Management device, distribution update method, and computer program

Abstract

A device according to one aspect of the present disclosure is a management device that manages operation of a vehicle, said management device comprising a storage unit that stores operation results including a target cost required in order to pass along a route from a starting point position of a target section to an ending point position thereof, and distribution results including the existing frequency distribution of the target cost, and a control unit that executes a distribution update of the target cost, wherein the distribution update includes: a process for generating a prior distribution of the target cost from the operation results; a process for extracting the existing frequency distribution to be applied to the prior distribution from the distribution results; a process for calculating a posterior distribution of the target cost from the prior distribution and the existing frequency distribution; and a process for replacing the existing frequency distribution with the posterior distribution.

Description

Management device, distribution update method, and computer program

[0001] This application claims priority to Japanese Patent Application No. 2023-89479 filed on May 31, 2023, and incorporates by reference all of the contents of that application.

[0002] Patent Literature 1 describes an estimation device that estimates the costs of multiple elements (nodes and links) that make up a network corresponding to a piping network or a road network using a preset model. The model is expressed by multiple column vectors corresponding to spatial and temporal modes. The estimation device includes a data receiving unit and an element cost calculation unit.

[0003] The data accepting unit receives observation data and prior distribution data set for each column vector. The element cost calculating unit generates a conditional posterior distribution for each column vector based on the likelihood function, the observation data, and the prior distribution data. The element cost calculating unit extracts a column vector from the conditional posterior distribution of each column vector for each of multiple samplings, and calculates estimated values ​​of the costs of multiple elements using the extracted column vector and the above-mentioned model.

[0004] JP 2018-112884 A

[0005] An apparatus according to one embodiment of the present disclosure is a management apparatus for managing vehicle operations, and includes a memory unit that stores operation records including target costs required to travel a route from the start position to the end position of a target section, and distribution records including an existing frequency distribution of the target costs, and a control unit that performs a distribution update of the target costs, wherein the distribution update includes a process of generating a prior distribution of the target costs from the operation records, a process of extracting the existing frequency distribution to be applied to the prior distribution from the distribution records, a process of calculating a posterior distribution of the target costs from the prior distribution and the existing frequency distribution, and a process of replacing the existing frequency distribution with the posterior distribution.

[0006] FIG. 1 is a network configuration diagram showing an example of an operation management system. FIG. 2 is a route diagram showing an example of a target section and a target cost. FIG. 3 is a block diagram showing an example of the internal configuration of a management server. FIG. 4 is a block diagram showing an example of the internal configuration of a shared server. FIG. 5 is an explanatory diagram showing an example of the data structure of operation results. FIG. 6 is an explanatory diagram showing an example of the data structure of distribution results. FIG. 7 is an explanatory diagram showing an example of the data structure of input / output information. FIG. 8 is an explanatory diagram showing an example of the contents of a distribution update of target costs. FIG. 9 is an explanatory diagram showing an example of the contents of a target cost estimation process.

[0007] <Problem to be Solved by the Present Disclosure> Patent Literature 1 estimates the costs of network elements (nodes and links) using a preset model, which poses a problem that the estimation logic does not hold when the network configuration changes. Therefore, even if the accuracy of the conditional posterior distribution is improved by increasing the amount of observation data, only a posterior distribution with poor versatility is obtained. In view of this problem, a first object of the present disclosure is to provide a management device and the like that can improve the accuracy of versatile distribution results.

[0008] In Patent Document 1, as a measure to increase the amount of observation data and improve the accuracy of the conditional posterior distribution, it is considered that multiple businesses jointly collect observation data. However, the observation data in Patent Document 1 is highly confidential observation data observed by sensors installed for each predetermined number of elements, making it difficult to improve the distribution accuracy by opening up the data. In view of this problem, a second object of the present disclosure is to provide a management device etc. that can improve the distribution accuracy by opening up the data.

[0009] <Effects of the Present Disclosure> According to the present disclosure, it is possible to improve the accuracy of versatile distribution results. According to the present disclosure, it is possible to improve distribution accuracy by opening up data.

[0010] <Outline of Embodiments of the Present Disclosure> An outline of an embodiment of the present disclosure will be described below. (1) A device according to the present embodiment is a management device that manages vehicle operation, and includes a storage unit that stores operation records including a target cost required to travel a route from a start point position to an end point position of a target section and a distribution record including an existing frequency distribution of the target cost, and a control unit that executes a distribution update of the target cost, wherein the distribution update includes a process of generating a prior distribution of the target cost from the operation record, a process of extracting the existing frequency distribution to be applied to the prior distribution from the distribution record, a process of calculating a posterior distribution of the target cost from the prior distribution and the existing frequency distribution, and a process of replacing the existing frequency distribution with the posterior distribution.

[0011] According to the management device of this embodiment, the operation record includes the target cost required to travel the route from the start point to the end point of the target section, and the distribution record includes a pre-existing frequency distribution of the target cost. Since the target cost does not depend on the road network configuration, the pre-existing frequency distribution of the target cost is also versatile data that does not depend on the road network configuration. Therefore, by performing the above-mentioned distribution update on the pre-existing frequency distribution of the target cost, the accuracy of the versatile distribution record can be improved. Thus, the first objective described above is achieved.

[0012] (2) In the estimation device of (1) above, the operation record and the distribution record may include an operation status indicating the state of the driving environment when traveling along the route of the target section, and the operation status may include the vehicle type of the vehicle.

[0013] According to the management device of this embodiment, the operation record and distribution record include operation status indicating the state of the driving environment when traveling along the route of the target section, and the operation status includes the vehicle model. Vehicle model is not particularly confidential information and can be said to be information suitable for opening up data. Therefore, if multiple businesses manage operation record and distribution record including vehicle model as operation status, it is possible to improve distribution accuracy by opening up data. Thus, the second objective described above is achieved.

[0014] (3) In the management device of (2) described above, the operation status may include an operation mode that indicates the mode of vehicle operation when traveling along the route of the target section. This is because the operation mode that indicates the mode of vehicle operation is not particularly confidential information and can be said to be information suitable for making data open. In addition, if operators with the same or similar operation modes share their operation records and distribution records, there is an advantage in that the accuracy of the existing frequency distribution can be further improved compared to when operators with different operation modes are included.

[0015] (4) In the management device of (2) or (3) above, the operation status may include an operation area that indicates a type of geographical area that includes the route of the target section. This is because the operation area that indicates a geographical area that includes the route of the target section is not particularly confidential information and can be said to be information that is suitable for making data open.

[0016] (5) In any of the management devices described in (2) to (4), the operation status may include a road type ratio that represents the ratio of the travel distance of a predetermined type of road traveled by the vehicle divided by the travel distance of the route in the target section. This is because the ratio of the road types that make up the route in the target section is not particularly confidential information and can be said to be information suitable for making data open.

[0017] (6) In any of the management devices described in (2) to (5) above, the operation status may include a day type when the target section of the route is traveled. The reason for this is that the day type when the target section of the route is traveled is not particularly confidential information and can be said to be information suitable for making the data open.

[0018] (7) In any of the management devices described above in (1) to (6), the control unit may be capable of executing a process of estimating the target cost, and the estimation process may include a process of determining an estimated value of the target cost based on a representative point of the existing frequency distribution extracted from the distribution history. In this way, the target cost can be easily estimated using the existing frequency distribution extracted from the distribution history.

[0019] (8) In any of the management devices described above in (1) to (7), the control unit may determine whether to use the prior distribution for the distribution update based on the result of accuracy evaluation of the prior distribution. In this way, by not using a prior distribution for which the result of accuracy evaluation is negative, it is possible to prevent the existing frequency distribution from being updated by a prior distribution with low accuracy.

[0020] (9) In any of the management devices described in (1) to (8) above, if the management device further includes a communication unit that communicates with a shared server operated jointly with other businesses, the communication unit may transmit the distribution results including the existing frequency distribution for which the distribution update has been performed to the shared server. In this way, the distribution results including the existing frequency distribution can be jointly managed by the shared server operated jointly with other businesses.

[0021] (10) In the management device of (9) above, when the control unit cannot extract the existing frequency distribution to be applied to the prior distribution from the distribution history, the control unit may cause the communication unit to send a distribution request requesting the existing frequency distribution from the shared server. In this way, even if the management server does not have an existing frequency distribution, it is possible to perform distribution update using an existing frequency distribution obtained from the shared server.

[0022] (11) The method according to this embodiment is a distribution update method executed by any one of the management devices described above in (1) to (10). Therefore, the distribution update method according to this embodiment has the same effects as any one of the management devices described above in (1) to (10).

[0023] (12) The computer program according to this embodiment is a program that causes a computer to function as any one of the management devices described above in (1) to (10). Therefore, the computer program according to this embodiment has the same effects as any one of the management devices described above in (1) to (10).

[0024] <Details of the embodiments of the present disclosure> Hereinafter, details of the embodiments of the present disclosure will be described with reference to the drawings. Note that at least some of the embodiments described below may be combined in any manner. [Definition of Terms] Before describing the details of the embodiments of the present disclosure, the main terms will first be defined.

[0025] "Vehicle": refers to any vehicle traveling on a road. Vehicles are primarily four-wheeled vehicles, but may also be other types of vehicles such as two-wheeled vehicles. The drive system of the vehicle is not limited to internal combustion engines, and electric vehicles and hybrid cars are also included in the vehicle. The vehicle may also be an autonomous vehicle. In this embodiment, when simply referring to a "vehicle," it includes both a probe vehicle that transmits probe information and a vehicle that does not transmit probe information.

[0026] "Transportation vehicle": A vehicle that travels on roads carrying cargo or passengers. Transportation vehicles include commercial trucks and commercial buses. Commercial trucks include, for example, small, medium, and large trucks that carry cargo. Commercial buses include, for example, route buses that operate on designated routes within cities or suburbs, express buses that operate between cities, and tourist buses that carry tourists.

[0027] "Probe information" refers to information that includes various data about a vehicle sensed by a probe vehicle traveling on a road. Probe information is also called probe data or floating car data. Probe information includes vehicle attribute data such as the probe vehicle's identification information, traveling position, traveling speed, traveling direction, and the time of occurrence of these (travel time).

[0028] "Probe vehicle": A vehicle that senses its own vehicle attribute data and transmits probe information including the sensed data to the outside. The probe vehicle in this embodiment is primarily the above-mentioned transportation vehicle. However, a general passenger vehicle that does not perform transportation operations can also be a probe vehicle.

[0029] "Target section": In the operation management of transport vehicles, this refers to a road section in which the starting point (hereinafter referred to as "first point") and the ending point (hereinafter referred to as "second point") are predefined. Specifically, the target section refers to a road section made up of multiple routes when multiple transport vehicles adopt a route that passes through the first point and the second point. In this case, the first point can be the departure point or relay point for the transport vehicles. Furthermore, the second point can be the relay point or arrival point for the transport vehicles.

[0030] "Target cost": The cost required for a transportation vehicle to travel the route from the first point to the second point of the target section. Specifically, the target cost is a cost that the operations manager has adopted as a management target among various costs required to travel the target section. The target cost may be, for example, the travel time (also called "travel time") (minutes) and travel distance (km) for the target section. If the transportation vehicle is an electric vehicle, the power consumption (kWh) required to travel the target section may also be adopted as the target cost.

[0031] [Overall System Configuration] Fig. 1 is a network configuration diagram showing an example of a traffic management system 100. As shown in Fig. 1, the traffic management system 100 of this embodiment includes a management server 1, a shared server 2, and a transportation vehicle 3 that communicates with the management server 1. The management server 1 is a collective term for multiple types of servers 1A, 1B, and 1C. Items common to these servers 1A, 1B, and 1C will be described in terms of the management server 1.

[0032] The transportation vehicle 3 is a general term for multiple types of vehicles 3A, 3B, and 3C. Common matters among these vehicles 3A, 3B, and 3C are described in terms of the transportation vehicle 3. The management server 1 is a computer (management device) that collects information about the transportation vehicle 3 and manages its operation. The management server 1 is operated by a transportation company or an IT (Information Technology) company. The management server 1 may be either an on-premise server or a cloud server.

[0033] The management server 1 includes the following multiple types of servers 1A, 1B, and 1C. The first operator, second operator, and third operator are different corporations that operate transportation or passenger services. The management server 1 may be of two types or four or more types depending on the number of operators. First server 1A: Management server 1 operated by the operations manager 4A of the first operator Second server 1B: Management server 1 operated by the operations manager 4B of the second operator Third server 1C: Management server 1 operated by the operations manager 4C of the third operator

[0034] The transport vehicles 3 include the following multiple types of vehicles 3A, 3B, and 3C. The number of types of transport vehicles 3 may be two or four or more depending on the number of businesses. First vehicle 3A: Transport vehicle 3 managed by the first server 1A Second vehicle 3B: Transport vehicle 3 managed by the second server 1B Third vehicle 3C: Transport vehicle 3 managed by the third server 1C

[0035] The transportation vehicle 3 has an on-board communication device that communicates wirelessly with wireless base stations 5 (e.g., mobile base stations) in various locations. The wireless base stations 5 can communicate with the management server 1 via a public network 6, which includes a mobile communication core network and the Internet. Therefore, the first server 1A communicates with the first vehicle 3A that is a management target, but does not communicate with the second vehicle 3B or the third vehicle 3C. Similarly, the second server 1B communicates with the second vehicle 3B that is a management target, but does not communicate with the first vehicle 3A or the third vehicle 3C. The same is true for the third server 1C.

[0036] The transport vehicle 3 is a probe vehicle. Therefore, the transport vehicle 3 wirelessly transmits a communication packet addressed to the management server 1, including uplink data, to the wireless base station 5. The uplink data includes probe information S1 sensed by the vehicle itself. The management server 1 collects the probe information S1 from the transport vehicles 3 under management, and individually manages the operation of the transport vehicles 3 under management using the collected probe information S1.

[0037] Specifically, the first server 1A independently manages operation based on the probe information S1 of one or more first vehicles 3A that it manages. Similarly, the second server 1B independently manages operation based on the probe information S1 of one or more second vehicles 3B that it manages, and the third server 1C independently manages operation based on the probe information S1 of one or more third vehicles 3C that it manages.

[0038] The shared server 2 is a computer (shared device) jointly operated by the businesses that operate the first server 1A, the second server 1B, and the third server 1C. The shared server 2 may be either an on-premise server or a cloud server. The shared server 2 stores predetermined common information agreed upon by the first business, the second business, and the third business. The common information includes, for example, "distribution results" ( FIG. 6 ) that include existing frequency distributions of target costs (travel time, travel distance, power consumption, etc.).

[0039] [Specific Examples of Target Sections and Target Costs] Figure 2 is a route diagram showing an example of a target section TR and a target cost TC. As shown in Figure 2, the target section TR is a road section that includes at least one route that the driver can select when the transportation vehicle 3 travels from a first point J1 to a second point J2. The first point J1 and the second point J2 can be set at, for example, a large intersection, an interchange on a highway, an entrance or exit of a highway, etc.

[0040] For example, assume that three types of transport vehicles 3A, 3B, and 3C take the following routes for transport work from a predetermined starting point to a destination point. Route of first vehicle 3A: Point a1 → Point a2 → Point a3 → Point a4 → Point a5 Route of second vehicle 3B: Point b1 → Point b2 → Point b3 → Point b4 → Point b5 Route of third vehicle 3C: Point c1 → Point c2 → Point c3 → Point c4 → Point c5 → Point c6

[0041] Furthermore, it is assumed that points a2, b2, and c2 are the same as the first point J1, and points a4, b4, and c5 are the same as the second point J2. In this case, the first server 1A determines that the route R1 traveled by the first vehicle 3A from the first point J1 to the second point J2 is the route of the target section TR. In other words, the first server 1A determines that the route R1 that follows "point a2 (J1) → point a3 → point a4 (J2)" is the route of the target section TR taken by the first vehicle 3A.

[0042] Similarly, the second server 1B determines that the route R2, which is the route taken by the second vehicle 3B from the first point J1 to the second point J2, that is, "point b2 (J1) → point b3 → point b3 (J2)," is the route of the target section TR for the second vehicle 3B. Furthermore, the third server 1C determines that the route R3, which is the route taken by the third vehicle 3C from the first point J1 to the second point J2, that is, "point c2 (J1) → point c3 → point c4 → point c5 (J2)," is the route of the target section TR for the third vehicle 3C.

[0043] 2, the target cost TC includes target costs TC1, TC2, and TC3 calculated individually by each of the servers 1A, 1B, and 1C. That is, when the first server 1A determines the route R1 of the target section TR for the first vehicle 3A, it calculates the target cost TC1 required to travel the route R1. Specifically, the first server 1A calculates, as the target cost TC1, at least one of the travel time, travel distance, and power consumption required to travel the route R1.

[0044] Similarly, when the second server 1B determines a route R2 of the target section TR for the second vehicle 3B, it calculates a target cost TC2 (at least one of travel time, travel distance, and power consumption) required to travel the route R2. Furthermore, when the third server 1C determines a route R3 of the target section TR for the third vehicle 3C, it calculates a target cost TC3 (at least one of travel time, travel distance, and power consumption) required to travel the route R3.

[0045] The routes R1, R2, and R3 are determined, for example, by map matching between time-series data of the travel position of the transport vehicle 3 and digital road map data. The travel time (minutes) is calculated from the travel time included in the probe information S1 of the transport vehicle 3. Specifically, the travel time can be calculated, for example, as the difference between the passing time of the first point J1 and the passing time of the second point J2.

[0046] The travel distance (km) is, for example, the integrated value of the lengths of the road links included in the routes R1, R2, and R3. The power consumption is, for example, the travel distance multiplied by the electricity cost (kWh / km) of the transport vehicle 3. In this way, the calculation of the target costs TC1, TC2, and TC3 is performed based on the probe information S1 of the transport vehicle 3. This calculation process is specifically performed by the control unit 11 (see FIG. 3) of each of the servers 1A, 1B, and 1C.

[0047] As described above, the target costs TC1, TC2, and TC3 are costs related to the routes R1, R2, and R3 when the transport vehicles 3A, 3B, and 3C pass through the first point J1 and the second point J2 of the target section TR, and therefore, unlike the number of vehicles and vehicle speeds measured by vehicle detectors, for example, these are data that do not depend on the road network configuration. Therefore, the frequency distribution of the target costs TC1, TC2, and TC3 is also versatile data in that it does not depend on the road network configuration.

[0048] [Configuration Example of Management Server] Fig. 3 is a block diagram showing an example of the internal configuration of the management server 1. As shown in Fig. 3, the management server 1 (servers 1A, 1B, 1C) includes a control unit 11, a storage unit 12, a communication unit 13, and multiple types of databases (DBs) 14, 15, 16. The multiple types of databases 14, 15, 16 are electronic data constructed in a predetermined data arrangement in the storage unit 12. However, some or all of the databases 14, 15, 16 may be constructed in an external storage device (not shown) connected to the management server 1.

[0049] The control unit 11 is an arithmetic processing device including a CPU (Central Processing Unit) and RAM (Random Access Memory). The control unit 11 may include an integrated circuit such as an FPGA (Field-Programmable Gate Array). The control unit 11 reads a computer program 17 stored in the storage unit 12 into the main memory (RAM) and executes various information processing in accordance with the program 17. This information processing includes determining the route of the target section TR and calculating the target cost TC.

[0050] The storage unit 12 is an auxiliary storage device including a nonvolatile memory such as a hard disk drive (HDD) and a solid state drive (SSD). The storage unit 12 may also include a flash read-only memory (ROM), a universal serial bus (USB) memory, an SD card, or the like.

[0051] The communication unit 13 is a communication interface capable of communication via the public network 6. When the communication unit 13 receives a communication packet including probe information S1 of the transportation vehicle 3 from the wireless base station 5, it outputs the received probe information S1 to the control unit 11. The control unit 11 uses the probe information S1 of the transportation vehicle 3 input from the communication unit 13 for route determination of the target section TR, calculation of the target cost TC, and the like.

[0052] The multiple types of databases 14, 15, 16 include an operation record database 14, a distribution record database 15, and an input / output information database 16. The operation record database 14 stores "operation record PR" (Fig. 5). The distribution record database 15 stores "distribution record DR" (Fig. 6). The input / output information database 16 stores "input / output information IO" (Fig. 7). The operation record PR and distribution record DR are used to update the distribution of the target cost TC. The input / output information IO is used in the estimation process of the target cost TC.

[0053] When the communication unit 13 receives a distribution request message from the shared server 2, it outputs the received division request message to the control unit 11. In response to the input distribution request, the control unit 11 extracts the distribution actual result DR ( FIG. 6 ) of the target cost TC from the distribution actual result database 15, and causes the communication unit 13 to transmit a distribution response message including the distribution actual result DR. As a result, the distribution actual result DR of the target cost TC held by the management server 1 is transferred to the shared server 2.

[0054] The computer program 17 stored in the storage unit 12 includes a program that causes the control unit 11 to execute distribution update of the target cost TC. The distribution update of the target cost TC is a process that includes, for example, the following steps: Step ST11: Generate prior distribution Step ST12: Extract distribution Step ST13: Calculate posterior distribution Step ST14: Replace distribution

[0055] The computer program 17 stored in the storage unit 12 includes a program that causes the control unit 11 to execute a process for estimating the target cost TC. The process for estimating the target cost TC includes, for example, the following steps: Step ST21: Input of information Step ST22: Cost estimation Step ST23: Cost output

[0056] Below, the data structures of "operation performance PR" and "distribution performance DR" necessary for updating the distribution of the target cost TC will be explained, followed by an example of the contents of the distribution update of the target cost TC (steps ST11 to ST14) (FIG. 8). Similarly, the data structure of "input / output information IO" necessary for the estimation process of the target cost TC will be explained, followed by an example of the contents of the estimation process of the target cost TC (steps ST21 to ST23) (FIG. 9).

[0057] [Configuration example of operation record] Figure 5 is an explanatory diagram showing an example of the data structure of the operation record PR. As shown in Figure 5, the operation record PR is a two-dimensional array of data including operation status Aij and target cost TCij. The row-wise variable i of the two-dimensional array is an identifier of the data content of the operation status Aij and target cost TCij. The column-wise variable j is an identifier of the type of operation status Aij and target cost TCij.

[0058] The operation status Aij is data representing the state of the driving environment when the transport vehicle 3 travels along the route of the target section TR. The driving environment may be either the internal environment or the external environment of the vehicle. The type (variable j) of the operation status Aij may be, for example, "vehicle type" (j = 0), "ratio of road types" (j = 1 to 12), "weather" (j = 21), "time period" (j = 22), "day type" (j = 23), "operation area" (j = 24), "operation mode" (j = 25), etc. The value of the type variable j may be set arbitrarily.

[0059] "Vehicle type" (j=0) represents the vehicle type of the transportation vehicle 3. The data content of the vehicle type may be, for example, "large vehicle," "medium vehicle," "compact vehicle," "luggage carrier," and "passenger vehicle."

[0060] The "ratio of road type" (j = 1 to 12) is the ratio obtained by dividing the travel distance on a road of a predetermined type by the travel distance of the route in the target section TR. The data content of the ratio is standardized to a standardized value, such as a rounded value to the first decimal place. The travel distance on a road of a predetermined type is determined by map matching between the time-series data of the travel position of the transport vehicle 3 and the digital road map data including the road type.

[0061] The road type may be, for example, one of the following: j=1: unsurveyed road j=2: national expressway j=3: urban expressway (including designated urban expressway) j=4: general national road j=5: major local road (prefectural road) j=6: major local road (designated city road)

[0062] j = 7: General prefectural road j = 8: General city road in designated city j = 9: Road with common name j = 10: Road other than the above j = 11: Toll road j = 12: Free road

[0063] "Weather" (j=21) indicates the type of weather when the transport vehicle 3 traveled along the route of the target section TR. The weather data content is determined, for example, by obtaining from the information providing server the weather at the travel location and travel time identified from the probe information S1. The weather data content may include, for example, "sunny," "cloudy," "rain," "heavy rain," "snow," "storm," and "frozen roads."

[0064] The "time period" (j=22) indicates the type of time period when the transport vehicle 3 traveled along the route of the target section TR. The type of time period is determined, for example, from the travel time of the transport vehicle 3 identified from the probe information S1. The data content of the time period may be, for example, "morning" (6:00 to 12:00), "afternoon" (12:00 to 18:00), and "night" (18:00 to 22:00), etc.

[0065] "Day type" (j=23) indicates the type of day when the transport vehicle 3 traveled the route of the target section TR. The data content of the day type is determined, for example, from the travel time of the transport vehicle 3 identified from the probe information S1. The data content of the day type may be, for example, "day of the week" (Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday), "weekday", and "holiday".

[0066] The "operation area" (j=24) represents the type of geographical area that includes the route of the target section TR. The type of operation area is determined by map matching between the time-series data of the traveling position of the transportation vehicle 3 and the digital road map data that includes the type of geographical area. The data content of the operation area may be, for example, "Osaka city," "Osaka suburbs," "Kobe city," and "Hyogo suburbs."

[0067] "Operation mode" (j=25) represents the mode of vehicle operation when the transport vehicle 3 travels along the route of the target section TR. The type of operation mode is determined based on the content of the transport work notified by the transport vehicle 3. The data content of the operation mode may include, for example, "urban operation," "long-distance operation," and "regular operation."

[0068] The target cost TCij is data representing the observed value of the target cost TC (travel time, travel distance, or power consumption) when the transport vehicle 3 passes through the target section TR. The method for calculating the target cost TC is as described above. The type (variable j) of the target cost TCij includes, for example, at least one of "travel time" (j=51), "travel distance" (j=52), and "power consumption" (j=53).

[0069] Hereinafter, the target cost TCi51 (travel time) may be abbreviated as “travel time Xi.” Similarly, the target cost TCi52 (travel distance) may be abbreviated as “travel distance Yi,” and the target cost TCi53 (power consumption) may be abbreviated as “power consumption Zi.”

[0070] [Configuration example of distribution record] Figure 6 is an explanatory diagram showing an example of the data structure of the distribution record DR. As shown in Figure 6, the distribution record DR is a two-dimensional array of data including operation status Akj and existing frequency distribution Fkj. The variable k in the row direction of the two-dimensional array is an identifier for the combination of data contents of Aj where Aij = Aj holds in the operation record PR (Figure 5). Hereinafter, the variable k will also be referred to as the "combination variable k." The variable j in the column direction is an identifier for the type of operation status Akj and frequency distribution Fkj.

[0071] The above type identifiers (variable j) are formatted in the same way as the type identifiers (variable j) of the operation status Aij and target cost TCij that make up the operation record PR (Figure 5). Therefore, frequency distribution Fk51 is the frequency distribution of target cost TCi51 (travel time Xi). Similarly, frequency distribution Fk52 is the frequency distribution of target cost TCi52 (travel distance Yi), and frequency distribution Fk53 is the frequency distribution of target cost TCi53 (power consumption Zi).

[0072] For example, the frequency distribution F1,51 of the travel time Xi is the frequency distribution of the travel time Xi when the combination variable k=1 and Aij=Aj in the operation record PR (FIG. 5). Similarly, the frequency distribution F2,51 of the travel time Xi is the frequency distribution of the travel time Xi when the combination variable k=2 and Aij=Aj in the operation record PR (FIG. 5). The same is true for the frequency distribution F3,51 of the travel time Xi.

[0073] The frequency distribution F1,52 of the travel distance Yi is the frequency distribution of the travel distance Yi when the combination variable k=1 and Aij=Aj in the operation record PR (FIG. 5). Similarly, the frequency distribution F2,52 of the travel distance Yi is the frequency distribution of the travel distance Yi when the combination variable k=2 and Aij=Aj in the operation record PR (FIG. 5). The same is true for the frequency distribution F3,52 of the travel distance Yi.

[0074] The frequency distribution F1, 53 of the power consumption Zi is the frequency distribution of the power consumption Zi when the combination variable k = 1 and Aij = Aj in the operation record PR (Fig. 5). Similarly, the frequency distribution F2, 53 of the power consumption Zi is the frequency distribution of the power consumption Zi when the combination variable k = 2 and Aij = Aj in the operation record PR (Fig. 5). The same is true for the frequency distribution F3, 53 of the power consumption Zi.

[0075] In Figure 6, the data format of the existing frequency distribution Fkj is a histogram, but the data format may also be a table. Furthermore, the frequencies of the frequency distribution Fkj may be absolute frequencies, relative frequencies, cumulative absolute frequencies, or cumulative relative frequencies. Figure 6 illustrates an example of the existing frequency distribution Fkj after at least one distribution update (Figure 8). However, for the frequency distribution Fkj at the beginning of operation, before any distribution update has been performed, an artificial distribution with a constant frequency may be used, for example.

[0076] [Configuration example of input / output information] Figure 7 is an explanatory diagram showing an example of the data structure of input / output information IO. As shown in Figure 7, the input / output information IO is a one-dimensional array of data including an operation status Aj, which is input information, and a target cost TCj, which is output information. The variable j in the one-dimensional array is an identifier of the type of operation status Aj and the target cost TCj.

[0077] The above type identifier (variable j) is set to the same format as the type identifier (variable j) of the operation status Akj and the frequency distribution Fkj of the target cost in the distribution record DR (FIG. 6). The data content to be input to the operation status Aj is determined, for example, based on the user's operational input and the test route. Specifically, the data content of "vehicle type" and "operation mode" is determined by the user's operational input, such as the operation managers 4A, 4B, and 4C.

[0078] The test route is, for example, a route for the transport vehicle 3 that is generated by a route search that is experimentally executed by the control unit 11. This route search is executed under the condition that the transport vehicle 3 passes through a first point J1 and a second point J2. After determining the route for the transport vehicle 3 by the route search, the control unit 11 determines the data contents to be stored in the "ratio of road types," "weather," "time period," "day of the week," and "operation area" included in the operation status Aj based on the determined route.

[0079] The target cost TCj, which is output information, includes at least one of travel time X, travel distance Y, and power consumption X. The estimated value of the cost to be output in the target cost TCj column is determined by the target cost TCj estimation process ( FIG. 9 ), which will be described later.

[0080] [Contents of Target Cost Distribution Update] FIG. 8 is an explanatory diagram showing an example of the contents of distribution management of the target cost TC executed by the control unit 11 of the management server 1. The prior distribution Fc, existing frequency distribution Fkj, and posterior distribution Fa in FIG. 8 are frequency distributions of the target cost TCij. The target cost TCij is at least one of travel distance Xi, travel time Yi, and power consumption Zi. The frequency may be an absolute frequency or a relative frequency. The frequency distribution may be in the form of a histogram or a table.

[0081] 8, the control unit 11 first generates a prior distribution (step ST11). The prior distribution generation is a process of generating a prior distribution Fc of the target cost TCi from the operation records PR(Aij, TCij) for a predetermined period (e.g., one month) stored in the operation record database 14. Specifically, the prior distribution generation includes the following processes.

[0082] Process 1: Sort Aij of the operation record PR in the row direction i, and number the combination variable k of the data content of Aj for which Aij = Aj holds. For example, let k = 1, 2...N. Process 2: Extract the data of the target cost TCij belonging to the combination k = 1. Process 3: Create a frequency distribution with TCij on the horizontal axis and the frequency of TCij on the vertical axis. Process 4: The created frequency distribution is used as the prior distribution Fc for k = 1. Process 5: Repeat processes 2 to 4 until k = N.

[0083] Next, the control unit 11 performs distribution extraction (step ST12). Distribution extraction is a process of extracting a pre-existing frequency distribution Fkj corresponding to the combination variable k from the distribution actual data DR(Akj, Fkj) stored in the distribution actual data database 15. Specifically, the control unit 11 searches for a distribution actual data DR(Akj, Fkj) whose data content of the combination variable k matches the data content of the operation status Akj, and determines the frequency distribution Fkj of the found distribution actual data DR as the pre-existing frequency distribution Fkj to be extracted.

[0084] In the distribution extraction (step ST12), if an existing frequency distribution Fkj corresponding to the combination variable k cannot be extracted from its own distribution record database 15, the control unit 11 may request the shared server 2 to extract the frequency distribution Fkj. Specifically, the control unit 11 may send a distribution request message including the data content of the operation status Aij of the combination variable k to the shared server 2.

[0085] In this case, if the shared server 2 can extract a pre-existing frequency distribution Fkj corresponding to the data content of the operation status Aij of the combination variable k from its own distribution history database 24, it transmits a distribution response control message including the frequency distribution Fkj to the management server 1. If the shared server 2 cannot extract the pre-existing frequency distribution Fkj from its own distribution history database 24, it transmits a message to the management server 1 indicating that extraction is not possible.

[0086] Next, the control unit 11 performs a posterior distribution calculation (step ST13). The posterior distribution calculation is a process of calculating the posterior distribution Fa based on the prior distribution Fc and the existing frequency distribution Fkj in the distribution history DB. Specifically, the control unit 11 calculates the posterior distribution Fa using, for example, an estimation formula of Bayes' theorem that applies the existing frequency distribution Fkj to the prior distribution Fc as a likelihood function. However, the estimation formula is not limited to Bayes' theorem, and may be an estimation formula based on a neural network or the like.

[0087] Next, the control unit 11 performs distribution replacement (step ST14). Distribution replacement is a process of replacing the existing frequency distribution Fkj extracted in step S12 with the posterior distribution Fa calculated in step ST13. As a result, the existing frequency distribution Fkj constituting the distribution actual result DR is updated to a frequency distribution that takes into account the trend of the prior distribution Fc for the most recent predetermined period. This makes it possible to improve the accuracy of the existing frequency distribution Fkj constituting the distribution actual result DR.

[0088] [Contents of the Target Cost Estimation Process] Figure 9 is an explanatory diagram showing an example of the contents of the target cost TC estimation process executed by the control unit 11 of the management server 1. As shown in Figure 9, the control unit 11 first performs information input (step ST21). The information input is a process of determining the data contents to be input to the operation status Aj from the input / output information IO (Aj, TCj) stored in the input / output information database 16.

[0089] As described above, the data contents of "vehicle type" and "operation mode" are determined by user input, and the data contents of "road type ratio," "weather," "time period," "day of the week," and "operation area" are determined based on a test route generated experimentally. Next, the control unit 11 performs cost estimation (step ST22). The cost estimation includes a process of extracting a known frequency distribution Fkj from the distribution history database 15 and a process of estimating a suitable target cost TC using the known frequency distribution Fkj.

[0090] Specifically, the control unit 11 extracts a frequency distribution Fkj from the distribution record DR(Akj, Fkj) included in the distribution record database 15, where the data content of the operational status Akj matches the operational status Aj of the input / output information IO. Then, the control unit 11 estimates the target cost TC as the cost value whose frequency of the frequency distribution Fkj corresponds to a predetermined feature point. The predetermined feature point may be, for example, the mode, mean, median, or maximum or minimum value of the convex portion of the frequency distribution Fkj.

[0091] Next, the control unit 11 performs cost output (step ST23). Cost output is a process of storing the estimated value of the target cost TC calculated in step ST22 in the target cost TCj, which is output information of the input / output information IO. When the estimated value is stored in the target cost TCj of the input / output information IO, the control unit 11 presents the estimated value of the target cost TCj to the user, for example, by displaying it on a screen.

[0092] [Example of Shared Server Configuration] Fig. 4 is a block diagram showing an example of the internal configuration of shared server 2. As shown in Fig. 4, shared server 2 includes control unit 21, storage unit 22, communication unit 23, and distribution performance database (DB) 24. Distribution performance database 24 is electronic data constructed in a predetermined data array in storage unit 12. However, part or all of database 24 may be constructed in an external storage device (not shown) connected to management server 1.

[0093] The control unit 21 is an arithmetic processing device including a CPU, RAM, etc. The control unit 11 may include an integrated circuit such as an FPGA. The control unit 21 reads a computer program 25 stored in the storage unit 12 into the main memory (RAM) and executes various information processing operations in accordance with the program 25. This information processing operation includes updating and managing the distribution performance DR.

[0094] The storage unit 22 is an auxiliary storage device including a non-volatile memory such as an HDD and an SSD. The storage unit 22 may also include a flash ROM, a USB memory, an SD card, etc. The communication unit 23 is a communication interface for communicating with the management server 1 (servers 1A, 1B, 1C) via the public network 6.

[0095] The control unit 21 receives a distribution result DR including an existing frequency distribution Fkj for which distribution update (FIG. 8) has been performed from each of the servers 1A, 1B, and 1C by exchanging distribution request and distribution response messages with each of the servers 1A, 1B, and 1C. The control unit 21 stores the latest distribution result DR of the received distribution result DR in its own distribution result database 24. As a result, of the distribution result DRs managed individually by the three servers 1A, 1B, and 1C, the latest distribution result DR is held by the shared server 2.

[0096] On the other hand, when the distribution result database 15 of each server 1A, 1B, 1C does not contain a distribution result DR or when the number of times distribution updates (FIG. 8) have been performed is small, the control unit 11 of each server 1A, 1B, 1C may request transmission of the distribution result DR from the shared server 2. In this way, each server 1A, 1B, 1C can obtain the distribution result DR including the accurate frequency distribution Fkj that has been updated by other companies from the start of operation of the distribution result DR.

[0097] [First Modification: Accuracy Evaluation of Prior Distribution] In the above-described embodiment, the control unit 11 of the management server 1 may evaluate the accuracy of the prior distribution Fc before performing the distribution update ( FIG. 8 ), and determine whether or not to use the prior distribution Fc for the distribution update based on the results of the accuracy evaluation. In this way, it is possible to prevent the existing frequency distribution Fjk from being updated by an inaccurate prior distribution Fa.

[0098] The accuracy evaluation can be performed, for example, by the following processes. Process 1: Execute the estimation process of the target cost TC (FIG. 9) using the prior distribution Fa as reference information instead of the existing frequency distribution Fkj. Process 2: Determine whether the target cost TCj (e.g., at least one of travel time, travel distance, and power consumption) output by the estimation process of process 1 is within a predetermined numerical range. Process 3: If the determination result of process 2 is positive, use the prior distribution Fa for distribution update. Process 4: If the determination result of process 2 is negative, do not use the prior distribution Fa for distribution update.

[0099] [Second Modification: Another Example of Shared Server] In the above embodiment, the case where shared server 2 is a server separate from management server 1 has been exemplified, but any one of first server 1A, second server 1B, and third server 1C, which are management servers 1 of each company, may serve as shared server 2. For example, in the case where an agreement is made such that first server 1A of a first business operator serves as shared server 2, second server 1B and third server 1C may transmit and receive distribution request and distribution response messages to first server 1A (shared server 2).

[0100] [Other Modifications] The embodiments disclosed herein are illustrative in all respects and are not restrictive. The scope of the present invention is not limited to the above-described embodiments, and includes all modifications within the scope of equivalents to the configurations described in the claims. In the above-described embodiments, the vehicle is a transportation vehicle, but the techniques disclosed herein, such as the distribution update and estimation processing of the target cost, can also be adopted for vehicles other than transportation vehicles.

[0101] REFERENCE SIGNS LIST 1 Management server (management device) 1A First server 1B Second server 1C Third server 2 Shared server (shared device) 3 Transport vehicle 3A First vehicle 3B Second vehicle 3C Third vehicle 4A Dispatch manager 4B Dispatch manager 4C Dispatch manager 5 Wireless base station 6 Public network 11 Control unit 12 Memory unit 13 Communication unit 14 Operation record database 15 Distribution record database 16 Input / output information database 17 Computer program 21 Control unit 22 Memory unit 23 Communication unit 24 Distribution record database 25 Computer program 100 Operation management system TR Target section TC Target cost PR (Aij, TCij) Operation record Aij Operation status TCij Target cost DR (Akj, Fkj) Distribution record Akj Operation status Fkj Existing frequency distribution

Claims

1. A management device for managing vehicle operation, a storage unit that stores operation records including target costs required to travel a route from a start point position to an end point position of a target section, and distribution records including a pre-existing frequency distribution of the target costs; a control unit that executes the distribution update of the target costs, The distribution update may be A process of generating a prior distribution of the target cost from the operation record; A process of extracting the existing frequency distribution to be applied to the prior distribution from the distribution record; A process of calculating a posterior distribution of the target cost from the prior distribution and the existing frequency distribution; and a process of replacing the existing frequency distribution with the posterior distribution.

2. The operation record and the distribution record are The driving status indicates the driving environment when traveling along the route of the target section, The operation status is as follows: The management device of claim 1 , further comprising a vehicle model.

3. The operation status is as follows: The management device according to claim 2 , further comprising an operation mode indicating a mode of vehicle operation when traveling along the route of the target section.

4. The operation status is as follows: The management device according to claim 2 , further comprising a service area indicating a type of geographical area that includes the route of the target section.

5. The operation status is as follows: The management device according to claim 2 , further comprising a road type ratio representing a ratio obtained by dividing a travel distance of a predetermined type of road traveled by the vehicle by a travel distance of the route of the target section.

6. The operation status is as follows: The management device according to claim 2 , further comprising a day type when the target section of the route is traveled.

7. The control unit The target cost estimation process is executable, The estimation process includes: The management device according to claim 1 , further comprising a process of determining an estimated value of the target cost based on a representative point of the existing frequency distribution extracted from the distribution results.

8. The control unit The management device according to claim 1 , further comprising: a determining unit configured to determine whether or not to use the prior distribution for the distribution update in accordance with a result of accuracy evaluation of the prior distribution.

9. Further comprising a communication unit for communicating with a shared server jointly operated with other businesses, The communication unit The management device according to claim 1 , further comprising: transmitting the distribution results including the existing frequency distribution for which the distribution update has been performed to the shared server.

10. The control unit 10. The management device according to claim 9, wherein, when the existing frequency distribution to be applied to the prior distribution cannot be extracted from the distribution record, the management device causes the communication unit to transmit a distribution request requesting the existing frequency distribution from the shared server.

11. A distribution update method executed by a management device that manages vehicle operation, comprising: a step of storing an operation record including a target cost required to travel a route from a start point position to an end point position of a target section, and a distribution record including a pre-existing frequency distribution of the target cost; performing a distribution update of the target costs; The distribution update may be A process of generating a prior distribution of the target cost from the operation record; A process of extracting the existing frequency distribution to be applied to the prior distribution from the distribution record; A process of calculating a posterior distribution of the target cost from the prior distribution and the existing frequency distribution; and replacing the existing frequency distribution with the posterior distribution.

12. A computer program that causes a computer to function as a management device that manages vehicle operation, The computer a storage unit that stores operation records including target costs required to travel a route from a start point position to an end point position of a target section, and distribution records including a pre-existing frequency distribution of the target costs; a control unit that executes the distribution update of the target cost; The distribution update may be A process of generating a prior distribution of the target cost from the operation record; A process of extracting the existing frequency distribution to be applied to the prior distribution from the distribution record; A process of calculating a posterior distribution of the target cost from the prior distribution and the existing frequency distribution; and replacing the existing frequency distribution with the posterior distribution.