermore, by performing two-stage clustering using the actual vehicle sequence data of each area AR extracted as a representative area in step S22, the number of learners can be reduced, thereby reducing the cost and time required for passenger demand forecasting. Specifically, if two-stage clustering is not performed, the number of learners required would be the number of area ARs extracted as representative areas (2,200), but by performing two-stage clustering and classifying the areas into a predetermined number of clusters, the number of learners required for learning can be reduced to the number of clusters (44).

[0113] The learning device for each cluster can use the actual vehicle sequence data of all area ARs classified into that cluster. That is, for example, when learning passenger demand forecast for area 1223, learning is generally performed using only the actual vehicle sequence data acquired in that area 1223. In contrast, with the present technology, if area 1223 is classified into cluster D-2, for example, and there are 468 area ARs belonging to cluster D-2, learning can be performed using the actual vehicle sequence data of the 468 area ARs, including area ARs other than area 1223. Therefore, learning can be performed for one learning device using a larger amount of data than the amount of data that can be acquired for one area AR, thereby improving prediction accuracy.

[0114] Furthermore, for unknown areas AR that are not extracted as representative areas in the learning prediction process, the cluster of the unknown area AR can be identified through the unknown area cluster classification process, and the predictor of the identified cluster can be used to predict the passenger demand for the unknown area AR.

[0115] In the above-mentioned steps S23 and S24, the learning parameters were adjusted and the predictor was trained using only the actual vehicle sequence data of each area AR extracted as a representative area. However, after clusters have been determined for all unknown area ARs included in the sales area, the learning parameters may also be adjusted and the predictor trained by adding the actual vehicle sequence data of the unknown area ARs.

[0116] Therefore, the prediction system 1 in Fig. 1 can learn and predict more efficiently, and can improve prediction accuracy with a small amount of data.

[0117] In the above-mentioned learning prediction process and unknown area cluster classification process, cluster classification and learning were performed using actual vehicle sequence data generated from all vehicle movement log data obtained from the vehicle management devices 22 of multiple taxis 11, regardless of the day of the week, weekday, holiday, etc.

[0118] However, the actual vehicle sequence data may be divided into categories such as day of the week, weekday, holiday, or weather, and cluster classification and learning may be performed for each category. This makes it possible to predict riding demand for each predetermined condition, such as day of the week, weekday, holiday, weather, or whether or not an event is taking place, and to display the prediction results on a display.

[0119] <7. Area AR combined display> Various display examples in which the demand prediction application of the terminal device 23 displays the predicted results of the riding demand on the display will be described below.

[0120] FIG. 13 shows a first display example of a demand forecast screen displayed by the demand forecast application.

[0121] On the demand forecast screen shown in Fig. 2, the demand forecast mesh 63 was configured by arranging areas AR of the same rectangular size in a matrix. Furthermore, the number of passengers in each area AR, which is the forecast result, was not displayed on the screen.

[0122] In contrast to this, in the demand forecast mesh 63 of FIG. 13, the number of passengers in each area AR, which is the forecast result, is displayed within the area AR.

[0123] Furthermore, the demand forecasting application combines multiple adjacent area ARs where the number of passengers is equal to or less than a predetermined threshold into a single area AR and displays the number of passengers. In the first display example of FIG. 13, multiple adjacent area ARs where the number of passengers is 10 or less are combined and displayed as a single area AR. Specifically, 2x2 area ARs where the number of passengers when displayed in the same rectangular size is "4," "2," "2," and "1" are combined into a single area AR and displayed as "9." Of course, depending on the number of passengers in the adjacent areas, they may not be combined even if they are 10 or less.

[0124] Since it is difficult to predict demand when the predicted number of passengers is small, such as 0, 1, or 2, the demand forecasting app can display demand forecasts in area AR units where the number of passengers is above a certain value. This improves the accuracy of predictions and provides more useful information to drivers.

[0125] The number of passengers displayed as a predicted result may be a value with a certain range, such as "10-13."

[0126] <8. Display of demand direction and frequency> FIG. 14 shows a second display example of the demand forecast screen displayed by the demand forecast application.

[0127] In FIG. 14, the display of colors and intensities according to the degree of riding demand for each area AR is omitted.

[0128] Figure 14 shows an example of a display that displays more detailed forecast results for an area AR that the driver is paying attention to (hereinafter referred to as the area of ​​interest AR) among the areas AR of the demand forecast mesh 63 superimposed on the map 41.

[0129] When the driver performs an operation to specify an area of ​​interest AR, such as by tapping (touching) a specific area AR from among the areas AR of the demand forecast mesh 63 superimposed on the map 41, the demand forecast app displays the specified area of ​​interest AR as shown in Figure 14.

[0130] 14, an area of ​​interest AR designated by the driver is displayed with an area of ​​interest frame 211 that is wider than the other areas AR. Arrows 212-1 to 212-8 pointing outward from the area of ​​interest frame 211 are also displayed. When there is no need to distinguish between the arrows 212-1 to 212-8, they will simply be referred to as arrows 212.

[0131] The direction of the arrow 212 represents the direction of movement of passengers boarding in the area of ​​interest AR, and the length of the arrow 212 represents the average travel distance of passengers boarding in the area of ​​interest AR and traveling in the direction of the arrow 212. Furthermore, the width of the arrow 212 (thickness in the direction perpendicular to the arrow direction) represents the proportion of passengers boarding in the direction indicated by the arrow 212 relative to all directions.

[0132] 14, the proportion of passengers boarding in the area of ​​interest AR is greater in the direction of arrow 212-3, and passengers boarding in the direction of arrow 212-4 travel a longer distance.Furthermore, for example, the proportion of passengers boarding in the area of ​​interest AR who travel in the direction of arrow 212-2 or 212-6 is smaller, and the travel distance is also shorter.

[0133] For example, when determining an area AR where the driver will do so-called "cruising" (looking for passengers while driving the taxi 11), the driver can set a specific area AR in the demand forecast mesh 63 as the area of ​​interest AR and display the arrow 212 to search for an area AR with many passengers going in the same direction as the driver's return.

[0134] The direction of movement of passengers in each area AR can be predicted by learning information including the direction (heading) of vehicle movement log data.

[0135] The number of arrows 212 to be displayed, in other words, the predicted number of passenger movement directions, may be a number other than the eight shown in Fig. 14. Furthermore, the proportion of passengers moving in the direction of the arrow 212 in all directions may be represented by a method other than the width of the arrow, such as by a different color or by a number.

[0136] <9. Display of pinpoint predictions> FIG. 15 shows a third display example of the demand forecast screen displayed by the demand forecast application.

[0137] FIG. 15 also shows a display example in which a more detailed prediction result is displayed when the driver selects a predetermined area AR as the area of ​​interest AR.

[0138] Within the area AR, which is created by dividing the operating area into specified units, there may be places where the boarding locations are fixed, such as taxi stands in front of stations or hotels, and where the number of boardings is higher than in other places.

[0139] If such boarding locations with a high number of boardings exist within the area of ​​interest AR, the demand forecasting app can pinpoint and predict and display the boarding locations with a high number of boardings and the number of boardings at those boarding locations, separately from the number of boardings for the entire area of ​​interest AR. Hereinafter, a boarding location with a high number of boardings identified within the area of ​​interest AR will be referred to as a pinpoint boarding location.

[0140] In Figure 15, a pinpoint boarding position mark 221 indicating the pinpoint boarding position is displayed at a predetermined position within the attention area AR, and a passenger count display 222 indicating the predicted number of passengers at the pinpoint boarding position mark 221 is displayed. In Figure 15, "43" displayed within the attention area frame 211 is the number of passengers for the entire attention area AR, and of that, "29" in the passenger count display 222 indicates the number of passengers at the pinpoint boarding position "Shinagawa Station Takanawa Exit Taxi Stand" of the pinpoint boarding position mark 221. In this way, by displaying the pinpoint boarding position and the predicted number of passengers there in addition to the number of passengers for the attention area AR, the actual vehicle rate can be increased.

[0141] The pinpoint boarding location can be estimated by learning using actual vehicle data, rather than investigating each and every location within the area AR where a boarding location is determined.

[0142] Specifically, as shown by the black circles on the left side of FIG. 16, the passenger's past boarding locations can be identified from the information on the departure points of the actual vehicle data. By learning the passenger's past boarding locations, the estimated value of the boarding location shown by the black circles and the probability (likelihood) of that boarding location are calculated, as shown on the right side of FIG. 16. The probability of the boarding location is expressed as a number ranging from 0 to 1, and is displayed near the boarding location in FIG. 16. For example, the demand forecasting app can display the estimated value of a boarding location whose probability is equal to or greater than a predetermined threshold (e.g., 0.8) as a pinpoint boarding location within the area of ​​interest AR.

[0143] <10. Display of estimated waiting time> FIG. 17 shows a fourth display example of the demand forecast screen displayed by the demand forecast application.

[0144] FIG. 17 also shows a display example in which a more detailed prediction result is displayed when the driver selects a predetermined area AR as the area of ​​interest AR.

[0145] In places where the boarding locations are fixed and the number of passengers is high, such as taxi stands in front of stations or hotels, there is a method known as "waiting on the taxis," in which passengers are obtained by waiting in a line of taxis 11 that will pick up passengers at the boarding location. The disadvantage of waiting on the taxis is that, for example, if there is a long line of taxis 11 at the taxi stand, it takes time for a taxi 11 to line up at the end of the long line and pick up a passenger.

[0146] Therefore, the demand forecasting application can display the time required for waiting at a boarding location with a high number of passengers (pinpoint boarding location), in other words, the time required to wait at the boarding location until passengers can be picked up.

[0147] Specifically, as shown in Fig. 17, when a pinpoint boarding position mark 221 in the area of ​​interest AR is a place for waiting, the demand prediction application displays a waiting start button 223 within a boarding number display 222 at the pinpoint boarding position mark 221. When the waiting start button 223 is tapped (touched), the demand prediction application displays a waiting display 224 indicating the time required for waiting (waiting time for waiting) if waiting for a taxi is performed. In the example of Fig. 17, "20 minutes" is displayed as the waiting time for waiting for a taxi.

[0148] For example, the driver can check the waiting time for pickup and select a location to wait for pickup by displaying the pickup waiting display 224 at the pinpoint pickup location. The pickup waiting time displayed on the pickup waiting display 224 may be a value with a certain range, such as "15 minutes to 20 minutes."

[0149] In the vehicle movement log data, it is possible to detect the boarding change point when the status changes from "empty" to "occupied" and the state in which the taxi 11 is moving slowly shortly before the boarding change point, and therefore it is possible to detect the waiting behavior of the taxi 11. For example, driving at a speed below a predetermined speed (5 km / h or less) within a predetermined period or distance before the time of the boarding change point can be detected as the waiting behavior. Therefore, by learning the waiting behavior, it is possible to predict the waiting time at a predetermined boarding location.

[0150] <11. Display of predicted length> FIG. 18 shows a fifth display example of the demand forecast screen displayed by the demand forecast application.

[0151] FIG. 18 also shows a display example in which a more detailed prediction result is displayed when the driver selects a predetermined area AR as the area of ​​interest AR.

[0152] Among the area ARs obtained by dividing the business area into predetermined units, there are area ARs and pickup locations where the proportion of pickups with long distances (longer than a predetermined distance) is high, for example, when the destination is Haneda Airport or Narita Airport. It is preferable for the driver to be able to determine the possibility of a long-distance passenger.

[0153] Therefore, the demand forecasting application can perform a long display 241 that displays the proportion of long-distance passengers in the attention area AR separately from the number of passengers in the entire attention area AR, as shown in FIG.

[0154] In the long display 241, the proportion (ratio) of rides with long ride distances out of the total number of rides in the area of ​​interest AR is displayed as the long degree. In addition, in the long display 241, the ride distance is divided into multiple segments, and the proportion of rides for each segment is displayed as a bar graph as a long segment. The "All" bar graph shown in FIG. 18 shows the proportion of rides for each segment in the entire operating area, and the "This" bar graph shows the proportion of rides for each segment in the area of ​​interest AR.

[0155] The long display 241 may display the longness and long category for the area of ​​interest AR, as shown in Figure 18, or may be displayed in conjunction with the pinpoint boarding position mark 221 to display the longness and long category for the pinpoint boarding position.

[0156] The bar graph of the long display 241 in Figure 18 divides the travel distance into multiple sections as long sections and shows the proportion of passengers for each divided section (travel distance), but it is also possible to divide the fare into multiple sections and show the proportion of passengers for each divided section (fare).

[0157] Furthermore, the long distance display 241 may predict the demand for riding for each time period and weather, and display the long distance degree and long distance category specialized for a predetermined time period and weather.

[0158] The degree of longness and long distance classification can be predicted by learning from actual vehicle data including the travel distance and fare items.

[0159] <12. Display of predicted riding distance> FIG. 19 shows a sixth display example of the demand forecast screen displayed by the demand forecast application.

[0160] FIG. 19 also shows a display example in which a more detailed prediction result is displayed when the driver selects a predetermined area AR as the area of ​​interest AR.

[0161] When a predetermined area AR is selected as the area of ​​interest AR, the demand forecasting application can perform a travel distance display 251 that displays the average travel distance of passengers boarding in the area of ​​interest AR and its confidence interval, as shown in Fig. 19. The confidence interval represents the range within which the average value of the population (population mean) is included with a predetermined degree of reliability.

[0162] The ride distance display 251 shows that the average ride distance in the area of ​​interest AR is "2.4 km," and that the confidence interval of the average ride distance at a confidence level of, for example, 95% is "1.1 km to 3.7 km." The confidence level of the confidence interval is not limited to 95% and can be set arbitrarily, such as 99%.

[0163] In this way, by displaying the average driving distance and its confidence interval of the area AR of interest, the driver can, for example, search for an area AR with a driving distance that is suitable for the remaining working hours, or search for an area AR with a long driving distance as a ``quick run'' route.

[0164] The ride distance display 251 may be displayed for the area of ​​interest AR as shown in Figure 19, or may be displayed in conjunction with the pinpoint ride location mark 221 to display the average ride distance and confidence interval for the pinpoint ride location.

[0165] Alternatively, instead of the average trip distance and confidence interval, the average trip fare and confidence interval may be shown.

[0166] Alternatively, instead of the average trip distance and confidence interval, the average trip time and confidence interval may be shown.

[0167] In addition, the ride distance display 251 may predict ride demand for each time period and weather, and display the average ride distance and confidence interval, the average ride fare and confidence interval, or the average ride time and confidence interval specialized for a specific time period and weather.

[0168] <13. Display of predicted fare> Some users refrain from using Taxi 11 because the fare is not confirmed until the ride. The demand forecasting app has a function that displays a predicted fare based on the current location and destination.

[0169] FIG. 20 shows an example of a fare prediction screen displayed by the demand prediction application.

[0170] The demand forecasting app displays the time and cost required to travel to the destination as a predicted result on a display, and also displays the time and cost required to travel for each division into specified units into which the travel route to the destination is divided as a predicted result on a display.

[0171] 20, individual display 261 shows the time and fare required for travelling for each division, and destination display 262 shows the time and fare required for travelling to the destination.

[0172] Learning the fare and travel time is difficult if the actual vehicle data shown in Figure 5 is used as is, because the same data is required for both the departure point and the arrival point. Therefore, the control unit 121 learns the time and fare required for travel for each division unit from the vehicle movement log data. The control unit 121 then calculates the time and fare required for travel to the destination by calculating the sum of the time and fare for each division unit included from the departure point to the arrival point. The division unit can be, for example, a unit divided using at least one or more of a predetermined distance, a predetermined time, a unit divided by road sections (blocks), a unit divided by traffic lights or intersections, etc.

[0173] The travel time and fare for each division unit, and the travel time and fare to the destination may be displayed with a predetermined range, such as "5 minutes - 10 minutes" or "300 yen - 500 yen."

[0174] 13 to 20, the drivers of the taxis 11 can operate more efficiently. That is, the demand prediction application can present prediction results that contribute to improving occupancy rates.

[0175] The various displays described with reference to Figures 13 to 20 can be appropriately set by the driver to turn the display on or off, set the display order, etc. on the setting screen displayed on the display by operating the setting button 64 of the demand forecasting app.

[0176] <14. Learning the boarding position> Next, learning and prediction of things other than passenger demands performed by the server 12 will be described.

[0177] FIG. 21 is a diagram for explaining learning of a boarding position relative to a building.

[0178] For example, a user (customer) of taxi 11 arranges taxi 11 from a predetermined location in building 271 using a taxi dispatch application 272 executed on a terminal such as a smartphone, and gets into taxi 11 at a predetermined location 273 such as the driveway of building 271. In this case, the taxi dispatch application 272 acquires the user's location information at the time taxi 11 is arranged from a GPS receiver in the terminal, and transmits this to server 12 as location information at the time of requesting taxi dispatch. In addition, boarding location information, which is the user's location information at the time the user gets into taxi 11, can be acquired from vehicle movement log data transmitted from vehicle management device 22 of taxi 11.

[0179] The server 12 learns the relationship between the location information at the time of the dispatch request and the location information at the time of boarding. As a result, when a user arranges a taxi 11 from a predetermined position in the building 271, the driver can learn the boarding position in the building 271, that is, where in the building 271 the taxi 11 should arrive. The server 12 stores the learning result as a boarding position list in the memory unit 122. The demand prediction application can display the learned boarding positions in the building 271 on the map 41. Furthermore, when a user specifies the building 271 as a destination, the driver can set the learned boarding position in the building 271 as the drop-off position.

[0180] In addition, the server 12 can infer the boarding location of buildings other than the building 271 where the taxi 11 was actually arranged, based on the learned relationship between the location information at the time of the dispatch request and the location information at the time of boarding, and display this on the map 41.

[0181] <15. Learning where to get off> FIG. 21 is a diagram for explaining learning of drop-off positions for buildings.

[0182] For example, a user gets off the taxi 11 at a predetermined location 274 and moves to a predetermined building 271, which is the destination. Location information at the time of getting off the taxi 11, which is the location information of the user at the time the user gets off the taxi 11, can be acquired from vehicle movement log data transmitted from the vehicle management device 22 of the taxi 11. In addition, the taxi dispatch app 272 acquires location information of the building 271 to which the user moved after getting off the taxi 11 from the GPS receiver in the terminal, and transmits this to the server 12 as post-movement location information.

[0183] The server 12 learns the relationship between the drop-off location information and the post-movement location information. As a result, when a user specifies the building 271 as a destination, the driver can learn the drop-off locations in the building 271 where to drop off the user. The server 12 stores the learning results as a drop-off location list in the memory unit 122. The demand prediction application can display the learned drop-off locations in the building 271 on the map 41. Furthermore, when a user uses the taxi dispatch application 272 to order a taxi 11 from a predetermined location in the building 271, the driver can also use the learned drop-off locations in the building 271 as a pickup location.

[0184] <16. Learning the boarding position> In the examples described with reference to Figures 21 and 22, it was explained that the learned boarding location is also displayed as the disembarking location, and the learned disembarking location is also displayed as the boarding location, but generally, the boarding location and disembarking location for a building are often places such as taxi pools, driveways, and entrances, and are often the same or nearby.

[0185] 23 , the server 12 learns the boarding position information and the disembarking position information, learns the optimal boarding position for the building 271, and stores the learned boarding position list in the storage unit 122. The demand prediction application can display the learned boarding positions for the building 271 on the map 41.

[0186] <17. Audio guide to passenger demand> Next, a description will be given of the audio guidance of the predicted results of the passenger demand, which is executed by the demand prediction application.

[0187] For safety reasons, the driver of taxi 11 cannot see the demand forecast screen displayed by the demand forecast app while driving. Therefore, the demand forecast app not only displays the results of predicted demand for rides on a map on the display, but also has a function to notify the driver of the predicted demand for rides by sound.

[0188] Since the demand forecasting app cannot notify by sound all of the forecast results displayed on the demand forecasting screen, it notifies by sound the forecast results corresponding to the current location of taxi 11 (hereinafter also referred to as the vehicle position) and the direction of travel of taxi 11.

[0189] In addition, the demand forecast application switches the display method of the demand forecast and the sound notification method depending on the scale of the map 41 that displays the demand forecast.

[0190] Below, we will explain the demand forecast display and sound output when the scale of the map 41 displaying the demand forecast is high, in other words, a wide-area map display, and the demand forecast display and sound output when the scale of the map 41 is low, in other words, a detailed map display.

[0191] Note that sound notifications from the demand forecasting app include both non-verbal sounds (also known as sound effects or electronic sounds) such as "beep, beep, beep," "bong," and "ping pong," as well as voice notifications (messages) consisting of words or sentences. However, for simplicity, in the following explanation, sound effects will simply be referred to as "sound," and verbal sound output will be referred to as "voice."

[0192] <Example of sound notification when displaying a wide-area map> FIG. 24 shows an example of a demand forecast screen in which the scale of the map 41 displaying the demand forecast is high, in other words, a wide-area map display.

[0193] 24, it is assumed that area 411 is an area AR predicted to have high demand for riding among the areas AR of the demand forecast mesh 63 superimposed on the map 41. In FIG. 24, the display of colors and intensities according to the degree of demand for riding for each area AR other than area 411 is omitted.

[0194] The taxi 11 is traveling at the location of the vehicle position mark 421, and in the direction of travel on National Route 1, there is an area 411 that is predicted to be a place with high demand for rides.

[0195] The demand prediction app detects that there is an area 411 with high demand for boarding in the direction of travel, and notifies the driver by sound or voice that there is an area 411 with high demand for boarding in the direction of travel. For example, the demand prediction app outputs by voice, "There is an area with high demand nearby in the direction of travel," and outputs three consecutive "beep beep" sounds.

[0196] Also, when taxi 11 is traveling on National Route 1 and is located at the location of vehicle position mark 422, the demand prediction application outputs, for example, a voice message saying, "High demand in the direction of travel," and outputs four consecutive beeps.

[0197] In this way, if there is an area 411 with high demand for boarding in the direction of travel, the demand prediction application notifies the driver of this fact by sound or voice at a predetermined timing. The timing of the notification can be set, for example, for each area AR of the demand prediction mesh 63. In this case, a notification by sound or voice is output every time the area AR in which the traveling taxi 11 is located changes. Alternatively, the demand prediction application may be set to notify at predetermined time intervals (for example, every one minute) or at predetermined distances (for example, every one kilometer). The notification timing can be changed on a setting screen that is displayed by operating the setting button 64.

[0198] The demand forecasting app changes the content of the voice message and the type of sound output depending on the distance (proximity) to the area AR with high passenger demand detected in the direction of travel.

[0199] In the above example, when the vehicle position is located at the vehicle position mark 421, which indicates that the vehicle is far from the high-demand area, the demand prediction application outputs a voice message saying, "There is a high-demand area nearby in your direction," followed by three consecutive beeps. When the vehicle is closer to the high-demand area, the demand prediction application outputs a voice message saying, "There is high demand nearby in your direction," followed by four consecutive beeps. That is, the voice message changes from "There is a high-demand area nearby in your direction" to "There is high demand in your direction." The sound message changes from three consecutive beeps to four consecutive beeps, with the number of consecutive beeps increasing as the vehicle approaches the high-demand area. In addition to increasing the number of consecutive sound effects, the duration of the sound may be changed, such as beeps, beeps, and beeps, or the volume may be gradually increased. A combination of at least two of the number of sounds, the duration, and the volume may also be used.

[0200] For example, if there is an intersection in the direction of travel and the high demand area is in the direction of turning right at the intersection, the demand forecasting app can output a voice message such as, "Turn right at the intersection; there is high demand."

[0201] The degree of ride demand at which the demand forecasting app notifies high-demand areas by sound or voice can also be changed on the setting screen displayed by operating the setting button 64. For example, as shown in Fig. 2, each area AR in the demand forecasting mesh 63 is displayed in a color or intensity corresponding to the degree of ride demand, but if the degree of ride demand distinguished by color or intensity on the demand forecasting screen in Fig. 2 is in five levels from level 1 to level 5, it is possible to set the app to notify when an area AR with the highest degree of ride demand of level 5 is located in the direction of travel, or to notify when an area AR with level 4 or higher is located in the direction of travel.

[0202] The driver can also set on the settings screen whether to provide guidance about high-demand areas by sound only, by voice only, or by both sound and voice. The demand forecasting app has a notification function (notification unit) that notifies the driver of specified areas AR where demand for rides is high by at least one of sound and voice.

[0203] In Figure 18, we explained about the long display 241, which displays the proportion (ratio) of long-distance rides as the longness degree when there is an area AR or boarding location where the proportion of long-distance rides is high.

[0204] When the demand forecasting app is displaying a wide-area map of the demand forecast screen on the display, it can also notify the driver of areas with high long-distance AR and the boarding location using sound or voice.

[0205] For example, if an area AR with a longness level equal to or greater than a predetermined value is present in the direction of travel, the demand forecasting app will output a voice message saying, "There is a long area nearby," along with a "beep" sound, which is a different type of sound from the notification of a high demand area.

[0206] <Example of sound notification when displaying a detailed map> FIG. 25 shows an example of a demand forecast screen in which the scale of the map 41 displaying the demand forecast is low, in other words, a detailed map display.

[0207] In the detailed map display of FIG. 25, the taxi 11 is traveling at the location of the vehicle position mark 441.

[0208] In the detailed map display, as a result of learning using actual vehicle sequence data, each boarding location within the display area where boarding demand exceeds a predetermined level is displayed as a circular demand point 451. Note that in Fig. 25, some of the reference symbols for the demand points 451 have been omitted to avoid cluttering the diagram.

[0209] For example, of the boarding locations learned within the display area shown on the display, the top 30 locations with the highest boarding demand are displayed as demand points 451. However, even if a predetermined boarding location is included in the top 30, it is excluded if it is selected because the total number of boarding locations is small. Therefore, the boarding locations displayed as demand points 451 are boarding locations whose predicted boarding demand is at least equal to or greater than a predetermined first level ThA and that are included in the top 30.

[0210] In the example of Figure 25, the demand points 451 include demand point 451A displayed as a dark circle, demand point 451B displayed as a light circle, and demand point 451C displayed with a pattern different from demand points 451A and 451B.

[0211] The difference in density between demand point 451A and demand point 451B represents the degree of predicted demand for rides. That is, when the predicted demand for rides is equal to or higher than the second level ThB, which indicates high demand for rides, demand point 451A is displayed with a high density, and when it is equal to or higher than the first level ThA but lower than the second level ThB, demand point 451B is displayed with a low density. In this way, by changing the density of demand point 451 according to the magnitude of demand for rides, for example, when demand point 451 is displayed alone without overlapping with other demand points 451, drivers can recognize the difference in demand for rides based on the difference in density. Furthermore, when demand point 451 is displayed overlapping with other demand points 451, the density of demand point 451 appears high, so that drivers can recognize a location where demand for rides is concentrated as a location with high demand for rides. Note that instead of distinguishing between demand point 451A and demand point 451B using two types of density, density may be changed continuously according to the degree of demand for rides.

[0212] The demand points 451C displayed with a different pattern represent the demand points 451 with a longness degree equal to or greater than a predetermined value among the demand points 451 selected as the top 30 with the highest demand for boarding. This allows the driver to recognize the demand points 451 with a high longness degree on the detailed map display.

[0213] In Figure 25, due to drawing constraints, demand point 451C, which has a high degree of longness, and the other demand points 451A and 451B are displayed with different patterns, but on a display capable of displaying in color, demand point 451C can be distinguished from demand points 451A and 451B by using different colors.

[0214] In the above example, the top 30 boarding locations within the display area of ​​the detailed map display are displayed as demand points 451, but the number of demand points 451 to be displayed can be changed on the setting screen. For example, on the setting screen, it is possible to select from the top 30, top 20, or top 10.

[0215] In the above example, the top 30 boarding locations are selected and displayed within the display area of ​​the detailed map display, but the top 30 may be selected and displayed in area AR units of the demand forecast mesh 63. In other words, the extraction unit for extracting the demand points 451 can be set appropriately.

[0216] Next, guidance by sounds and voice in a detailed map display such as that shown in FIG. 25 will be described.

[0217] The taxi 11 is traveling in a predetermined direction (for example, toward Shinagawa Station) from the location of the vehicle position mark 441. If a demand point 451 is present within a predetermined distance in the traveling direction, the demand prediction application outputs a voice message saying, "High demand in the traveling direction." The demand prediction application also outputs a beep every time the taxi 11 passes a demand point 451. One sound is output for each demand point 451.

[0218] In the detailed map display, a sound is output when a demand point 451 passes, so a "beep" sound is generated continuously on roads where many demand points 451 are displayed. This allows the driver to recognize that the road they are currently traveling on is a road with high demand for rides, contributing to improved recognition of area ARs and roads with high demand for rides, and ultimately contributing to an increase in occupancy rates. The volume of the "beep" sound may be changed and output depending on the degree of demand for rides at each demand point 451.

[0219] Unlike the example in Fig. 25, when there is no demand point 451 in the traveling direction of the taxi 11 and there are many demand points 451 in the opposite direction to the traveling direction, the demand prediction app outputs a voice message such as, "There is a high demand point in the opposite direction." Also, when there is an intersection in the traveling direction and the road with many demand points 451 is in the direction of turning right at the intersection, the demand prediction app outputs a voice message such as, "There is high demand in the direction of turning right at the intersection."

[0220] FIG. 26 shows another example of a detailed map display.

[0221] In the detailed map display of FIG. 26, part of the demand point 451 displayed in the detailed map display of FIG. 25 is replaced with demand points 451D to 451F.

[0222] As shown in Fig. 4, the vehicle movement log data also records the traveling direction of the taxi 11, and therefore the traveling direction of the taxi 11 boarded at the demand point 451 is also learned. The traveling direction of the taxi 11 at the demand point 451 is learned as eight directions, for example, similar to the arrows 212-1 to 212-8 in Fig. 14. If there is a traveling direction that accounts for a ratio of 50% or more of the eight traveling directions of the taxi 11 boarded at each demand point 451, the demand prediction application can display that direction (hereinafter referred to as the dominant direction).

[0223] Demand point 451D represents demand point 451 that indicates a dominant direction. Demand point 451E is displayed when a predetermined number or more of demand points 451D having the same dominant direction exist within a predetermined range. In other words, demand point 451E represents a set of a predetermined number or more of demand points 451D.

[0224] Demand point 451D is displayed, for example, as shown in Fig. 26, by a mountain-shaped (V-shaped) symbol, and the direction of the corner indicates the dominant direction. Demand point 451E is displayed by a symbol that is an enlarged version of the symbol for demand point 451D. Demand points 451D and 451E may also be displayed by other symbols that can indicate direction, such as arrow symbols.

[0225] Demand point 451F represents demand point 451 that has a dominant direction and a high degree of longness. In other words, when demand point 451D that has a dominant direction is demand point 451 that has a high degree of longness, like demand point 451C in Fig. 25, it is displayed with a different pattern or color, like demand point 451F, to simultaneously represent that it is demand point 451 that has a high degree of longness. Similarly, when demand point 451E has a high degree of longness, it is displayed with a different pattern or color.

[0226] If there is a dominant direction at each demand point 451 on the detailed map display, displaying that dominant direction can help the driver decide which direction to take when, for example, "driving," which can contribute to improving the occupancy rate. Also, even at intersections where various directions are possible, the driver can decide which direction to take with the highest demand by referring to the dominant direction around the intersection.

[0227] The notification method using sounds and voices in the detailed map display in FIG. 26 is the same as in FIG. 25, so a description thereof will be omitted.

[0228] Even in the detailed map display, whether the guidance for the demand points 451 is given by sound only, by voice only, or by both sound and voice can be changed according to the setting value on the setting screen. The demand prediction application has a notification function (notification unit) that notifies the driver of predetermined demand points 451, which are locations with high demand for rides, by at least one of sound and voice.

[0229] As described above, the demand prediction application can contribute to improving the occupancy rate by notifying the driver of the predicted results of passenger demand for the direction of travel of the taxi 11 by sound or voice.

[0230] The way in which ride demand is displayed differs depending on whether the demand forecast screen is a wide-area map display (Figure 24) or a detailed map display (Figures 25 and 26), and accordingly, the sound and voice notification methods also differ.

[0231] The driver can switch between the wide-area map display and the detailed map display, for example, by operating the zoom button 62 or by changing the settings on the setting screen. Alternatively, instead of operating the touch panel, the demand prediction application may switch by recognizing (by voice recognition) a voice instruction for "wide-area display" or "detailed display" issued by the driver. Furthermore, the demand prediction application may automatically switch. For example, the demand prediction application can change to the detailed map display when it is determined that the taxi 11's own position has entered a high-demand area guided by the wide-area map display, and change to the wide-area map display when it is determined that the taxi 11 has left the high-demand area.

[0232] <Voice guidance control processing> The above-mentioned audio and voice guidance of the predicted demand for rides is necessary when the taxi 11 is empty and has no passengers on board, but is not necessary when the taxi is carrying a passenger. Furthermore, audio and voice notifications are a distraction to passengers. Therefore, the demand prediction application can obtain the status of "occupied" or "empty" detected by the fare meter 21 and control the on / off of audio and voice guidance in conjunction with the status.

[0233] FIG. 27 is a flowchart of a voice guidance control process for controlling voice guidance.

[0234] First, in step S51, the demand prediction application determines whether the status of the taxi 11 (vehicle) has been changed to "actual vehicle."

[0235] If it is determined in step S51 that the status of the taxi 11 has changed to "actual vehicle," the process proceeds to step S52, where the demand prediction application controls to turn off the voice guidance. After step S52, the process returns to step S51.

[0236] On the other hand, if it is determined in step S51 that the status of taxi 11 has not changed to "occupied vehicle", the processing proceeds to step S53, and the demand forecasting application determines whether the status of taxi 11 has changed to "vacant vehicle".

[0237] If it is determined in step S53 that the status of the taxi 11 has changed to "vacant," the process proceeds to step S54, where the demand prediction application controls to turn on the voice guidance. After step S54, the process returns to step S51.

[0238] On the other hand, if it is determined in step S53 that the status of the taxi 11 has not been changed to "vacant", the process also returns to step S51.

[0239] The voice guidance control process in FIG. 27 is started when the demand prediction application is started or when the voice guidance is set to ON on the setting screen, and is repeated until the demand prediction application is terminated.

[0240] As described above, the demand prediction application can control the on / off of voice guidance in conjunction with the status of "occupied vehicle" or "vacant vehicle" detected by the fare meter 21. By controlling the status of "occupied vehicle" or "vacant vehicle", voice guidance can be automatically provided (without driver operation) only when necessary by the driver.

[0241] 27 is an example of a voice guidance notification control process that outputs boarding demand in the direction of travel as a voice message, but notifications by sound and notifications by both sound and voice are also controlled in the same way. The status of "occupied vehicle" or "vacant vehicle" may be obtained by the demand prediction application directly from the fare meter 21 or indirectly via the vehicle management device 22.

[0242] <Other audio outputs> In addition to notifying drivers of the above-mentioned ride demand through sound and voice, the demand forecasting app can also notify drivers of the following information through voice:

[0243] For example, the demand prediction application can obtain train service suspension and resumption information in real time from the server 12 and output it as a voice message. This allows drivers to quickly move to areas where demand for trains is increasing according to the train service information.

[0244] For example, the demand prediction application can obtain the end and start timings of events being held in the business area of ​​the taxi 11 from the server 12 in real time and output the timings as voice (messages). This allows the driver to quickly move to an area where demand for rides increases in response to the end or start of an event.

[0245] The demand prediction application can, for example, obtain weather information in the operating area of ​​the taxi 11, particularly information about sudden weather changes, from the server 12 in real time and output it as a voice message. This allows the driver to quickly move to an area where demand for rides increases in response to weather changes.

[0246] The server 12 acquires train operation information, event information, weather information, and the like from the servers of affiliated information providers, and transmits the information to the demand forecasting application of each terminal device 23.

[0247] <18. Recommended Route Presentation Processing> In the above-described embodiment, an example was described in which the driver himself determines the direction of travel of the taxi 11 and drives the taxi 11 based on the ride demand forecast displayed on the display of the terminal device 23 and the audio output by the demand forecast application.

[0248] The following describes a recommended route presentation function in which the demand prediction application searches for routes with high demand for rides based on ride demand prediction data transmitted from the server 12 and presents recommended routes.

[0249] The demand prediction application has a navigation function (navigation processing unit) that searches for a route to a destination set by any method based on the vehicle's position. In addition to the functions of a general navigation system installed in a vehicle (hereinafter referred to as a car navigation system), the demand prediction application also has a function that searches for a route to a destination taking into account locations predicted to have high demand for rides. The destination may be a predetermined pinpoint location, or may be a destination area such as one or more areas AR divided by a demand prediction mesh 63. In the following description, a location predicted to have high demand for rides will be referred to as a ride point. A ride point is the same as the location referred to as a ride position or a demand point in the above description.

[0250] The recommended route presentation function of the demand forecasting app is used when the taxi 11 is empty, i.e., when it is "cruising." Using the recommended route presentation function allows the taxi 11 to catch passengers quickly and shorten the time the taxi is empty and the driving distance while empty. This function is particularly useful for new drivers who have only recently started working as a taxi driver because they do not have a specific area (region) where they are good at driving and do not have knowledge of areas with high passenger demand. Even veteran drivers who have been working as a taxi driver for a long time and drivers who achieve above-average sales have areas where they are good at driving and areas where they are not good at driving. Therefore, when driving in an area where they are not good at driving, the taxi 11 can catch passengers quickly and shorten the time the taxi is empty and the driving distance while empty. By setting the destination in a good area or location during route search, the taxi 11 can return to the desired area or location while passing through a route with high demand (a pickup point) more frequently. By using the recommended route presentation function to repeatedly pass through a route with high demand (a pickup point), the taxi 11 can overcome its weak areas and expand its strong areas.

[0251] The preferred area may be registered (set) in advance by the driver on a setting screen or the like, or may be determined (automatically registered) by the demand forecasting application using the driver's vehicle movement log data and actual vehicle data. Preferred locations may be, for example, the driver's office, pinpoint pick-up locations frequently used by the driver, or locations where "waiting" is performed, such as taxi stands frequently used by the driver.

[0252] When a demand forecasting application uses data to determine areas where drivers are good at driving, the application may create a heat map, for example, as shown in FIG. 28, for sales regions distributed on an XY plane, with the Z-axis representing cumulative time. Areas with a time value equal to or greater than a predetermined threshold, for example, 70% of the total time, may be extracted as areas where the driver is good at driving. The areas where the driver is good at driving may be all of the areas or more extracted as being equal to or greater than the threshold, or the area with the longest cumulative time, or multiple areas with the highest cumulative time. The display unit for the areas where the drivers are good at driving may be the extracted area itself, or, to make it easier for the driver to recognize, the display unit may be the city, ward, town, or village with the largest area in the extracted area, or the city, ward, town, or village with the point with the longest cumulative time.

[0253] When the original data for calculating the cumulative time in the Z-axis direction is vehicle movement log data, the areas where the driver spent most of their time traveling can be determined as their specialty areas. Also, when the original data for calculating the cumulative time in the Z-axis direction is only vehicle movement log data with a status of "actual vehicle," the areas where the driver spent most of their time carrying passengers can be determined as their specialty areas. When the original data for calculating the cumulative time in the Z-axis direction is actual vehicle data, the areas where the driver spent most of their time carrying passengers can be determined as their specialty areas.

[0254] FIG. 29 shows an example of a recommended route presentation screen displayed when the recommended route presentation function is executed in the demand forecasting application.

[0255] The recommended route presentation screen of Figure 29 is displayed after a recommended route is searched for, for example, by the driver tapping a recommended route search button or the like displayed on the demand forecast screen of Figure 2 or the like.

[0256] The recommended route presentation screen in Fig. 29 displays a map 41 on which a demand forecast mesh 63 is superimposed. As described with reference to Fig. 2, each area AR of the demand forecast mesh 63 is displayed classified by color or intensity according to the degree of ride demand, but in the example of Fig. 29, the display by color or intensity according to the degree of ride demand is omitted to make the diagram easier to read. The same applies to Figs. 32 to 44 described later, in which the display by color or intensity according to the degree of ride demand is omitted.

[0257] A route search result 501 for a recommended route is displayed on the map 41. Also displayed on the map 41 are a detail button 511 that is operated to reduce the scale of the map 41, a wide area button 512 that is operated to increase the scale of the map 41, and a full screen display button 513 that is operated to switch to full screen display.

[0258] A predicted time display area 502 is located below the map 41. The predicted time display area 502 displays the predicted time of demand forecast when a recommended route is searched for, and also allows the predicted time of demand forecast to be changed, similar to the predicted time setting area 42 in FIG.

[0259] A recommended route information presentation section 503 is disposed to the right of the map 41 and the predicted time display section 502. The words "recommended route" are displayed at the top of the recommended route information presentation section 503, indicating that this is a recommended route presentation screen.

[0260] The recommended route information presentation section 503 displays "<Destination mode>", indicating that the route search result 501 displayed on the map 41 was searched in the "Destination mode" among multiple search modes. The demand forecasting application has three search modes for route search that take demand points into consideration: "Destination mode", "Immediate pickup mode", and "Nearby pickup point mode".

[0261] The "destination mode" is a mode in which a destination is set and a route that passes through a demand point is searched for from among multiple routes to the destination. The destination may be set by the driver, or the driver's specialty area may be set as the destination (destination area). For example, the "destination mode" is suitable when the current location of the taxi 11 is outside the specialty area and the driver wants to catch a passenger while returning to the specialty area.

[0262] The "immediate pickup mode" does not have a specific destination, but searches for a route from the current location that passes through the demand point. The "immediate pickup mode" can be used when the current location of the taxi 11 is within its specialty area.

[0263] The "nearby pick-up point mode" is a mode in which there is no specific destination and a route is searched for that passes through demand points with high evaluation values ​​around the current location. The "nearby pick-up point mode" can be used whether the current location of taxi 11 is within or outside the preferred area.

[0264] Below "<Destination Mode>" in the recommended route information presentation section 503, "Return to area "Setagaya Ward"" is displayed, indicating that the destination of the route search in "Destination Mode" is Setagaya Ward.

[0265] Furthermore, the recommended route information presentation unit 503 displays information such as "recommended level: 98 points," "boarding points: 108 locations," and "total demand forecast level: 6 levels." "Recommended level: 98 points" indicates that the recommended level of the route search result 501 displayed on the map 41 is 98 points. The recommended level may be a score out of 100, or a score display in which a higher number indicates a higher degree of recommendation. "Boarding points: 108 locations" indicates that the route of the route search result 501 displayed on the map 41 passes through 108 boarding points. In the demand forecast screen of FIG. 2, an example was described in which the degree of boarding demand for each area AR, distinguished by color or intensity, is displayed in five levels, from level 1 to level 5. However, "total demand forecast level: 6 levels" indicates that the total boarding demand levels of one or more areas AR passed by the route search result 501 displayed on the map 41 is 6 levels.

[0266] A recommended route presenting process for searching for a recommended route that takes demand points into consideration and presenting the route to the driver will be described with reference to the flowchart of FIG.

[0267] This process is started by, for example, tapping a recommended route search button or the like displayed on the demand forecast screen such as in Fig. 2. The ride demand forecast data used in the recommended route presentation process can be data periodically transmitted from the server 12, or can be obtained by requesting the server 12 as needed.

[0268] First, in step S71, the demand forecasting application determines whether a search mode has been specified. The search mode can be specified, for example, by selecting one of the search mode buttons, "destination mode," "immediate pickup mode," or "nearby pickup point mode," on a mode selection screen that is displayed after pressing a route search button. Alternatively, the application may be set in advance so that "destination mode" is selected when the vehicle is located outside the preferred area, and "immediate pickup mode" or "nearby pickup point mode" is selected when the vehicle is located within the preferred area.

[0269] The process of step S71 is repeated until it is determined that the search mode has been designated, and if it is determined that the search mode has been designated, the process proceeds to step S72.

[0270] Then, in step S72, the demand prediction application determines whether the specified search mode is the "destination mode," the "immediate boarding mode," or the "nearby boarding point mode."

[0271] If it is determined in step S72 that the specified search mode is the "destination mode," the process proceeds to step S73, and steps S73 to S75 are executed. If it is determined that the specified search mode is the "immediate pickup mode," the process proceeds to step S76, and steps S76 and S77 are executed. If it is determined that the specified search mode is the "nearby pickup point mode," the process proceeds to step S78, and steps S78 to S80 are executed.

[0272] If it is determined that the specified search mode is the "destination mode," the demand forecasting application sets the destination (destination) to the specialty area in step S73. The destination is set to a pre-registered specialty area by default, but can be changed by the driver's operation.

[0273] In step S74, the demand forecasting application searches for multiple (predetermined number of) routes based on the vehicle position and destination using a route search algorithm such as Dijkstra's algorithm or A* algorithm. This process is similar to the function of a general car navigation system.

[0274] In step S75, the demand forecasting application calculates the total score SUMscore of each of the searched routes.

[0275] The total route score SUMscore can be calculated by, for example, calculating score Sc = area AR passenger demand level [1-5] × boarding point [0,1] × direction agreement [cosθ] for each area AR that the searched route passes through, and then summing the scores Sc for each area AR from the vehicle's location to the destination. The area AR's passenger demand level [1-5] represents the degree of passenger demand for that area AR and is a value between level 1 and 5. The boarding point [0,1] is set to "1" if there is a boarding point in that area AR and "0" if there is not. The direction agreement [cosθ] is the angle (cosθ) between the vehicle's location and the direction from the vehicle's location to the area AR relative to the direction from the vehicle's location to the destination, and the closer the direction is to 1, the closer the value is to 1. For the demand forecast used in the calculation formula for the score Sc described above, the predicted time for the area AR is sequentially updated depending on the distance from the vehicle's location, for example, the number of times the area AR is passed through. For example, if one area AR is 500m x 500m and the bus travels at 30km / h for 10 minutes, it will travel 5km in 10 minutes, so the predicted time for the area AR is advanced by 10 minutes for every 10 areas traveled, and the demand forecast for the route is updated. Note that the boarding point [0,1] in the calculation formula for the score Sc described above may be replaced by the number of boarding points present in that area AR.

[0276] The total score SUMscore of the route may be calculated not in units of areas AR as described above, but for each boarding point on the route. Also, if the direction of a part of the searched route is the opposite direction to the destination, the score Sc may not be added or may be subtracted.

[0277] The total score SUMscore of the route is not limited to the above example, and any calculation method can be used that will result in a larger total score SUMscore for a route with a high level of demand forecast and many boarding points.

[0278] In addition, other aspects may be added to the route total score SUMscore to calculate the score Sc or the total score SUMscore. For example, a coefficient or score Sc according to the length of the boarding point may be added, so that the longer the boarding point, the higher the total score SUMscore or score Sc. Alternatively, a coefficient or score Sc may be added depending on whether the route passed through on the searched route is a main road or a narrow road. The direction of travel of passengers when boarding at each boarding point can be identified from the departure and arrival points of past boarding history (actual vehicle data). The closer the driver's destination and the direction of travel of the boarding point are, the higher the total score SUMscore or score Sc may be.

[0279] After calculating the total score SUMscore of each route in step S75, the process proceeds to step S81, which will be described later.

[0280] On the other hand, if it is determined that the specified search mode is "immediate pickup mode," in step S76, the demand forecasting application determines multiple routes for the specified route search area using a depth-limited search based on graph theory. More specifically, with nodes in graph theory representing intersections and edges representing roads (paths) between intersections, multiple routes are determined within the route search area using a depth-limited search. The route search area may be the specialty area if the vehicle's position is within the specialty area, or may be within a certain radius of the vehicle's position, or may be a predetermined number of areas AR centered on the vehicle's position.

[0281] In step S77, the demand forecasting application calculates a total score SUMscore for each of the searched routes. For example, the demand forecasting application calculates and adds up the score Sc for each boarding point on the searched route, for example, as score Sc = boarding demand level [1-5] for the area AR that includes the boarding point, to calculate the total score SUMscore.

[0282] The total score SUMscore of the route is not limited to the above example, and any calculation method can be used that will result in a larger total score SUMscore for routes with a high level of demand forecast and many boarding points.Addition or subtraction of the reverse score Sc, and updating of the predicted time of the demand forecast, can be performed in the same way as in the "destination mode" calculation.

[0283] After calculating the total score SUMscore of each route in step S77, the process proceeds to step S81, which will be described later.

[0284] On the other hand, if it is determined that the specified search mode is the "nearby boarding point mode," in step S78, the demand forecasting application extracts boarding points that are within a certain distance from the vehicle position and assigns a score Sc to each extracted boarding point.

[0285] The score Sc for each boarding point in the "nearby boarding point mode" can be assigned, for example, as follows:

[0286] For example, the demand prediction application assigns a larger score Sc to each boarding point the closer the boarding date and time is to the current date and time.The demand prediction application assigns a larger score Sc to each boarding point the greater the number of boardings (number of boardings).The demand prediction application assigns a larger score Sc to each boarding point the greater the total boarding time (total boarding time for each boarding) or the longer the average boarding time.

[0287] The demand forecasting app assigns a score Sc to each boarding point, with a higher score for locations that can be reached by going straight or turning left from the vehicle's position, and a lower score for locations that can be reached by turning right.

[0288] For example, as shown in FIG. 31, assume that taxi 11 is traveling at the location of vehicle position mark 505, and there are three boarding points: boarding point HS1 that can be reached by going straight, boarding point HS2 that can be reached by turning left, and boarding point HS3 that can be reached by turning right. In Japan, where traffic is on the left side of the road, boarding point HS1 by going straight and boarding point HS2 by turning left are relatively easy to reach, but boarding point HS3 by turning right, which requires crossing the oncoming lane, often takes a long time to reach because it depends heavily on the timing of traffic lights and the timing of vehicles traveling in the oncoming lane. Therefore, the demand forecasting app can assign a higher score Sc to boarding points HS1 and HS2 than to boarding point HS3.

[0289] In foreign countries where traffic drives on the right, a higher score Sc is assigned to locations that can be reached by going straight or turning right, and a lower score Sc is assigned to locations that can be reached by turning left. In other words, a higher score Sc is assigned to boarding points that go straight or turn in a direction that does not require crossing an oncoming lane, and a lower score Sc is assigned to boarding points that require crossing an oncoming lane.

[0290] The demand prediction application assigns a higher score Sc to each boarding point the closer it is to the vehicle's position.

[0291] The demand forecasting application assigns a score Sc to each boarding point, which increases the larger the proportion of the boarding point within the area AR in which the boarding point is located (e.g., the number of boardings at that boarding point divided by the total number of boardings within the area AR).

[0292] When a destination is set, the demand prediction application assigns a score Sc at each boarding point that increases as the degree of coincidence with the direction of the destination increases.

[0293] The sum of the scores Sc assigned to each boarding point in this way will be the final score Sc for each boarding point. Note that the above is just an example of how to assign the score Sc to each boarding point, and the score Sc may be assigned based on other criteria.

[0294] In step S79, the demand forecasting application refers to the score Sc assigned to each boarding point, and searches for multiple routes that use a predetermined number of boarding points with high scores Sc as via points. More specifically, first, the application refers to the score Sc assigned to each boarding point, and extracts a predetermined number of boarding points with high scores Sc. Then, multiple (a predetermined number) routes are searched for using a route search algorithm so as to pass through the extracted boarding points.

[0295] In step S80, the demand forecasting application calculates a total score SUMscore for each searched route. For example, the demand forecasting application calculates the total score SUMscore by adding up the scores Sc of each boarding point on the searched route. Alternatively, the total score SUMscore may be the number of boarding points on the route.

[0296] After calculating the total score SUMscore of each route in step S80, the process proceeds to step S81.

[0297] In step S81, the route with the highest total score (SUMscore) among multiple routes calculated in one of the search modes, "destination mode," "immediate pickup mode," or "nearby pickup point mode," is displayed on the display as a recommended route. The recommended route presentation screen shown in Figure 29 is an example displayed in the "destination mode" search mode.

[0298] The recommended route presentation process of FIG. 30 is executed as described above. After the recommended route is presented, route guidance begins, similar to a typical car navigation system. That is, in addition to the route display on the display, the route is instructed to the driver by voice guidance such as "Turn right at the next intersection." When the route is displayed on the display, the symbol, color, pattern, etc. of the boarding point displayed on the route may be changed according to the value (magnitude) of the boarding point's score Sc, as in the demand point 451 shown in FIG. 26.

[0299] When the driver operates the touch panel or the like to stop the guidance of the recommended route, the guidance of the recommended route is terminated. Note that the demand forecasting application of the terminal device 23 may acquire the status of "occupied vehicle," "vacant vehicle," or "pick-up vehicle" from the fare meter 21 or the vehicle management device 22, and may terminate (automatically terminate) without any operation by the driver when the status becomes other than "vacant vehicle," i.e., "occupied vehicle" or "pick-up vehicle."

[0300] <19. No riding area guide signs> Taxi drivers need to be aware of no-ride zones, where it is prohibited to pick up passengers outside of taxi stands. If a driver who is not familiar with no-ride zones, such as a new driver, picks up a passenger in a no-ride zone, he or she will be subject to severe penalties. For example, in the Kanto region, no-ride zones are set up in Ginza and Shinbashi. In the Kansai region, no-ride zones are set up in Kitashinchi and Minamichi.

[0301] The demand forecasting app can display no-boarding areas on the demand forecast screen.

[0302] FIG. 32 shows an example of a demand forecast screen displaying no-boarding areas.

[0303] In FIG. 32, parts corresponding to those in FIG. 29 etc. are given the same reference numerals, and the description of those parts will be omitted as appropriate.

[0304] 32, a detail button 511, a wide area button 512, and a full screen display button 513 are displayed on a map 41 on which a demand forecast mesh 63 is superimposed. Also displayed on the map 41 is a current location button 514 that switches the display of the map 41 to a display based on the vehicle's position.

[0305] Furthermore, a no-ride zone display 521 is displayed in an area of ​​the map 41 that corresponds to a no-ride zone. A detailed display 522 that displays detailed information about the no-ride zone is also displayed near the no-ride zone display 521. The detailed display 522 includes the text "No-ride zone 10 PM to 1 AM," which indicates the hours during which the no-ride zone applies, and an [OFF] button for erasing the no-ride zone display 521. Because the no-ride zone display 521 is superimposed on the map 41, if the display indicating the demand forecast pick-up point is difficult to see or if a driver does not need the no-ride zone display, the no-ride zone display 521 can be erased by operating the [OFF] button. The demand forecasting app erases the no-ride zone display 521 and the detailed display 522 outside of the hours during which the no-ride zone applies.

[0306] 32, a predicted time display section 502 and an additional information display section 531 are provided in an area different from the display area of ​​the map 41. The additional information display section 531 displays, for example, train operation information, event information, weather information, and the like.

[0307] The information on the no-ride areas may be stored in advance in the demand prediction application (terminal device 23), or may be acquired from the server 12 or a server of another information provider.

[0308] <20. Placement indication> In the display of the waiting time forecast shown in Figure 17, it has been explained that there is a method called "waiting at a taxi," in which passengers are obtained by lining up for taxis 11 at a taxi stand and leaving the taxis 11 waiting. Places where "waiting at a taxi" takes place (hereinafter referred to as "waiting places") include taxi stands in front of stations or hotels, and in front of the entrances of designated office buildings. Some of these waiting places are limited to certain taxi companies. Taxis 11 of other taxi companies cannot use a waiting place that is limited to a designated taxi company. The demand forecasting app has a function to display that the waiting place is exclusively for a designated taxi company.

[0309] FIG. 33 shows an example of a demand forecast screen displaying the pickup locations for each taxi company.

[0310] On the demand forecast screen of FIG. 33, a detail button 511, a wide area button 512, a full screen display button 513, and a current location button 514 are displayed on a map 41 on which a demand forecast mesh 63 is superimposed.

[0311] Furthermore, at each of the predetermined boarding points 541A and 541B on the map 41, pickup location indicators 551A and 551B are displayed, indicating that the boarding points are pickup locations. Pickup location indicator 551A is represented by a solid circular figure, and pickup location indicator 551B is represented by a dashed circular figure, with pickup location indicator 551A and pickup location indicator 551B being displayed in different ways. This difference in display method indicates that different taxi companies can use the pickup locations. When there is no particular distinction between pickup location indicator 551A and pickup location indicator 551B, they are simply referred to as pickup location indicator 551.

[0312] The driver can display detailed information 553 of a predetermined pickup location by selecting (tapping) the pickup location display 551. In the example of Fig. 33, detailed information 553 about pickup location display 551A of boarding point 541A is displayed.

[0313] Detailed information 553 displays information such as the name of the stop, whether the taxi 11 can use the parking spot, the time periods during which the taxi 11 can pick up passengers, the name of the taxi company that can use the parking spot, and other available parking spots displayed on the map 41 on the demand forecast screen.

[0314] Detailed information 553 of pickup location display 551A in Figure 33 displays "Hotel Shinagawa" as the name of the stop, "X (unavailable)" indicating that the pickup location is unavailable for taxi 11, "all day" as the time period during which passengers can be picked up, "Km Taxi Only" as the name of the taxi company that can use the pickup location, and "Osaki Think Building" as the name of another available pickup location displayed on map 41 of the demand forecast screen, i.e., pickup point 541B of pickup location display 551B.

[0315] The demand forecasting app can identify the driver's taxi company by registering (inputting) the company ID that identifies the company and the driver ID that identifies the driver working in taxi 11 on the login screen or settings screen when the demand forecasting app is launched.

[0316] 33, the pickup location displays 551A and 551B are represented by circular shapes surrounding the boarding point, but the display method for indicating the pickup location is not limited to this. The pickup location display 551 may change its display (color or symbol) depending on whether the taxi 11 is available for use, or the pickup location display 551 may be displayed only at pickup locations where the taxi 11 is available.

[0317] Information about the pickup location may be stored in advance in the demand prediction app, or may be acquired from the server 12 or a server of another information provider. The vehicle movement log data includes a company ID that identifies the company to which the taxi 11 belongs. Therefore, the server 12 not only acquires known pickup location information, but also estimates the boarding point from the history of the vehicle movement log data, and can identify whether the boarding point is a pickup location and, if so, which taxi company is available. Whether the boarding point is a pickup location can be determined by detecting the taxi 11's waiting behavior as described above. Therefore, the boarding change point at which the status changes from "empty" to "occupied" after the waiting behavior can be determined as the pickup point of the pickup location.

[0318] The demand prediction application's pickup location display 551 prevents drivers from going to unnecessary pickup locations, enabling efficient sales.

[0319] <21.Train time display> During the operating hours of a certain railway line, at the station where the last train (the so-called last train) departs, the demand for taxis 11 increases due to people who miss that train. Also, at the station where the last train arrives, other transfer lines and buses often stop running at that time, so the demand for taxis 11 increases. Alternatively, on railway lines where the number of trains is low (the interval between trains is long), people who get off at a certain station often use taxis 11. Therefore, if train time information such as the last train time and arrival time can be provided to drivers, drivers can acquire passengers by heading to the taxi stand in front of the station based on that train time information.

[0320] The demand forecasting application has a function of displaying predetermined train times, such as the last train times for stations on the railway line displayed on the map 41, on the demand forecast screen.

[0321] FIG. 34 shows an example of a demand forecast screen displaying the time of the last train.

[0322] Map 41 on the demand forecast screen in Figure 34 shows Shinagawa Station of Keikyu Corporation, Shinagawa Station of East Japan Railway Company (JR East), and Takanawadai Station on the Toei Asakusa Line, and displays the train times for the last trains at each station.

[0323] Specifically, the time display 571 displays "Shinagawa Station 0:23 bound for Kanazawa Bunko," indicating that the departure time of the last train bound for Kanazawa Bunko from Shinagawa Station operated by Keikyu Corporation is 0:23.

[0324] The time display 572 displays "Shinagawa Station 0:46 bound for Osaki," indicating that the departure time of the last train bound for Osaki from Shinagawa Station of East Japan Railway Company (JR East Japan) is 0:46.

[0325] Time display 573 displays "Takanawadai Station 0:31 bound for Nishi-Magome," indicating that the last train bound for Nishi-Magome from Takanawadai Station on the Toei Asakusa Line departs at 0:31.

[0326] On the right side of the map 41 on the demand forecast screen, a list display section 581 is displayed, which displays the train time information displayed on the map 41 in a list.

[0327] The list display section 581 displays the same information as the time displays 571 to 573 in a predetermined order, such as in ascending or descending order of train times, or in descending or descending order of distance from the vehicle's location to the station. The sort button 582 is operated to change the order displayed in the list display section 581, such as from earliest to latest train times, from closest to farthest distance from the vehicle's location, or from train time to distance from the vehicle's location. The train times may be either arrival times or departure times.

[0328] By having a function for displaying specific train times, such as the last train time, the demand forecasting app increases the chances of attracting passengers who have missed the last train or who have gotten off the last train. The train time display function can be turned on or off by setting. Instead of displaying train times for all stations displayed on the map 41 of the demand forecasting screen, it may be possible to limit the display to stations with a large number of users (with a certain number of users or more), terminal stations where multiple lines serve, or stations that are the first or last stop. Instead of stations displayed on the map 41, it may be possible to display stations within a certain distance from the vehicle's position (for example, within a 2.5 km radius). Alternatively, it may be possible to display only stations that are in the direction of travel of the vehicle. The conditions for the stations for which train times are to be displayed may be set on the setting screen.

[0329] As described above, on railway lines where trains run infrequently (with long intervals between trains), all train times may be displayed, not just the last train or the first train.

[0330] The demand prediction application can acquire train time display data from the server 12 together with the ride demand prediction data or as part of the ride demand prediction data, and display it on the demand prediction screen. The demand prediction application may display the train time display data in conjunction with the current location and time of the taxi 11, for example, a predetermined time before the train time, or may provide a train time display button on the demand prediction screen and display the train time display based on the driver's operation.

[0331] <22. Reverse boarding point display> When a passenger uses a taxi 11 bound for Haneda Airport or Narita Airport, the ride distance is expected to be long. For this reason, the driver may request passengers to go to a specific destination. Since the actual vehicle data records the departure and arrival points, by collecting actual vehicle data with arrival points at specific locations, it is possible to analyze pick-up points (departure points) with destinations at specific locations.

[0332] Therefore, the demand forecasting application has a reverse lookup boarding point display function that allows the driver to specify a specific location as a destination, and displays only the boarding points where passengers actually boarded the specified location as their destination in past actual vehicle data. A reverse lookup boarding point refers to a boarding point where the destination is limited to a specific location. Places that can be specified as destinations include, for example, Haneda Airport, Narita Airport, and Tokyo Disney Resort (registered trademark) (hereinafter referred to as TDR).

[0333] FIG. 35 shows an example of a demand forecast screen in which the reverse lookup boarding point display function has been executed.

[0334] The demand forecast screen of FIG. 35 is displayed, for example, when the driver operates a reverse lookup boarding point display button or the like displayed on the demand forecast screen.

[0335] 35, a reverse lookup boarding point display section 601 is placed next to (to the right of) the map 41 on which the demand forecast mesh 63 is superimposed. Below the map 41, a predicted time display section 602 is placed.

[0336] The reverse lookup boarding point display section 601 displays a list of passenger destinations for which demand forecasts are to be displayed, and an execute button 611 for displaying the demand forecast. In the example of Fig. 35, three destinations, Haneda Airport, Narita Airport, and TDR, are displayed, and the execute button 611A is touched (selected) to display the demand forecast for passengers heading to Haneda Airport, the execute button 611B is touched (selected) to display the demand forecast for passengers heading to Narita Airport, and the execute button 611C is touched (selected) to display the demand forecast for passengers heading to TDR.

[0337] The predicted time display section 602 displays the predicted time of the demand forecast, and like the predicted time setting area 42 in FIG. 2, the predicted time of the demand forecast can be changed.

[0338] FIG. 36 shows an example of a demand forecast screen for passengers heading to Haneda Airport when the execute button 611A is touched. The magnification of the map 41 on the demand forecast screen can be changed by operating the detail button 511 or the wide area button 512. In the example of FIG. 36, the scale of the demand forecast map 41 initially displayed by pressing the execute button 611A is set to a high magnification (wide area map). However, the scale of the demand forecast map 41 initially displayed by pressing the execute button 611A can also be set to the same magnification as the map 41 on the demand forecast screen of FIG. 35 at the time of execution. The demand forecast screen of FIG. 36 may display, for each boarding point and destination, the probability that the passenger is going to that destination.

[0339] Demand forecasts for passengers heading to a specific location can be classified by conditions such as time of day, weather, and day of the week (weekday, day before a holiday, holiday), and reverse lookup boarding points that match the conditions at the time the demand forecast is executed can be displayed.

[0340] The one or more destinations displayed in the reverse lookup boarding point display unit 601 may be preset locations such as Haneda Airport, Narita Airport, and TDR shown in Fig. 35, or may be locations that are frequently visited from the current location of the taxi 11. In addition to the above-mentioned examples, other theme parks, concert venues, event venues, baseball stadiums, etc. may also be set as destinations.

[0341] <23. Display of demand forecast classification for dispatch / driving / waiting> When a user rides in a taxi 11, there are three ways to obtain a taxi 11: "trip request," "cruising vehicle," and "customer waiting at stand." "Trip request" is a method of arranging a taxi 11 through a taxi company's call center or app, etc., and having the taxi 11 come to a specified location. "Cruising" is a method of catching an empty taxi 11. "Cruising" is a method of moving to a taxi location and getting into a taxi 11 that is waiting at a taxi. When the demand forecasting app displays the demand forecast on the demand forecast screen, it can display the demand forecast by distinguishing between the riding methods of "trip request," "cruising vehicle," and "customer waiting at stand." In other words, the demand forecasting app has a function of displaying the demand forecast by distinguishing between the riding methods of "trip request," "cruising vehicle," and "customer waiting at stand."

[0342] FIG. 37 shows an example of a demand forecast screen that displays demand forecasts by distinguishing between different boarding methods: "dispatch," "driving," and "waiting."

[0343] The demand forecast screen of FIG. 37 includes a map 41 on which a demand forecast mesh 63 is superimposed, a forecast time display section 502 that allows the user to specify the forecast time for demand forecasting, and an additional information display section 531.

[0344] On the map 41 of the demand forecast screen, in addition to a details button 511, a wide area button 512, a full screen display button 513, and a current location button 514, there are provided an attached display button 641, a passing display button 642, and a dispatch display button 643. Also, to make it possible to distinguish between the "dispatch," "passing," and "waiting for" boarding methods, boarding points 645 are displayed in different ways, such as with different colors, patterns, and mark shapes, on the map 41.

[0345] The add-on display button 641 is operated when displaying the boarding point 645 for the "waiting to pick up" boarding method on the map 41. The cruising display button 642 is operated when displaying the boarding point 645 for the "cruising" boarding method on the map 41. The dispatch display button 643 is operated when displaying the boarding point 645 for the "dispatch" boarding method on the map 41. The add-on display button 641, cruising display button 642, and dispatch display button 643 are toggle buttons, and each time they are operated, the display of the boarding point 645 can be turned on or off for the specified boarding method. Any combination of "dispatch," "cruising," and "waiting to pick up" is also possible. For example, when both "dispatch" and "cruising" are turned on, the map 41 displays the boarding points 645 for both the "dispatch" boarding method and the "cruising" boarding method.

[0346] The demand forecast for boarding points for different boarding methods, "dispatch," "driving," or "waiting," can be obtained by forecasting the boarding demand for each boarding method, "dispatch," "driving," or "waiting." Actual vehicle data for "dispatch" can be collected by collecting actual vehicle data for the "actual vehicle" immediately after the status of taxi 11 becomes "pick-up." Actual vehicle data for "waiting" can be collected by collecting actual vehicle data for the "actual vehicle" after the waiting operation. Actual vehicle data for "driving" can be actual vehicle data other than "dispatch" and "waiting."

[0347] By displaying ride demand (boarding point) with a distinction between different boarding methods, such as "dispatch," "driving," or "waiting," it is possible to present drivers with ride demand that matches their business style.

[0348] <24. Display of predicted fare> In FIG. 18, an example has been described in which, when a predetermined area AR is selected as the area of ​​interest AR, the long display 241 is displayed, displaying the proportion of long-distance passengers in the area of ​​interest AR.

[0349] Also, in FIG. 19, an example has been described in which, when a predetermined area AR is selected as the attention area AR, a travel distance display 251 is performed that displays the average travel distance and its confidence interval of passengers who board in the attention area AR.

[0350] In addition, in the explanation of Figure 19, it was explained that the demand forecasting application may display the average fare (fare) and confidence interval instead of the average travel distance and confidence interval, and that it can predict travel demand by time of day and weather.

[0351] FIG. 38 shows an example of a display in which the average fare (fare) and confidence interval of passengers boarding in the area of ​​interest AR are displayed.

[0352] As shown in FIG. 38, the demand forecasting application can display fare information 711, which displays the average fare and its confidence interval for passengers boarding in the area of ​​interest AR, in addition to the number of passengers in the entire area of ​​interest AR.

[0353] The fare display 711 shows that the average fare in the area of ​​interest AR is "2,400 yen," and that the confidence interval of the average fare at a confidence level of, for example, 70% is "1,110 yen to 3,700 yen." The confidence level of the confidence interval is not limited to 70% and can be set arbitrarily, such as 80%.

[0354] Further, the fare display 711 indicates that the predicted fare displayed is based on actual vehicle data specific to the "18:00-18:30 time period, weekday, rainy, October" period.

[0355] In this way, by displaying the average fare and its confidence interval for the area of ​​interest AR, the driver can, for example, search for an area AR where a high fare is expected.

[0356] The fare display 711 may be displayed for the area of ​​interest AR as shown in Figure 38, or may be displayed in conjunction with the pinpoint boarding location mark 221 (Figure 15) to display the average fare and confidence interval for the pinpoint boarding location.

[0357] In the above explanation, each area AR of the demand forecast mesh 63 is displayed with different colors and intensities according to the degree of demand for travel, as explained with reference to Fig. 2. However, the color and intensities may be changed depending on the expected fare. In this case, the driver can, for example, select a route where a high fare is expected and drive in a "cruising" manner.

[0358] <25. Display of real-time number of available vehicles> The demand forecasting app predicts and displays demand for rides at a specified time (time period), but if demand for rides of 10 vehicles is predicted in the attention area AR, but there are 20 taxis 11 wanting to pick up passengers there, the 10 taxis 11 will not be able to acquire passengers. In other words, whether or not passengers can be acquired also depends on the relationship between supply and demand.

[0359] A taxi company manages the current location and status, such as "occupied," "empty," or "pick-up," of each taxi 11 in operation in real time at a dispatch center or the like. By combining operation data, including the current location and status of each taxi 11, acquired in real time, with a ridership demand forecast, drivers can operate efficiently, taking into account the relationship between supply and demand described above. Note that "real time" includes a slight time lag (for example, several minutes) required for collecting information on the current location and status of each taxi 11 in operation, transmitting operation data to a demand forecasting app, etc.

[0360] FIG. 39 shows an example of a demand forecast screen that displays real-time vacant vehicle information based on operation data.

[0361] The demand forecast screen of FIG. 39 includes a map 41 on which a demand forecast mesh 63 is superimposed, a forecast time display section 502 that allows the user to specify the forecast time for demand forecasting, and an area information display section 741.

[0362] In each area AR divided by the demand forecast mesh 63 of the map 41, information 742 about available taxis 11 present in that area AR is displayed in real time. The information 742 about available taxis 11 indicates the number of taxis 11 currently moving within that area AR and having an "available" status.

[0363] Of the areas AR divided by the demand forecast mesh 63, an area of ​​interest AR designated by the driver has an area of ​​interest frame 211 displayed. Detailed information about the area of ​​interest frame 211 is displayed as area information in the area information display section 741. The area information display section 741 displays, for example, the predicted number of vehicles for passenger demand in the area AR, the long-distance passenger rate in the area AR, and the like. The example in FIG. 39 shows that there are currently five "empty" taxis 11 in the area of ​​interest AR where the area of ​​interest frame 211 is displayed, while the predicted number of vehicles for passenger demand is zero.

[0364] FIG. 40 shows an example of a demand forecast screen display showing real-time vacant vehicle information when the map 41 is displayed at a low scale, in other words, in a detailed map display.

[0365] Fig. 39 shows an example of a demand forecast screen display showing real-time vacant vehicle information when the scale of the map 41 is high, in other words, when a wide-area map is displayed. When a wide-area map is displayed, as shown in Fig. 39, the number of "vacant" taxis 11 is displayed as vacant vehicle information for each area AR of the demand forecast mesh 63.

[0366] On the other hand, in the case of a detailed map display, as shown in FIG. 40, an icon 751 indicating the presence of an "empty" taxi 11 is displayed at the position where the "empty" taxi 11 is present.

[0367] At the boarding point of the waiting place, the number of vehicles waiting at the boarding point of the waiting place can also be displayed.

[0368] By having a demand forecasting app with the ability to display real-time information on available vehicles based on operation data, drivers can select areas AR where they have a high probability of acquiring passengers, thereby increasing the probability of acquiring passengers.

[0369] Furthermore, if the demand prediction application can acquire operation data, it can search for and present recommended routes, including real-time vacant taxi information from the operation data, in the recommended route presentation process described above. That is, when selecting a predetermined route as part of a recommended route, the demand prediction application selects as a recommended route a route where the predicted number of taxis with passenger demand is equal to or greater than the number of "vacant" taxis, or assigns a large score Sc to the route. Furthermore, the demand prediction application may include a process for selecting as a recommended route if no "vacant" taxis 11 have been passing along the predetermined route for a certain period of time from the present until a certain time ago, since there may be passenger demand.

[0370] <26. Display of daily sales evaluation> The demand forecasting app can have a function to output sales evaluation information that evaluates the driver's work for the day after the driver has finished working for the day. Sales evaluation can be based on excellent drivers, for example, drivers with high average daily sales. By using excellent drivers as the evaluation standard driver, it is possible to provide drivers with information to increase sales.

[0371] 41 shows an example of an evaluation screen that outputs one-day business evaluation information. This evaluation screen is displayed, for example, when an operation to close one day's business is performed in the demand forecasting app.

[0372] A title display 811 is displayed at the top of the evaluation screen in Figure 41. In the example in Figure 41, "10 / 11 Crew Score" is displayed, indicating that the evaluation information is for the business day on October 11th.

[0373] The evaluation screen also includes an evaluation score display 812, a radar chart 813, a business income graph 814, and a route history display button 815.

[0374] The evaluation score display 812 shows the overall evaluation of the driver for the entire day, expressed as a value with a standard driver being given a score of 100. By referring to this overall evaluation value, it is possible to confirm how close the driver has become to the standard driver.

[0375] The radar chart 813 shows the evaluation results, which show the driver's overall evaluation value for the entire day, broken down into multiple categories. In the example of FIG. 41, the evaluation is divided into five categories: operating income, on-board vehicle rate, empty vehicle time, number of business trips, and business range. Operating income represents an evaluation value from the perspective of operating income (sales) per actual driving time. The on-board vehicle rate represents an evaluation value from the perspective of "on-board" driving time / total driving time. Empty vehicle time represents an evaluation value from the perspective of "empty vehicle" time / total driving time. Number of business trips represents an evaluation value from the perspective of the number of times a customer was picked up. Business range represents an evaluation value from the perspective of the size of the area traveled.

[0376] The operating income graph 814 displays the daily operating income trend, with the horizontal axis representing the business hours (business hours) and the vertical axis representing the operating income. The solid line 821 displayed in the operating income graph 814 represents the driver's actual sales. Meanwhile, the dashed line 822 displayed in the operating income graph 814 represents ideal hypothetical sales based on the driver's actual driving route and the driving data of other taxis 11. For example, with the driver's actual driving route and the driving data of other taxis 11, it is possible to analyze cases where the driver actually went straight at a certain intersection, but if the driver had turned left, it is expected that the driver would have acquired a passenger. By analyzing such assumptions with the actual driving route, it is possible to predict the ideal operating income that could have been obtained with a slight change in the driving route. Such ideal operating income is displayed as the dashed line 822. In addition, comments are displayed on the dashed line 822 for points (locations) that could have increased operating income, such as "If the driver had turned at Higashi Ginza 7-chome" or "If the driver had turned at Shinbashi 5-chome."

[0377] The route history display button 815 is a function for displaying the driver's actual driving history for one day on a map.

[0378] FIG. 42 shows an example of the operation history screen that is displayed when the route history display button 815 is operated.

[0379] As shown in Fig. 42, the operation history screen displays the route traveled by the taxi 11 from the start of business to the end of business in one day, the status of "occupied vehicle," "empty vehicle," or "pick-up," the locations where passengers picked up and dropped off, and the times. Note that although the example in Fig. 42 does not display the map 41, in reality, the screen is superimposed on the map 41. Also, although the example in Fig. 42 displays only a portion of the operation route for one day, the driver can check all or part of the operation route for one day by changing the display magnification of the map 41.

[0380] The operation history screen of Figure 42 may be displayed on a single screen together with the evaluation screen of Figure 41. The operation history screen of Figure 42 can be used as a reference for the next operation by referring to the points that had the potential to increase operating income in the operating income graph 814 on the evaluation screen of Figure 41.

[0381] 27. Displaying additional information taking distance and direction into account When displaying additional information such as train operation information, event information, weather information, etc., the additional information display unit 531 may be provided in an area different from the display area of ​​the map 41, as shown in the example of Figure 32, or the additional information may be displayed on the map 41 taking into account distance and direction.

[0382] FIG. 43 shows an example of a demand forecast screen that displays additional information taking into account distance and direction.

[0383] In the example of FIG. 43, additional information 831 and additional information 832 are displayed on the map 41.

[0384] Additional information 831 is information notifying that service has been suspended at Tamachi Station on the Yamanote Line. Additional information 831 is displayed at a position corresponding to the direction of Tamachi Station, based on vehicle position mark 505, which indicates the current location of taxi 11. In the example of FIG. 43, Tamachi Station is located outside the display area of ​​map 41 on the demand forecast screen, so only additional information 831 is displayed at a position corresponding to the direction of Tamachi Station. However, if Tamachi Station were located on map 41 on the demand forecast screen, a symbol such as an X or a △, indicating that service has been suspended, would be displayed in the area of ​​Tamachi Station on map 41 along with additional information 831. Alternatively, only a symbol may be displayed as additional information 831, and detailed information may be displayed when the symbol is tapped (selected).

[0385] Additional information 832 is information notifying that a train delay has occurred at Togoshi Station on the Toei Asakusa Line. Additional information 832 is displayed at a position corresponding to the direction of Togoshi Station, based on vehicle position mark 505 indicating the current location of taxi 11.

[0386] The direction based on the vehicle position mark 505 may be a precise angle in units of one degree, or may be an angle that converges to a predetermined range such as four directions or eight directions.

[0387] Similarly, when the distance to the position related to the additional information is far from the current location, it is displayed far from the vehicle position mark 505 , and when it is close, it is displayed close to the vehicle position mark 505 .

[0388] For example, if the additional information is information regarding a train delay, the direction and distance from the vehicle position mark 505 can be calculated for a station that is calculated to be affected as the position related to the additional information.

[0389] For example, if the additional information is information relating to an event, the location where the event is held can be used as the location relating to the additional information, and the direction and distance from the vehicle position mark 505 can be calculated.

[0390] For example, if the additional information is information about weather such as a sudden downpour, the location of the weather phenomenon can be used as the location related to the additional information to calculate the direction and distance from the vehicle position mark 505.

[0391] As described above, by displaying the additional information on the map 41 according to the distance and direction of the additional information and presenting it to the driver, the driver can intuitively understand the additional information.

[0392] In addition, on the demand forecast screen of FIG. 43, the display of the boarding point 645 (similar to the demand point 451) displayed on the map 41 and the display of the forecasted time display section 502 are omitted.

[0393] <28. Displaying information according to the direction of travel> Among the forecast information such as boarding points in demand forecasts and additional information such as event information and train delay information shown in Fig. 43 and the like, information in the direction of travel of the taxi 11 is important, but information in the opposite direction to the direction of travel is not so important. The same is true for the route information on the map 41.

[0394] Therefore, when displaying the map 41 on the demand forecast screen, the demand forecast application can display information about the traveling direction in a manner that displays a larger amount of information than information about the opposite direction to the traveling direction.

[0395] A in FIG. 44 shows a display example in a head-up mode in which the traveling direction of the taxi 11 is at the top (upper edge) of the screen.

[0396] In head-up mode, the vehicle position mark 505 is positioned so that the right area R1 and the left area L1 are identical or approximately identical in the left-right direction, and the upper area U1 is larger than the lower area D1 in the up-down direction, relative to the entire area of ​​the map 41.

[0397] FIG. 44B shows a display example in North Up mode, in which the north direction is at the top (upper edge) of the screen, regardless of the traveling direction of the taxi 11.

[0398] In the north-up mode, the allocation between the right region R1 and the left region L1, and the allocation between the upper region U1 and the lower region D1, differ depending on the traveling direction of the taxi 11, and Fig. 44B shows a display example in which the traveling direction of the taxi 11 is northeast. In this case, the right region R1 is larger than the left region L1 in the left-right direction, and the upper region U1 is larger than the lower region D1 in the up-down direction.

[0399] Other illustrations are omitted, but for example, when the taxi 11 is traveling in a southwesterly direction, the left area L1 is larger than the right area R1 in the left-right direction, and the lower area D1 is larger than the upper area U1 in the up-down direction.

[0400] As described above, by displaying information in the traveling direction so that the amount of information displayed is greater than the amount of information displayed in the opposite direction to the traveling direction, more useful information can be displayed to the driver.

[0401] <29. Computer Configuration Example> The above-described series of processes can be executed by hardware or software. When the series of processes is executed by software, the programs that make up the software are installed on a computer. Here, the term "computer" includes microcomputers built into dedicated hardware, and general-purpose personal computers, for example, that can execute various functions by installing various programs.

[0402] FIG. 45 is a block diagram showing an example of the hardware configuration of a computer when the computer executes the processes executed by the server 12, the vehicle management device 22, or the terminal device 23 using a program.

[0403] In the computer, a CPU (Central Processing Unit) 301, a ROM (Read Only Memory) 302, and a RAM (Random Access Memory) 303 are interconnected by a bus 304.

[0404] An input / output interface 305 is further connected to the bus 304. To the input / output interface 305, an input unit 306, an output unit 307, a storage unit 308, a communication unit 309, and a drive 310 are connected.

[0405] The input unit 306 includes operation buttons, a keyboard, a mouse, a microphone, a touch panel, an input terminal, etc. The output unit 307 includes a display, a speaker, an output terminal, etc. The storage unit 308 includes a hard disk, a RAM disk, a non-volatile memory, etc. The communication unit 309 includes a network interface, etc. The drive 310 drives a removable recording medium 311 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory.

[0406] In the computer configured as above, the CPU 301 performs the above-described series of processes by, for example, loading a program stored in the storage unit 308 into the RAM 303 via the input / output interface 305 and the bus 304 and executing the program. The RAM 303 also stores data necessary for the CPU 301 to execute various processes as needed.

[0407] The program executed by the computer (CPU 301) can be provided by being recorded on a removable recording medium 311 such as a package medium, for example. The program can also be provided via a wired or wireless transmission medium such as a local area network, the Internet, or digital satellite broadcasting.

[0408] In a computer, the program can be installed in the storage unit 308 via the input / output interface 305 by inserting the removable recording medium 311 into the drive 310. The program can also be received by the communication unit 309 via a wired or wireless transmission medium and installed in the storage unit 308. Alternatively, the program can be installed in the ROM 302 or the storage unit 308 in advance.

[0409] In this specification, a system refers to a collection of multiple components (devices, modules (components), etc.), regardless of whether all the components are contained in the same housing. Therefore, multiple devices housed in separate housings and connected via a network, and a single device with multiple modules housed in a single housing, are both systems.

[0410] In addition, in this specification, the steps described in the flowcharts may be performed in chronological order in the order described, but they do not necessarily have to be processed in chronological order, and may be performed in parallel or at any necessary timing, such as when a call is made.

[0411] The embodiments of the present technology are not limited to the above-described embodiments, and various modifications are possible without departing from the spirit of the present technology.

[0412] The above-described embodiment has been described as an example of a prediction system that predicts demand for taxis as commercial vehicles, but the present invention can also be applied to systems that predict demand for other commercial vehicles that carry passengers (people), specifically, buses, trains, airplanes, ships, helicopters, etc., as well as commercial vehicles that carry goods (luggage), trucks, dump trucks, etc. Furthermore, the commercial vehicle may be an unmanned transport vehicle such as a drone.

[0413] For example, it is possible to adopt a form in which all or part of the above-described embodiments are combined as appropriate.

[0414] For example, this technology can be configured as cloud computing, in which a single function is shared and processed collaboratively by multiple devices via a network.

[0415] Furthermore, each step described in the above flowchart can be executed by one device, or can be shared and executed by multiple devices.

[0416] Furthermore, when one step includes multiple processes, the multiple processes included in that one step can be executed by one device or can be shared and executed by multiple devices.

[0417] The effects described in this specification are merely examples and are not limiting, and there may be effects other than those described in this specification.

[0418] The present technology can also be configured as follows. (1) The system divides the business area of ​​the commercial vehicle into a plurality of areas, and among the plurality of areas for which passenger demand for each area has been predicted, includes a display control unit that displays, on a display unit, the direction of movement and distance of movement of passengers in a target area, which is a target area among the plurality of areas for which passenger demand for each area has been predicted. Information processing device. (2) The display control unit causes the display unit to display the movement direction and the movement distance using an arrow pointing outward from the attention area. The information processing device according to (1) above. (3) The direction of the arrow indicates the direction of travel, the length of the arrow indicates the distance traveled, and the thickness of the arrow indicates the proportion of passengers in the direction of travel in all directions. The information processing device according to (2) above. (4) The display control unit causes the display unit to display, as a prediction result, boarding locations in the attention area where boarding is frequent and the number of boardings at the boarding locations. The information processing device according to any one of (1) to (3). (5) The display control unit causes the display unit to display, as prediction results, boarding positions in the attention area with a high number of boardings, the number of boardings at the boarding positions, and the number of boardings in the attention area. The information processing device according to any one of (1) to (4). (6) The display control unit causes the display unit to display, as prediction results, a predetermined boarding position within the attention area, the number of boarding passengers at the boarding position, and a time required to wait at the boarding position and pick up passengers. The information processing device according to any one of (1) to (5). (7) The display control unit causes the display unit to display, as a prediction result, a ratio of trips in the area of ​​interest in which the trip distance is equal to or greater than a predetermined distance. The information processing device according to any one of (1) to (6). (8) The display control unit divides a travel distance in the area of ​​interest into a plurality of sections, and causes the display unit to display a travel ratio for each divided section as a prediction result. The information processing device according to any one of (1) to (7). (9) The display control unit also displays, on the display unit, the riding ratio for each of the sections across the plurality of areas as a prediction result. The information processing device according to (8). (10) The display control unit causes the display unit to display, as prediction results, the time and fare required to travel to the destination, and the time and fare required to travel for each division unit obtained by dividing the travel route to the destination into predetermined units. The information processing device according to any one of (1) to (9). (11) The display control unit causes the display unit to display the average riding distance of the passengers in the area of ​​interest and its confidence interval as a prediction result. The information processing device according to any one of (1) to (10). (12) The display control unit causes the display unit to display the average fare of the passenger in the area of ​​interest and its confidence interval as a prediction result. The information processing device according to any one of (1) to (10). (13) The display control unit causes the display unit to display the average riding time of the passengers in the area of ​​interest and its confidence interval as a prediction result. The information processing device according to any one of (1) to (10). (14) The display control unit combines, among the divided areas, adjacent areas in which the number of passengers in each area is equal to or less than a predetermined threshold, into one area and displays the prediction result on the display unit. The information processing device according to any one of (1) to (10). (15) The vehicle further includes a notification unit that notifies the driver by sound of locations with high demand for boarding that exist in the direction of travel. The information processing device according to any one of (1) to (14). (16) The notification unit notifies the area unit of the location where the demand for riding is high. The information processing device according to (15) above. (17) The notification unit changes the type of sound depending on the scale of the prediction result displayed on the display unit to notify the user of a location with high demand for boarding. The information processing device according to (15) or (16). (18) The notification unit notifies the user of the location with high demand for riding by changing the type of sound depending on the distance to the location with high demand for riding. The information processing device according to any one of (15) to (17). (19) The notification unit changes the type of sound depending on the level of the demand for riding, and notifies the user of the location where the demand for riding is high. The information processing device according to any one of (15) to (18). (20) The notification unit issues a notification by sound every time the vehicle passes through a location with high demand for boarding. The information processing device according to any one of (15) to (19). (twenty one) The notification unit turns notifications on and off in conjunction with the status of "occupied vehicle" or "vacant vehicle." The information processing device according to any one of (15) to (20). (twenty two) The sound is a sound effect or a voice message. The information processing device according to any one of (15) to (21). (twenty three) The notification unit further notifies at least one of train operation information, event information, and weather information by voice message. The information processing device according to any one of (15) to (22). (twenty four) The display control unit causes the display unit to display a recommended route based on a prediction result of the passenger demand. The information processing device according to any one of (1) to (23). (twenty five) The display control unit searches for a route to the set destination and displays it as the recommended route on the display unit. The information processing device according to (24). (26) The destination is an area or location where the driver is a specialist. The information processing device according to (25) above. (27) The area where the driver is good at is an area where the driver has traveled for a predetermined time or more. The information processing device according to (26) above. (28) The area where the driver is good at is an area where the driver has been carrying passengers for a predetermined period of time or more. The information processing device according to (26) or (27). (29) The display unit of the area in which the driver is good at is a city, ward, town, or village, The display unit of the location where the driver is good at is the pickup location where the number of passengers is large. The information processing device according to any one of (26) to (28). (30) The display control unit searches for a route that passes through a location where a demand for riding is predicted in the vicinity of the current location, and displays the route as the recommended route on the display unit. The information processing device according to (24). (31) The display control unit causes the display unit to display a route with a high total score obtained by adding up the scores of the routes that the user passes through as the recommended route. The information processing device according to any one of (24) to (30). (32) The total score is calculated by summing the scores for each area based on the level of rider demand. The information processing device according to (31). (33) The total score is calculated by summing the scores for each location where ride demand is predicted. The information processing device according to (31). (34) The score for a location where it is necessary to cross an oncoming lane is set lower than the score for a location where it is not necessary to cross an oncoming lane. The information processing device according to any one of (31) to (33). (35) The closer the predicted passenger direction is to the destination direction, the higher the score. The information processing device according to any one of (31) to (34). (36) The predicted time for predicting the riding demand in a predetermined area when searching for the recommended route is changed according to the distance from the current location. The information processing device according to any one of (24) to (35). (37) A route where the predicted number of passenger demands is equal to or greater than the number of vacant commercial vehicles is displayed as the recommended route. The information processing device according to any one of (24) to (36). (38) The display control unit further displays a no-ride area on the display unit. The information processing device according to any one of (1) to (37). (39) The display control unit further causes the display unit to display a location where "waiting for attachment" is to be performed. The information processing device according to any one of (1) to (38). (40) The display control unit further displays on the display unit the name of a company that can use the location where the "waiting" is performed. The information processing device according to (39). (41) The display control unit further displays on the display unit other "waiting" locations where the commercial vehicle can be used. The information processing device according to (39) or (40). (42) The display control unit further displays train times for stations on the map displayed on the display unit. The information processing device according to any one of (1) to (41). (43) The display control unit displays a list of train times at the plurality of stations in order of arrival time or in order of distance from the vehicle position to the station. The information processing device according to (42). (44) The display control unit displays only the boarding point where the passenger boarded the vehicle, with the designated location as the passenger's destination. The information processing device according to any one of (1) to (43). (45) The display control unit displays a plurality of the destinations and displays only the boarding point of the selected destination. The information processing device according to (44). (46) The display control unit distinguishes between the boarding methods of "dispatch," "driving," and "waiting," and causes the display unit to further display the boarding demand forecast. The information processing device according to any one of (1) to (45). (47) The display control unit turns on and off the display of the ride demand forecast for each of the ride methods, namely, "dispatch," "driving," and "waiting." The information processing device according to (46) above. (48) The display control unit causes the display unit to display the average fare of the passenger in the area of ​​interest and its confidence interval as a prediction result. The information processing device according to any one of (1) to (47). (49) The display control unit causes the display unit to display the average fare of the passenger at a predetermined boarding location and its confidence interval as a prediction result. The information processing device according to any one of (1) to (47). (50) The display control unit further displays real-time vacant vehicle information on the display unit. The information processing device according to any one of (1) to (49). (51) The display control unit displays the number of vacant vehicles for each area as the vacant vehicle information. The information processing device according to (50). (52) The display control unit displays an icon of the vacant commercial vehicle as the vacant vehicle information. The information processing device according to any one of (50) and (51). (53) The display control unit further causes the display unit to display business evaluation information that evaluates the business of the day after the business of the day is finished. The information processing device according to any one of (1) to (52). (54) The display control unit displays actual sales and virtual sales as part of the sales evaluation information. The information processing device according to (53). (55) The display control unit displays a driving route including a status of an occupied vehicle and an empty vehicle as part of the business evaluation information. The information processing device according to (53) or (54). (56) The display control unit further displays the additional information on the map of the display unit according to the distance or direction of the additional information. The information processing device according to any one of (1) to (55). (57) The display control unit causes the display unit to display information about the traveling direction so that an amount of information displayed is greater than an amount of information about the opposite direction to the traveling direction. The information processing device according to any one of (1) to (56). (58) The information processing device The operating area of ​​the commercial vehicle is divided into a plurality of areas, and the travel direction and travel distance of passengers in a target area, which is a target area among the plurality of areas for which passenger demand for each area has been predicted, are displayed on a display unit as a prediction result. Information processing methods. (59) On the computer, The operating area of ​​the commercial vehicle is divided into a plurality of areas, and the travel direction and travel distance of passengers in a target area, which is a target area among the plurality of areas for which passenger demand for each area has been predicted, are displayed on a display unit as a prediction result. A program for executing a process. [Explanation of symbols]

[0419] 1 Prediction system, 11 Taxi, 12 Server, 22 Vehicle management device, 23 Terminal device, 63 Demand forecast mesh, 121 Control unit, 131 Data generation unit, 132 Learning unit, 133 Prediction unit, 141 Control unit, 142 Operation unit, 143 Display unit, 145 Speaker, 146 Microphone, 211 Attention area frame, 212 Arrow, 221 Pinpoint boarding location mark, 222 Number of boardings display, 223 Waiting start button, 224 Waiting display, 241 Long display, 251 Travel distance display, 261 Individual display, 262 Destination display, 301 CPU, 302 ROM, 303 RAM, 306 Input unit, 307 Output unit, 308 Memory unit, 309 Communication section, 310 Drive, 521 No-boarding area display, 531 Additional information display section, 551 Attached location display, 553 Detailed information, 581 List display section, 582 Sort button, 601 Reverse lookup boarding point display section, 641 Attached display button, 642 Smooth display button, 643 Dispatch display button, 711 Boarding fare display, 741 Area information display section, 742 Vacant vehicle information, 751 Icon, 812 Evaluation score display, 813 Radar chart, 814 Operating income graph, 815 Route history display button, 821, 822 Solid line, 831, 832 Additional information

Claims

1. The system includes a control unit that searches for a route to be presented based on the route score, which is based on passenger demand forecast data for commercial vehicles, and on operational data for commercial vehicles including vacant vehicle information. Information processing system.

2. a control unit that searches for a route to be presented based on a score for a route on the route, the score being based on passenger demand forecast data for a commercial vehicle; The score is set based on the width of the road along the route. Information processing system.

3. a control unit that searches for a route to be presented based on a score for a route on the route, the score being based on passenger demand forecast data for a commercial vehicle; The score for a location where it is necessary to cross an oncoming lane is set lower than the score for a location where it is not necessary to cross an oncoming lane. Information processing system.

4. a control unit that searches for a route to be presented based on a score for a route on the route, the score being based on passenger demand forecast data for a commercial vehicle; The closer the predicted passenger direction is to the destination direction, the higher the score. Information processing system.

5. The control unit is provided for searching for a route that passes through a location where passenger demand is predicted around the current location based on the score of the route on the route, which is based on passenger demand prediction data for the commercial vehicle, as a route to be presented. Information processing system.

6. a control unit that searches for a route to be presented based on a score for a route on the route, the score being based on passenger demand forecast data for a commercial vehicle; The predicted time for predicting the demand for rides in a predetermined area when searching for the presented route is changed according to the distance from the current location. Information processing system.

7. The system is equipped with a control unit that searches for a route to be presented, the route being a score based on passenger demand forecast data for commercial vehicles, and in which the number of predicted passenger demands is equal to or greater than the number of vacant commercial vehicles, based on the score for the route. Information processing system.

8. The route is based on the passenger demand forecast data for commercial vehicles, and a control unit is provided that searches for a route on which no vacant commercial vehicles have passed for a certain period of time as a route to be presented, based on the score for the route on the route. Information processing system.

9. The control unit compares the scores of the routes on each route and searches for the route to be presented. The information processing system according to any one of claims 1 to 8.

10. The control unit compares the total scores obtained by adding up the scores for each route along the route to search for the route to be presented. The information processing system according to claim 9 .

11. The control unit searches for a route with a high total score as the route to be presented. The information processing system according to claim 10.

12. The total score is calculated by adding up the scores for each boarding point, which is a location where boarding demand is predicted. The information processing system according to claim 10.

13. The score of the boarding point is set based on at least one of the date and time of boarding, the number of boardings, the total boarding time, the average boarding time, and the percentage of boardings whose boarding distance is equal to or greater than a predetermined distance at the boarding point. The information processing system according to claim 12.

14. The total score is calculated by adding up the scores for each area into which the sales area of ​​the sales vehicle is divided. The information processing system according to claim 10.

15. The control unit controls the display of a map including the route to be presented. The information processing system according to any one of claims 1 to 8.

16. The presented route is a route recommended to the driver of the commercial vehicle. The information processing system according to any one of claims 1 to 8.

17. The commercial vehicle is a taxi The information processing system according to any one of claims 1 to 8.

18. The information processing system A route to be presented is searched for based on the route score on the route, which is based on the passenger demand forecast data of the commercial vehicle, and the operation data of the commercial vehicle including the vacant vehicle information. Information processing method including.

19. The information processing system A route to be presented is searched for based on the score of the route on the route, which is based on the passenger demand forecast data of the commercial vehicle. Including, The score is set based on the width of the road along the route. Information processing methods.

20. The information processing system A route to be presented is searched for based on the score of the route on the route, which is based on the passenger demand forecast data of the commercial vehicle. Including, The score for a location where it is necessary to cross an oncoming lane is set lower than the score for a location where it is not necessary to cross an oncoming lane. Information processing methods.

21. The information processing system A route to be presented is searched for based on the score of the route on the route, which is based on the passenger demand forecast data of the commercial vehicle. Including, The closer the predicted passenger direction is to the destination direction, the higher the score. Information processing methods.

22. The information processing system The route is based on the passenger demand forecast data for the commercial vehicle, and based on the score for the route, a route that passes through a location where passenger demand is predicted around the current location is searched for as a route to be presented. Information processing method including.

23. The information processing system A route to be presented is searched for based on the score of the route on the route, which is based on the passenger demand forecast data of the commercial vehicle. Including, The predicted time for predicting the demand for rides in a predetermined area when searching for the presented route is changed according to the distance from the current location. Information processing methods.

24. The information processing system A route is searched for as a presented route, in which the number of predicted passenger demand vehicles is equal to or greater than the number of vacant commercial vehicles, based on the score of the route on the route, which is based on the passenger demand forecast data of the commercial vehicles. Information processing method including.

25. The information processing system The route is based on the passenger demand forecast data for commercial vehicles, and based on the route score, a route on which no vacant commercial vehicles have passed for a certain period of time is searched for as a route to be presented. Information processing method including.

26. The information processing system includes: The route to be presented is searched by comparing the route scores of each route. The information processing method according to any one of claims 18 to 25.

27. The information processing system includes: The route to be presented is searched for by comparing the total score obtained by adding up the scores for each route along the route.

27. The information processing method according to claim 26.

28. The information processing system includes: A route with a high total score is searched for as the route to be presented.

28. The information processing method according to claim 27.

29. The total score is calculated by adding up the scores for each boarding point, which is a location where boarding demand is predicted.

28. The information processing method according to claim 27.

30. The score of the boarding point is set based on at least one of the date and time of boarding, the number of boardings, the total boarding time, the average boarding time, and the percentage of boardings whose boarding distance is equal to or greater than a predetermined distance at the boarding point.

30. The information processing method according to claim 29.

31. The total score is calculated by adding up the scores for each area into which the sales area of ​​the sales vehicle is divided.

28. The information processing method according to claim 27.

32. The information processing system, Controlling the display of a map including the proposed route The information processing method according to any one of claims 18 to 25, further comprising:

33. The presented route is a route recommended to the driver of the commercial vehicle. The information processing method according to any one of claims 18 to 25.

34. The commercial vehicle is a taxi The information processing method according to any one of claims 18 to 25.

35. The system includes a control unit that controls the display of a map including search results for a route to be presented based on the route score on the route, which is based on passenger demand forecast data for commercial vehicles, and on the operation data of commercial vehicles including vacant vehicle information. Information processing system.

36. The control unit compares the scores of the routes on each route and controls the display of a map including the search results of the route to be presented.

36. The information processing system according to claim 35.

37. The control unit compares the scores of the routes on the route with the total score obtained by adding up the scores of the routes for each route, and controls the display of the map.

37. The information processing system according to claim 36.

38. The information processing system A score based on passenger demand forecast data for commercial vehicles, and a map display including search results for routes to be presented is controlled based on the route score and operational data of commercial vehicles including vacant vehicle information. Information processing method including.

39. The information processing system includes: Compare the route scores for each route and control the display of the map containing the route search results.

39. The information processing method according to claim 38.

40. The display of the map is controlled by comparing a total score obtained by adding up the scores for each route along the route.

40. The information processing method according to claim 39.

Images