WO2021144923A1

WO2021144923A1 - Energy consumption estimation program, energy consumption estimation method, and energy consumption estimation device

Info

Publication number: WO2021144923A1
Application number: PCT/JP2020/001291
Authority: WO
Inventors: 拓工藤; 拓郎池田
Original assignee: 富士通株式会社
Priority date: 2020-01-16
Filing date: 2020-01-16
Publication date: 2021-07-22
Also published as: JPWO2021144923A1; JP7400837B2

Abstract

This invention makes it possible to estimate energy consumption for a route that has not been traveled on by causing a computer to execute processing in which: for each of a plurality of sets of trip data including a travel start point, travel destination, travel distance, and energy consumption, map data expressing roads as link sets is referred to and a link set indicating a travel route for the trip data is estimated; a storage unit storing road feature information is referred to and feature information corresponding to each link included in the estimated link sets is extracted; the feature information extracted for each link and the travel distance and energy consumption included in each of the plurality of sets of trip data are used to learn a regression model having an energy consumption rate that is the amount of energy consumption per unit of distance traveled as the dependent variable and the feature information as an independent variable; and energy consumption for a designated travel route is estimated on the basis of the feature information for each link composing the designated travel route and the regression model.

Description

Energy consumption estimation program, energy consumption estimation method and energy consumption estimation device

The present invention relates to an energy consumption estimation program, an energy consumption estimation method, and an energy consumption estimation device.

Since an electric vehicle has a shorter cruising distance than a gasoline-powered vehicle, a technique for presenting the cruising distance and the like to a user by estimating power consumption is required (for example, Patent Document 1, Patent Document 2, etc.).

Japanese Unexamined Patent Publication No. 2014-182035 Japanese Unexamined Patent Publication No. 2013-50367

However, the conventional power consumption estimation technology is premised on being applied to a route with a travel record, and it is considered that the effectiveness for a route without a travel record is low.

It is considered that such a situation applies not only to electric vehicles but also to mobile objects that use energy other than electricity, such as gasoline.

Therefore, on one side, the purpose is to make it possible to estimate the energy consumption of routes that have not been traveled.

In one plan, for each of a plurality of trip data including the movement start point, the movement end point, the movement distance, and the energy consumption, the movement route related to the trip data is referred to by referring to the map data representing the road by a set of links. The feature information corresponding to each link included in the estimated set of links is extracted by estimating the set of links indicating the above and referring to the storage unit that stores the feature information of the road, and the feature extracted for each link. Based on the information and the travel distance and energy consumption included in the plurality of trip data, a regression model is learned in which the energy consumption rate, which is the energy consumption per unit travel distance, is used as the objective variable and the feature information is used as the explanatory variable. Then, the computer is made to execute a process of estimating the energy consumption of the designated movement path based on the feature information of each link constituting the designated movement path and the regression model.

According to one aspect, it is possible to estimate the energy consumption of a route that has no record of movement.

It is a figure which shows the configuration example of the power consumption estimation system 1 in embodiment of this invention. It is a figure which shows the hardware configuration example of the power consumption estimation apparatus 10 in embodiment of this invention. It is a figure which shows the functional structure example of the power consumption estimation apparatus 10 in embodiment of this invention. It is a figure which shows the structural example of the trip data storage part 121. It is a flowchart for demonstrating an example of the processing procedure which the power consumption estimation apparatus 10 executes at the time of learning the power consumption rate estimation model m1. It is a figure for demonstrating the outline of the movement distance estimation process. It is a figure which shows an example of the specific result of the estimation route. It is a figure for demonstrating the learning process of the 1st configuration example of the power consumption rate estimation model m1. It is a figure which shows the expression example of the estimated path. It is a figure for demonstrating the learning process of the 2nd configuration example of the power consumption rate estimation model m1. It is a flowchart for demonstrating an example of the processing procedure executed by the power consumption estimation apparatus 10 at the time of estimating the power consumption when the electric vehicle 20 moves in a certain movement path. It is a figure for demonstrating the estimation process of the power consumption of the planned route when the power consumption rate estimation model m1 of the 1st configuration example is used. It is a figure for demonstrating the estimation process of the power consumption of a planned route when the power consumption rate estimation model m1 of the 2nd configuration example is used. It is a flowchart for demonstrating an example of the processing procedure at the time of estimating the optimum route from the viewpoint of electricity cost. It is a figure which shows the map data for demonstrating an example of the processing procedure at the time of estimating the optimum route from the viewpoint of electricity cost.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing a configuration example of the power consumption estimation system 1 according to the embodiment of the present invention. In the power consumption estimation system 1 shown in FIG. 1, the power consumption estimation device 10 is via a network including an information processing device (vehicle-mounted device) mounted on each of a plurality of electric vehicles 20 and a mobile communication network or the like. Be connected.

The power consumption estimation device 10 learns a regression model for estimating the power consumption rate based on the actual data for each trip of each electric vehicle 20, and uses the regression model to determine a newly designated movement route. One or more computers that estimate power consumption. The trip is a section from when the power of the electric vehicle 20 is turned on and the movement is started until the power is turned off and the movement is completed. In the case of a gasoline-powered vehicle, the trip is the section from when the ignition is turned on to when it is turned off.

FIG. 2 is a diagram showing a hardware configuration example of the power consumption estimation device 10 according to the embodiment of the present invention. The power consumption estimation device 10 of FIG. 2 has a drive device 100, an auxiliary storage device 102, a memory device 103, a CPU 104, an interface device 105, and the like, which are connected to each other by a bus B, respectively.

The program that realizes the processing by the power consumption estimation device 10 is provided by the recording medium 101. When the recording medium 101 on which the program is recorded is set in the drive device 100, the program is installed in the auxiliary storage device 102 from the recording medium 101 via the drive device 100. However, the program does not necessarily have to be installed from the recording medium 101, and may be downloaded from another computer via the network. The auxiliary storage device 102 stores the installed program and also stores necessary files, data, and the like.

The memory device 103 reads and stores the program from the auxiliary storage device 102 when the program is instructed to start. The CPU 104 executes the function related to the power consumption estimation device 10 according to the program stored in the memory device 103. The interface device 105 is used as an interface for connecting to a network.

An example of the recording medium 101 is a portable recording medium such as a CD-ROM, a DVD disc, or a USB memory. Further, as an example of the auxiliary storage device 102, an HDD (Hard Disk Drive), a flash memory, or the like can be mentioned. Both the recording medium 101 and the auxiliary storage device 102 correspond to computer-readable recording media.

FIG. 3 is a diagram showing a functional configuration example of the power consumption estimation device 10 according to the embodiment of the present invention. In FIG. 3, the power consumption estimation device 10 includes a route estimation unit 11, a variable value extraction unit 12, a model learning unit 13, a power consumption estimation unit 14, and the like. Each of these parts is realized by a process of causing the CPU 104 to execute one or more programs installed in the power consumption estimation device 10. The power consumption estimation device 10 also uses storage units such as a trip data storage unit 121, a map data storage unit 122, and a weather log data storage unit 123. Each of these storage units can be realized by using, for example, an auxiliary storage device 102, a storage device that can be connected to the power consumption estimation device 10 via a network, or the like.

The trip data storage unit 121 stores data including actual information related to the trip (hereinafter referred to as "trip data") for each past trip of each electric vehicle 20.

FIG. 4 is a diagram showing a configuration example of the trip data storage unit 121. In FIG. 4, one line corresponds to one trip data. Each trip data includes values such as vehicle ID, movement start place, movement end place, movement distance, average speed, movement start date and time, movement end date and time, power consumption and acceleration / deceleration frequency.

The vehicle ID is identification information of each electric vehicle 20. The movement start location is the position information of the position where the movement (trip) is started. The movement end location is the position information of the position where the movement (trip) has ended. These location information may be accurate latitude and longitude, and from the viewpoint of personal information protection, they are information indicating the area corresponding to one of the meshes when the map is divided into meshes (for example, 100 m square). You may.

The travel distance is the total travel distance [km] in the trip. The average speed is the average speed [km / h] in the trip. The move start date and time is the start date and time of the move (trip). The move end date and time is the end date and time of the move (trip). The power consumption is the power consumption [kW] in the trip. The acceleration / deceleration frequency is the ratio of the sum of the durations of the states in which the absolute value of the acceleration of ± is equal to or greater than the threshold value to the period from the movement start date / time to the movement end date / time. The acceleration / deceleration frequency is an example of a parameter indicating a tendency (habit) of how the driver drives. Different parameters to the same effect may be used instead of the acceleration / deceleration frequency. Further, among the parameters not shown in FIG. 4, such as the usage rate of air conditioning and the load capacity of luggage, the electric vehicle 20 can measure the power consumption rate (per unit mileage) of the electric vehicle 20. The trip data may include parameters (having a correlation) that affect (power consumption) [Wh / km].

In this way, trip data is data that summarizes one trip, and does not include instantaneous position information, power consumption, etc. in the time series. For example, the trip data may be transmitted from each electric vehicle 20 to the power consumption estimation device 10 at a predetermined timing. Alternatively, the trip data may be acquired from each electric vehicle 20 in the dealer, and the trip data may be transmitted from the terminal in the dealer to the power consumption estimation device 10.

The route estimation unit 11 estimates one movement route for each trip data stored in the trip data storage unit 121. That is, the trip data does not include information indicating the travel route of the trip. Therefore, the route estimation unit 11 trips related to the trip data based on the movement start location, movement end location, and travel distance included in the trip data and the map data stored in the map data storage unit 122. Estimate the movement route in. Map data is data that represents a road by a set of links and nodes, and data that includes information on various facilities (POI (Point Of Interest), etc.). The estimation result of the movement route (hereinafter referred to as "estimated route") by the route estimation unit 11 is also represented by a set of links (a shape in which a plurality of links are connected to one) as in the case of map data.

The variable value extraction unit 12 learns the power consumption rate estimation model m1 from the trip data, the map data storage unit 122, and the weather log data storage unit 123 for each estimation route (that is, for each trip data for which the movement route is estimated). Extract the data. In the present embodiment, the power consumption rate estimation model m1 is a regression model in which the power consumption rate is used as an objective variable and the characteristic information of the link (road), the driving tendency, the weather conditions, and the like are used as explanatory variables. Therefore, the actual values of the objective variable and the explanatory variable are extracted as learning data. The feature information (gradient, etc.) of the link can be acquired from the map data storage unit 122. The meteorological conditions can be acquired from the meteorological log data storage unit 123. That is, the meteorological log data storage unit 123 stores log data of meteorological conditions in each region (hereinafter referred to as "meteorological data").

The model learning unit 13 learns the power consumption rate estimation model m1 based on the actual value group of the explanatory variables and the objective variables of the power consumption rate estimation model m1 extracted by the variable value extraction unit 12.

The power consumption estimation unit 14 uses the learned power consumption rate estimation model m1 to estimate the power consumption when traveling on a designated movement route (for example, a route that the user plans to travel).

Hereinafter, the processing procedure executed by the power consumption estimation device 10 will be described. FIG. 5 is a flowchart for explaining an example of a processing procedure executed by the power consumption estimation device 10 at the time of learning the power consumption rate estimation model m1.

In step S110, the route estimation unit 11 executes a movement route estimation process for each trip data stored in the trip data storage unit 121.

FIG. 6 is a diagram for explaining the outline of the movement distance estimation process. The trip data to be processed in FIG. 6 is referred to as “target trip data”.

As shown in FIG. 6, the route estimation unit 11 first searches for a route by referring to the target trip data and the map data (S111). Specifically, the route estimation unit 11 searches for a route between the movement start location and the movement end location of the target trip data with reference to the map data. Here, multiple routes can be searched. Subsequently, the route estimation unit 11 compares the travel distance estimated for each searched route with the travel distance included in the target trip data, and identifies one estimated route based on the error rate of the travel distance. do.

Regarding the movement distance estimation process, two examples of the first method and the second method will be described more specifically. However, the movement route may be estimated by another method. The step numbers in the following description correspond to the step numbers in FIG.

[First method of moving distance estimation processing]
In step S111, the route estimation unit 11 searches for a predetermined number of routes in ascending order of required time between the movement start point and the movement end point of the target trip data with reference to the map data, and the movement distance of each route. Calculate the estimated value of. The estimated value of the travel distance of the route can be calculated, for example, by obtaining the sum of the distances of the links constituting the route.

In step S112, the route estimation unit 11 refers to the route related to the estimated value having the highest error rate with the travel distance of the target trip data among the estimated values of the travel distance calculated for each searched route with respect to the target trip data. Specify as an estimated route. Here, the error rate is a value obtained by the following equation when the travel distance of the target trip data is D ₁ and the estimated value of the travel distance of the searched route is D _2.
Error rate = | D _2- D ₁ | ÷ D ₁
A high error rate means that the value of the error rate is small. When there are a plurality of estimated values having the highest error rate, one estimated value may be selected by an arbitrary method, and the route related to the estimated value may be specified as the estimated route.

[Second method of moving distance estimation processing]
In step 111, the route estimation unit 11 searches for the shortest route between the movement start location and the movement end location of the target trip data with reference to the map data, and calculates the estimated value of the travel distance of the shortest route. For example, a known method such as Dijkstra's algorithm may be used to search for the shortest path.

In step S112, if the error rate between the estimated value of the travel distance of the shortest route and the travel distance of the target trip data is less than the threshold value, the route estimation unit 11 uses the shortest route searched in step S111 as the estimation route of the target trip data. And. If the error rate is equal to or greater than the threshold value, there is no estimation route for the target trip data. Therefore, in this case, the target trip data is not used in the subsequent processing.

FIG. 7 is a diagram showing an example of a specific result of the estimated route. FIG. 7 shows that the estimated route is specified for each trip data. However, when the second method is adopted, there may be trip data in which the estimated route is not specified. In FIG. 7, for convenience, each estimated route has the same shape, but in reality, each estimated route has a different shape.

Return to Fig. 5. Following step S110, the variable value extraction unit 12 sets the training data of the power consumption rate estimation model m1 (each explanatory variable and the objective variable of the power consumption rate estimation model m1) for each estimation path specified in step S110. The extraction process of the value (actual value)) is executed (S120).

The objective variable of the power consumption rate estimation model m1 is the power consumption rate [Wh / km]. The power consumption rate is obtained by dividing the power consumption of the trip data by the travel distance.

The explanatory variables of the power consumption rate estimation model m1 are a variable that depends on the road (link of map data) (hereinafter referred to as "characteristic variable of the link") and a variable that does not depend on the road (hereinafter referred to as "link-independent"). There are variables. The value of the feature variable of the link can be said to be the feature information (or feature amount) of the link. On the other hand, it can be said that the value of the variable that does not depend on the link is information that does not depend on the characteristics of the road and that affects (correlates with) energy consumption (energy consumption rate). Each is illustrated below.

[Example of link feature variables]
・ Gradient of road corresponding to link ・ Type of road corresponding to link (expressway, national road, etc.)
-The width of the road corresponding to the link-The number of traffic lights around the road corresponding to the link-The number of stores around the road corresponding to the link In this embodiment, the position information of the link is not regarded as a characteristic variable of the link. This is so that the power consumption rate estimation model m1 does not depend on the position information of the link. Since the power consumption rate estimation model m1 does not depend on the position information of the link, the power consumption rate estimation model m1 enables the estimation of the power consumption rate even on a road (link) that has not been moved in the past. Can be done. For example, it is possible to estimate the power consumption rate in consideration of the movement record by further adding the information including the position information of the link to the process shown in the present embodiment.

[Example of variable that does not depend on the link]
-Estimated route travel distance-Acceleration / deceleration frequency-Air-conditioning usage rate-Luggage load-Temperature-Wind direction-Wind speed-Weather Link characteristic variable values are obtained from the map data storage unit 122 based on the link position information. It can be extracted (acquired). That is, the variable value extraction unit 12 acquires the value of the feature variable of the link for each link constituting the estimated route from the map data based on the position information of the link for each estimated route. Regarding the number of traffic lights in the vicinity and the number of stores in the vicinity, for example, the map data may be divided into a grid and the number of traffic lights or stores in the grid including the target link may be counted.

On the other hand, the variable value extraction unit 12 extracts (acquires) the value of the variable that does not depend on the link from the trip data, the meteorological data, or the like for each estimated route. For example, the variable value extraction unit 12 extracts (acquires) the weather conditions such as the movement start date and time of the trip data corresponding to the estimated route and the temperature, wind direction / speed, and weather at the movement start location for each estimated route. Then, the extracted value is given to each link constituting the estimated route. Further, the variable value extraction unit 12 extracts (acquires) the travel distance, acceleration / deceleration frequency, air conditioning usage rate, load capacity of luggage, etc. from the trip data corresponding to the estimated route for each estimated route. A value is given to each link constituting the estimated route. That is, the value of the variable that does not depend on the link is given the same value to each link belonging to the same estimated route. It is also possible that trip data exists for long-distance travel. In this case, it is possible that the weather conditions and the like may change during the movement. Therefore, for the explanatory variables whose values are extracted from the meteorological data, the values may be extracted for each link. In this case, the date and time when the electric vehicle 20 passes through each link in the estimated route is obtained by proportionally dividing the elapsed time from the movement start date and time to the movement end date and time of the trip data corresponding to the estimated route based on the distance of each link. You may be asked. Meteorological conditions and the like based on the meteorological data corresponding to the date and time and the position information of the link obtained in this way may be added to the link.

Following step S120, the model learning unit 13 learns the power consumption rate estimation model m1 using the values of the explanatory variables and the objective variables extracted for each link of each estimation path by the variable value extraction unit 12. Is executed (S130).

In this embodiment, two configuration examples, a first configuration example and a second configuration example, are shown for the power consumption rate estimation model m1. Since the learning process of the power consumption rate estimation model m1 is different for each configuration example, it will be described below for each configuration example.

[Learning process of the first configuration example of the power consumption rate estimation model m1]
FIG. 8 is a diagram for explaining the learning process of the first configuration example of the power consumption rate estimation model m1. In FIG. 8, the power consumption rate estimation model m1 includes a link-independent model m21, a link-dependent model m22, and a composite model m23.

The link-independent model m21 is a model that does not depend on the characteristics of the link (road). That is, the link-independent model m21 is a regression model in which a variable that does not depend on the link but affects the power consumption rate (has a correlation) is used as an explanatory variable, and the power consumption rate is used as an objective variable. The link-dependent model m22 is a model that depends on the characteristics of the link (road). That is, the link-dependent model m22 is a regression model in which the characteristic variable of the link is used as the explanatory variable and the power consumption rate is used as the objective variable. The composite model m23 calculates, for example, the power consumption rate as the final output of the power consumption rate estimation model m1 by giving weights to each of the estimated value of the link-dependent model m22 and the estimated value of the link-independent model m21. It is a model to do. For the optimization of each weight, a known technique such as grid search can be used.

FIG. 8 shows that the value of the link-independent variable extracted for the estimated route corresponding to a certain trip data and assigned to each link of the estimated route is input to the link-independent model m21. There is. Further, it is shown that the value of the feature variable of the link extracted for each link of the estimated route and assigned to each link is input to the link-dependent model m22. Further, it is shown that the outputs (estimated values) from each of the link-independent model m21 and the link-dependent model m22 are input to the composite model m23. Furthermore, it is shown that the power consumption rate estimation model m1 is trained by comparing the output (estimated value) from the composite model m23 with the actual value of the power consumption rate derived from the trip data. There is.

When the power consumption rate estimation model m1 includes three models as shown in FIG. 8, two learning methods, a first learning method and a second learning method, can be given as an example as learning methods for the three models. The first learning method is a method of learning the synthetic model m23 by using the trained link-independent model m21 and the trained link-dependent model m22 after learning the link-independent model m21 and the link-dependent model m22. .. The second learning method is a method of learning three models at the same time.

First, the first learning method will be explained. In the first learning method, the learning of the link-independent model m21 and the link-dependent model m22 can be executed independently (in parallel).

The link-independent model m21 is learned by the model learning unit 13 by executing the following steps (1) and (2).

(1) The objective variable of each estimated route is as follows (that is, the power consumption rate of the entire trip [Wh / km]).

However, _Pi is the power consumption of the link i when the estimated route is expressed as shown in FIG. Further, _Di is the distance of the link i. m is the total number of links included in the estimated route. note that,

Is the power consumption of the trip data (FIG. 4) corresponding to the estimated route.

(2) The following data set is obtained as training data by the value of the variable independent of the link extracted for each estimated route in step S120 and the objective variable in (1). Therefore, the model learning unit 13 learns the link-independent model m21 by using the learning data, for example, by a method of multiple regression or nonlinear multivariate regression.
Estimated path 1 learning data: Objective variable value, explanatory variable 1 value ... Explanatory variable k value Estimated path 2 learning data: objective variable value, explanatory variable 1 value ... Explanatory variable k value・
・
・
Learning data of the estimated path n: the value of the objective variable, the value of the explanatory variable 1 ... the value of the explanatory variable k The explanatory variable k is the kth variable among the variables that do not depend on the link given to the link m. It is a variable.

On the other hand, the link-dependent model m22 is represented by the following function F (for example, a neural network).

However, α _i is the weight of the link i, and each X _i is a characteristic variable of the link given to the link i. j is the total number of feature variables of the link. The weight αi corresponds to the power consumption rate of the link _{_{i (P i ÷ D i)}} .

In this case, learning is performed by the model learning unit 13 executing the following procedure (1).

(1) The model learning unit 13 optimizes the learning parameters (neural network learning parameters) of the function F so that the following equations are satisfied.

The value of the learning parameter of the function F can be optimized, for example, by an error backpropagation method or the like with the following absolute value as a loss.

By learning parameters of the function F is optimized, by entering the D _i and each variable X _i of any link i to a function F, it is possible to return the power consumption rate of the link i. That is, the function F corresponds to the link-dependent model m22. However, the output per estimated path of the link-dependent model m22 shown in FIG. 8 is as follows.

However, m is the total number of links included in the estimated route.

When the learning of the link-independent model m21 and the link-dependent model m22 is completed, the model learning unit 13 learns the synthetic model m23. Specifically, the model learning unit 13 executes the following steps (1) to (3) for each estimation route.

(1) The model learning unit 13 inputs the value of the variable that does not depend on the link into the link-independent model m21, inputs the value of the feature variable of each link of the estimated route to be processed into the link-dependent model m22, and two of them. Get an estimate of the power consumption rate from the model.

(2) The model learning unit 13 inputs two estimated values into the composite model m23, and obtains a value (estimated value of power consumption rate) output from the composite model m23.

(3) The model learning unit 13 determines the parameters of the synthetic model m23 based on the difference between the estimated value output from the synthetic model m23 and the actual value of the power consumption rate based on the trip data corresponding to the estimated path to be processed. (For example, the weight for each of the link-independent model m21 and the link-dependent model m22) is updated.

When the difference converges, the learning process of the synthetic model m23 ends.

Next, the second learning method will be described. In the second learning method, the comparison target with the actual value of the power consumption rate is only the estimated value output from the synthetic model m23. The model learning unit 13 optimizes each model by back-calculating the errors of the link-independent model m21 and the link-dependent model m22 from the error of the estimated value of the synthetic model m23 with respect to the actual value of the power consumption rate.

For example, the composite model m23 is a model that outputs the weighted average of the link-independent model m21 and the link-dependent model m22, and each parameter will be described as follows.
Output of link-independent model m21: a
Output of link-dependent model m22: b
Output of synthetic model m23: c
Weight for link-independent model m21: w1
Weight for link-dependent model m22: w2
In this case, the output c of the synthetic model m23 is as follows.
c = a × w1 + b × w2
Here, assuming that the actual value of the power consumption rate is x, the error Ec of the composite model m23 is as follows.
Ec = cx
In the second learning method, it is considered that the error Ec is the sum of the output error Ea of the link-independent model m21 and the error Eb of the link-dependent model m22 in the composite model m23.

Therefore, the model learning unit 13 executes the following calculation.
Ea = w1 / (w1 + w2) * Ec
Eb = w2 / (w1 + w2) * Ec
Then, the error in the link-independent model m21 and the error in the model link-dependent model m22 can be calculated from the error and the weight in the composite model m23.

The model learning unit 13 repeatedly optimizes the link-independent model m21 and the link-dependent model m22 while updating the weights w1 and w2, and finds the combination of w1 and w2 having the smallest Ec, whereby the link-independent model m21 , The link-dependent model m22 and the synthetic model m23 can be optimized (learned).

[Learning process of the second configuration example of the power consumption rate estimation model m1]
FIG. 10 is a diagram for explaining the learning process of the second configuration example of the power consumption rate estimation model m1. In FIG. 10, the power consumption rate estimation model m1 is one model that is not divided into a plurality of models as shown in FIG. The power consumption rate estimation model m1 in FIG. 10 is a regression model (for example, a neural network, a decision tree, a multiple regression model, etc.) in which a link-independent variable and a link feature variable are used as explanatory variables and the power consumption rate is used as an objective variable. Is.

In this case, the model learning unit 13 uses the power consumption rate estimation model as the value of the variable that does not depend on the link extracted for the estimated route and the value of the characteristic variable of the link extracted for the estimated route for each estimation route. Enter in m1. The model learning unit 13 compares the estimated value of the power consumption rate output from the power consumption rate estimation model m1 with the actual value of the power consumption rate that can be obtained from the trip model corresponding to the estimated route, and obtains the comparison result. Based on this, the learning parameters of the power consumption rate estimation model m1 are updated. The model learning unit 13 updates the learning parameters until the comparison result converges. As a result, when the value of the variable that does not depend on the link of the arbitrary route and the value of the feature variable of each link constituting the route are input, the power consumption rate estimation model m1 that outputs the estimated value of the power consumption rate is learned. Will be done.

Next, an example of the processing procedure executed using the learned power consumption rate estimation model m1 will be described. FIG. 11 is a flowchart for explaining an example of a processing procedure executed by the power consumption estimation device 10 when estimating the power consumption when the electric vehicle 20 moves along a certain movement route.

In step S210, the power consumption estimation unit 14 receives input of each position information (for example, latitude and longitude) of the departure place (planned movement start place) and the destination (planned movement end place). The starting point and the destination may be specified by, for example, a user of a user terminal connected to the power consumption estimation device 10 via a network.

Subsequently, the power consumption estimation unit 14 determines the movement route from the departure point to the destination (hereinafter, referred to as "planned route") (S220). For example, the power consumption estimation unit 14 searches for a plurality of routes from the departure point to the destination with reference to the map data storage unit 122, and the route selected by the user from the plurality of routes is determined. It may be determined as a planned route. Alternatively, the shortest route from the starting point to the destination may be determined as the planned route. The planned route is represented by a set of links and nodes.

Although the example in which the power consumption estimation unit 14 searches for the planned route is shown here, the movement route searched by another computer may be input to the power consumption estimation unit 14 as the planned route.

Subsequently, the power consumption estimation unit 14 acquires the value of the explanatory variable of the power consumption rate estimation model m1 with respect to the planned route (S230). Specifically, the power consumption estimation unit 14 acquires the value of the feature variable of the link for each link included in the planned route. The power consumption estimation unit 14 also acquires the value of a variable that does not depend on the link for the planned route. The current date and time may be applied to variables such as meteorological data for which the date and time need to be specified for the acquisition of values. Alternatively, the date and time when the planned route is scheduled to be moved may be input to the power consumption estimation unit 14 together with the planned route. Basically, the method of acquiring the value of each explanatory variable may be performed in the same manner as the extraction of each variable by the variable value extraction unit 12. However, when the scheduled movement date and time is in the future, the weather forecast data corresponding to the future date and time may be used as the weather data.

Subsequently, the power consumption estimation unit 14 calculates the estimated value of the power consumption [kW] of the planned route by inputting the acquired explanatory variable value into the learned power consumption rate estimation model m1 (S240). ..

Regarding step S240, two examples will be described, one is the case where the power consumption rate estimation model m1 is the first configuration example (FIG. 8) and the other is the case where the power consumption rate estimation model m1 is the second configuration example (FIG. 10).

FIG. 12 is a diagram for explaining the power consumption estimation process of the planned route when the power consumption rate estimation model m1 of the first configuration example is used.

When the power consumption estimation unit 14 inputs the value of the link-independent variable into the trained link-independent model m21, the link-independent model m21 outputs an estimated value of one power consumption rate for the planned route ( S241a).

Further, when the power consumption estimation unit 14 inputs the value of the feature variable (feature amount of the link) acquired for each link of the planned route into the trained link-dependent model m22, the link-dependent model m22 becomes , The estimated value of the power consumption rate for the link is output for each link of the planned route (S242a). In FIG. 12, the link-dependent model m22 is shown for each link, but this does not indicate that the link-independent model m21 exists for each link, and the link-independent model m21 for each link does not exist. Indicates that the calculation is performed by. This point is the same for the synthetic model m23.

Following steps S241a and S242 a, the power consumption estimation unit 14 outputs the estimated value output from the link-independent model m21 for the planned route and the link-dependent model m22 for the link for each link of the planned route. When the calculated estimated value is input to the trained synthetic model m23, the synthetic model m23 is obtained by synthesizing two estimated values for each link of the planned route, and is an estimated value of the power consumption rate for the link. Is output (S243a). The same value is applied to all links in the output from the link-independent model m21.

Subsequently, the power consumption estimation unit 14 calculates the product of the estimated value output from the synthetic model m23 for the link and the distance of the link for each link of the planned route, so that the power consumption [kW] of the link is calculated. ] Is calculated (S244a). The distance of each link can be obtained from the map data.

Subsequently, the power consumption estimation unit 14 calculates the estimated value of the power consumption of the planned route by calculating the sum of the estimated values of the power consumption calculated for each link (S245a).

On the other hand, FIG. 13 is a diagram for explaining the power consumption estimation process of the planned route when the power consumption rate estimation model m1 of the second configuration example is used.

When the power consumption estimation unit 14 inputs the value of the variable that does not depend on the link and the value of the characteristic variable of the link into the learned power consumption rate estimation model m1 for each link of the planned route, the power consumption rate estimation model m1 outputs an estimated value of the power consumption rate for each link (S241b).

Subsequently, the power consumption estimation unit 14 calculates the product of the estimated value output from the power consumption rate estimation model m1 for the link and the distance of the link for each link of the planned route, thereby consuming the link. Calculate the estimated value of power [kW] (S242b)
Subsequently, the power consumption estimation unit 14 calculates the estimated value of the power consumption of the planned route by calculating the sum of the estimated values of the power consumption calculated for each link (S243b).

Next, the processing procedure for estimating the optimum route from the viewpoint of the electric cost of the electric vehicle 20 will be described using the learned power consumption rate estimation model m1. FIG. 14 is a flowchart for explaining an example of a processing procedure when estimating the optimum route from the viewpoint of electricity cost. Further, FIG. 15 is a diagram showing map data for explaining an example of a processing procedure when estimating the optimum route from the viewpoint of electricity cost.

In step S310, the power consumption estimation unit 14 acquires all the map data within the range designated as the search range of the movement route from the map data storage unit 122. As a result, information as shown in (1) of FIG. 15 is acquired. The search range may be determined to cover the starting point and the destination at a minimum, or may be, for example, the entire area of Japan if computational resources permit.

Subsequently, the power consumption estimation unit 14 acquires the value of the explanatory variable of the power consumption rate estimation model m1 (S320). The method of acquiring the value may be the same as in step S230 of FIG. However, in step S320, the values of the explanatory variables are acquired for all the links in the search range. When the value of the variable that does not depend on the link such as meteorological data is unknown, the value of the variable that does not depend on the link may be calculated based on the past actual value such as the average temperature in the area of the search range. ..

Subsequently, the power consumption estimation unit 14 calculates the power consumption rate for each link by inputting the value of the explanatory variable to the learned power consumption rate estimation model m1 for each link (S330). ..

Subsequently, the power consumption estimation unit 14 calculates the power consumption for each link by multiplying the power consumption rate calculated in step S330 by the distance of the link (S340). As a result, information as shown in (2) of FIG. 15 can be obtained. That is, in (2) of FIG. 15, the thickness of the line indicating each link indicates the magnitude of power consumption.

Subsequently, the power consumption estimation unit 14 estimates the optimum route by solving the route optimization problem with the power consumption calculated for each link as the cost (weight) of each link (S350). In FIG. 15 (3), an example of the optimum route is shown by a broken line. As a method for solving the route optimization problem, a known algorithm such as Dijkstra's algorithm or A * method may be used.

Although the example in which the power consumption rate estimation model m1 is used has been described above, only the link-dependent model m22 in the first configuration example (FIG. 8) is used instead of the entire power consumption rate estimation model m1. May be good. In this case, in step S320, the value of the feature variable of the link may be acquired. In step S330, the power consumption rate of each link may be calculated by inputting the value of the feature variable of the link into the link-dependent model m22.

Further, steps S310 to S340 may be executed asynchronously (for example, in advance) with step S350, or may be executed synchronously (in real time) with step S350.

However, when the weight (power consumption) of each link is calculated in advance (that is, when steps S310 to S340 are executed in advance), when the weight is calculated, it is at the date and time when the optimum route is moved. It may not be possible to obtain the values of variables that do not depend on road links such as weather conditions. Therefore, if the sum of the weights of each link included in the optimum route is calculated as the power consumption of the optimum route, the estimated value of the power consumption may deteriorate.

Therefore, when it is necessary to calculate the power consumption of the optimum route, the power consumption estimation unit 14 calculates the estimated value of the power consumption of the optimum route by executing steps S230 and S240 of FIG. 11 for the optimum route. You may. By doing so, it can be expected that the accuracy of the estimated value will be improved.

As described above, according to the present embodiment, the regression model is learned based on the characteristic information of the link included in the movement route of the past trip (movement) and the power consumption rate of the trip. At this time, the position information of the link is not included in the feature amount. Therefore, using the regression model, it is possible to estimate the power consumption even if the route has no past travel record.

This makes it possible, for example, to provide information indicating whether or not the electric vehicle 20 can reach the destination with the current remaining battery level, and to present an appropriate route to a user who wants to save power consumption.

Further, according to the present embodiment, a variable that does not depend on the link (feature amount that does not depend on the link) is also used as an explanatory variable of the regression model. As a result, in the estimation of power consumption, it is possible to consider the situation (weather conditions, etc.) that does not depend on the link, and it is expected that the estimation accuracy will be improved.

Further, according to the present embodiment, for example, it does not require various time series data acquired from various sensors or the like in real time (for example, periodically) within one trip, and returns based on the trip data. You can train the model. This eliminates the need to acquire data from the electric vehicle 20 in real time, and can reduce the cost of constructing a network for acquiring data.

In the present embodiment, an example of estimating the energy (electric power) consumed by the electric vehicle 20 has been described, but a moving body (for example, a gasoline vehicle, a diesel vehicle, etc.) that moves by consuming other energy has been described. The present embodiment may be applied. In this case, the power consumption rate may be replaced by the energy consumption rate corresponding to each moving body (energy consumption per unit travel distance (for example, consumption of gasoline)), and the power consumption may be replaced by the energy consumption.

In the present embodiment, the power consumption estimation device 10 is an example of the energy consumption estimation device. The route estimation unit 11 is an example of the first estimation unit. The variable value extraction unit 12 is an example of the extraction unit. The model learning unit 13 is an example of the learning unit. The power consumption estimation unit 14 is an example of the second estimation unit.

Although the embodiments of the present invention have been described in detail above, the present invention is not limited to such specific embodiments, and various modifications are made within the scope of the gist of the present invention described in the claims.・ Can be changed.

1 Power consumption estimation system 10 Power consumption estimation device 11 Route estimation unit 12 Variable value extraction unit 13 Model learning unit 14 Power consumption estimation unit 20 Electric vehicle 100 Drive device 101 Recording medium 102 Auxiliary storage device 103 Memory device 104 CPU
105 Interface device 121 Trip data storage unit 122 Map data storage unit 123 Meteorological log data storage unit B Bus m1 Power consumption rate estimation model m21 Link-independent model m22 Link-dependent model m23 Composite model

Claims

For each of a plurality of trip data including a movement start point, a movement end point, a movement distance, and energy consumption, a set of links indicating a movement route related to the trip data with reference to map data representing a road by a set of links. Estimate and
By referring to the storage unit that stores the characteristic information of the road, the characteristic information corresponding to each link included in the estimated set of the links is extracted.
Based on the feature information extracted for each of the links and the travel distance and energy consumption included in the plurality of trip data, the energy consumption rate, which is the energy consumption per unit travel distance, is set as the objective variable, and the feature information is described. Learn the regression model as a variable and
The energy consumption of the specified travel path is estimated based on the feature information of each link constituting the designated travel path and the regression model.
An energy consumption estimation program characterized by having a computer perform processing.
For each of the plurality of trip data, a computer is made to execute a process of extracting information that does not depend on the characteristics of the road and that affects energy consumption.
The learning process learns the regression model based on the information and
The estimation process estimates energy consumption based on the information.
The energy consumption estimation program according to claim 1.
The energy consumption is estimated using the regression model for each link included in the search range of the movement route between the specified starting point and the specified destination.
Using the energy consumption estimated for each link included in the search range as the weight of each link, the optimum route between the starting point and the destination is estimated.
The process of estimating the energy consumption of the movement path estimates the energy consumption of the optimum path.
The energy consumption estimation program according to claim 1 or 2, wherein the processing is executed by a computer.
For each of a plurality of trip data including a movement start point, a movement end point, a movement distance, and energy consumption, a set of links indicating a movement route related to the trip data with reference to map data representing a road by a set of links. Estimate and
By referring to the storage unit that stores the characteristic information of the road, the characteristic information corresponding to each link included in the estimated set of the links is extracted.
Based on the feature information extracted for each of the links and the travel distance and energy consumption included in the plurality of trip data, the energy consumption rate, which is the energy consumption per unit travel distance, is set as the objective variable, and the feature information is described. Learn the regression model as a variable and
The energy consumption of the specified travel path is estimated based on the feature information of each link constituting the designated travel path and the regression model.
An energy consumption estimation method characterized by a computer performing processing.
For each of the plurality of trip data, the computer executes a process of extracting information that does not depend on the characteristics of the road and that affects energy consumption.
The learning process learns the regression model based on the information and
The estimation process estimates energy consumption based on the information.
The energy consumption estimation method according to claim 4, wherein the energy consumption is estimated.
The energy consumption is estimated using the regression model for each link included in the search range of the movement route between the specified starting point and the specified destination.
Using the energy consumption estimated for each link included in the search range as the weight of each link, the optimum route between the starting point and the destination is estimated.
The process of estimating the energy consumption of the movement path estimates the energy consumption of the optimum path.
The energy consumption estimation method according to claim 4 or 5, wherein the processing is executed by a computer.
For each of a plurality of trip data including a movement start point, a movement end point, a movement distance, and energy consumption, a set of links indicating a movement route related to the trip data with reference to map data representing a road by a set of links. The first estimation part that estimates
An extraction unit that extracts feature information corresponding to each link included in the estimated set of links by referring to a storage unit that stores road feature information, and an extraction unit.
Based on the feature information extracted for each of the links and the travel distance and energy consumption included in the plurality of trip data, the energy consumption rate, which is the energy consumption per unit travel distance, is set as the objective variable, and the feature information is described. A learning unit that learns a regression model as a variable,
A second estimation unit that estimates the energy consumption of the designated movement path based on the feature information of each link constituting the designated movement path and the regression model, and
An energy consumption estimation device characterized by having.
The extraction unit extracts information that does not depend on the characteristics of the road and affects energy consumption for each of the plurality of trip data.
The learning unit learns the regression model based on the information and
The second estimation unit estimates energy consumption based on the information.
The energy consumption estimation device according to claim 7.
A third estimation unit that estimates energy consumption using the regression model for each link included in the search range of the movement route between the specified starting point and the specified destination.
It has a fourth estimation unit that estimates the optimum route between the starting point and the destination, using the energy consumption estimated for each link included in the search range as the weight of each link.
The second estimation unit estimates the energy consumption of the optimum path.
The energy consumption estimation device according to claim 7 or 8.