Control device, learning device, control system, device control method, and program

Abstract

To provide a control device, a learning device, a control system, an apparatus control method, and a program capable of appropriately controlling the apparatus on the basis of a position and an action pattern of a user.SOLUTION: A cloud server 1 includes: a moving state determination unit 112 which determines a moving route of a user and a stay period of a user in an at least one position being previously set and contained in the moving route; an estimation unit 114 which estimates operation contents of respective apparatuses 9 by using a learned model which estimates the operation contents of respective apparatuses 9 from the moving route and the stay period in at least one position determined by the moving state determination unit 112; and an apparatus control unit 115 which controls respective apparatuses 9 on the basis of the operation contents of respective apparatuses 9 estimated by the estimation unit 114.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present disclosure relates to a control device, a learning device, a control system, a device control method, and a program.

Background Art

[0002] There has been proposed a cooperation system including a cooperation server for cooperating devices, and an adapter for connecting the cooperation server and the devices, which controls the devices to provide a service according to a user in accordance with a scenario incorporated in advance on a scenario platform triggered by an unconscious action, an environmental change, or a physical change of the user (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, in the cooperation system described in Patent Document 1, based on the user's route search log and the history of position information, a pattern related to the user's behavior is specified, and a scenario corresponding to the specified behavior pattern is incorporated into the scenario platform. For example, when the user has a behavior pattern of returning home, devices such as an air conditioner and lighting fixtures installed in the user's home are controlled to operate according to the user's return time.

[0005] The present disclosure has been made in view of the above circumstances, and an object thereof is to provide a control device, a learning device, a control system, a device control method, and a program capable of appropriately controlling a device based on the position and behavior pattern of a device user.

Means for Solving the Problems

[0006] To achieve the above object, the control device according to the present disclosure includes: a movement situation determination unit that determines a user's movement route and the user's stay time at at least one preset location included in the movement route; From the movement route determined by the movement situation determination unit and the stay time at the at least one location, the operation content of each of at least one device installed at a location different from the at least one location and operation time Using a learned model that estimates the operation content of each of the at least one device installed at a location different from the at least one location and operation time an estimation unit that estimates the operation content of each of the at least one device installed at a location different from the at least one location; Based on the operation content of each of the at least one device estimated by the estimation unit and operation time a device control unit that controls each of the at least one device.

Advantages of the Invention

[0007] According to the present disclosure, the estimation unit uses the above-mentioned learned model to determine, from the user's movement route and the user's stay time at at least one location included in the user's movement route, the operation content of each of at least one device installed at a location different from the at least one location and operation time to estimate. Then, the device control unit controls each of the at least one device based on the operation content of each of the at least one device estimated by the estimation unit and operation time Thereby, each of the at least one device can be appropriately controlled based on the position and action pattern of the user of the device.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Embodiments for Carrying Out the Invention

[0009] Hereinafter, a control system according to each embodiment of the present disclosure will be described with reference to the drawings.

[0010] (Embodiment 1) The control device according to the present embodiment includes a movement situation determination unit that determines the movement route of the user and the stay time of the user at at least one preset location included in the movement route, an estimation unit that estimates the operation content of each of at least one device, and a device control unit. The estimation unit estimates the operation content of each of at least one device using a learned model that estimates the operation content of each of at least one device from the movement route determined by the movement situation determination unit and the stay time at at least one location. The device control unit controls each of at least one device based on the operation content of each of at least one device estimated by the estimation unit.

[0011] As shown in FIG. 1, the control system according to the present embodiment includes a plurality (two in FIG. 1) of devices 9 installed in a building H, a cloud server 1, a location management server 2, a relay server 3, and a terminal device 5. Here, the terminal device 5 has a function of receiving GNSS (Global Navigation Satellite System) radiated from an artificial satellite SAT. In the building H, a router 82 connected to a local network NW2 constructed in the building H and a data line termination device 81 connected to the router 82 and a wide area network NW1 are installed. The wide area network NW1 is, for example, the Internet. The local network NW2 is, for example, a wired LAN (Local Area Network) or a wireless LAN. The data line termination device 81 is an ONU (Optical Network Unit), a modem, a gateway, or the like.

[0012] The device 9 is an air conditioner, a refrigerator, a rice cooker, etc., and can communicate with the cloud server 1 via the local network NW2 and the wide area network NW1. Every time the device 9 is manually operated by the user or every time control information is acquired from the cloud server 1, the device 9 generates operation information indicating the operation content and transmits it to the cloud server 1. Here, the operation information includes operation device information indicating a number for identifying the device 9 operated by the user, operation time information indicating the time when the user operated the device 9, operation information indicating a number for identifying the content of the operation performed by the user on the device 9, and season information indicating the month when the user operated the device 9.

[0013] The location management server 2 manages the location information indicating the location of the terminal device 5 transmitted from the terminal device 5 in association with the terminal device identification information for identifying the terminal device 5 that is the transmission source of the location information. Then, when the preset location information acquisition timing arrives, the location management server 2 transmits location information request information for requesting the transmission of the location information of the terminal device 5 to the terminal device 5, and thereby acquires the location information transmitted from the terminal device 5. Further, when the location of the terminal device 5 transitions from a state where it exists inside the area including the user's home preset to a state where it exists outside, the location management server 2 determines that the user has gone out while carrying the terminal device 5, and starts the periodic transmission of the location information of the terminal device 5 to the cloud server 1. On the other hand, when the location management server 2 acquires return home notification information notifying that the user has returned home from the cloud server 1, it stops the transmission of the location information of the terminal device 5 to the cloud server 1.

[0014] When the relay server 3 acquires various information transmitted from the terminal device 5 to the cloud server 1, it transmits the acquired various information to the cloud server 1. Further, when the relay server 3 acquires various information transmitted from the cloud server 1 to the terminal device 5, it transmits the acquired various information to the terminal device 5.

[0015] The terminal device 5 is, for example, a smartphone, and as shown in FIG. 2, includes a CPU (Central Processing Unit) 501, a main memory unit 502, an auxiliary storage unit 503, a display unit 504, an input unit 505, a wide-area communication unit 506, a GNSS reception unit 508, and a bus 509 that connects these to each other. The main memory unit 502 has a volatile memory such as a RAM (Random Access Memory) and is used as a working area for the CPU 501. The auxiliary storage unit 503 is a non-volatile memory such as a semiconductor flash memory and stores programs for the CPU 501 to execute various processes. The display unit 504 is a display device such as a liquid crystal display or an organic EL (Electro-Luminescence) display. The input unit 505 is, for example, a transparent touch pad arranged over the display unit 504. The wide-area communication unit 506 is connected to the wide-area network NW1, transmits information transferred from the CPU 501 to the relay server 3 via the wide-area network NW1, and transfers information received from the relay server 3 via the wide-area network NW1 to the CPU 501. The GNSS reception unit 508 receives GNSS signals radiated from the artificial satellite SAT, generates reference position information indicating the position of the artificial satellite SAT, time information, etc. based on the GNSS signals, and transfers them to the CPU 501.

[0016] By reading out and executing the program stored in the auxiliary storage unit 503 in the main memory unit 502, the CPU 501 functions as a reception unit 511, a display control unit 512, a positioning unit 513, a position information notification unit 514, an operation content acquisition unit 515, a change request generation unit 516, and a change request transmission unit 517, as shown in FIG. 3. Further, as shown in FIG. 3, the auxiliary storage unit 503 shown in FIG. 2 has an image storage unit 531 that stores information on images constituting an operation screen image to be displayed on the display unit 504. Also, as shown in FIG. 3, the main memory unit 502 shown in FIG. 2 has a position storage unit 521 that temporarily stores position information indicating the position of the terminal device 5.

[0017] When the reception unit 511 receives an operation performed by the user on the input unit 505, it notifies the display control unit 512 of information indicating the received operation content. Also, when the reception unit 511 receives an operation for starting the positioning of the terminal device 5 by the user on the input unit 505, it notifies the positioning unit 513 of the positioning start command information. The display control unit 512 generates an operation screen image using the information stored in the image storage unit 531 based on the information notified from the reception unit 511 and causes the display unit 504 to display it.

[0018] The positioning unit 513 estimates the position of the terminal device 5 based on the reference position information, time information, etc. input from the GNSS reception unit 508. Then, the positioning unit 513 updates the position information stored in the position storage unit 521 with the position information indicating the estimated position of the terminal device 5. When the position information notification unit 514 acquires position information request information from the position management server 2, it transmits the position information stored in the position storage unit 521 to the position management server 2. The operation content acquisition unit 515 acquires operation information indicating the content of the operation performed by the cloud server 1 on the device 9 transmitted from the cloud server 1 and notifies the display control unit 512 of the acquired operation information. Here, when the display control unit 512 is notified of the operation information from the operation content acquisition unit 515, it generates an operation screen image including the notified operation information and causes the display unit 504 to display it.

[0019] The change request generation unit 516 generates change request information for requesting a change in the operation content of the device 9 from the cloud server 1. Here, assume that the user performs an operation for changing the operation content via the input unit 505 when the operation content of the device 9 displayed on the display unit 504 is different from the intended operation content. In this case, the reception unit 511 notifies the change request generation unit 516 of the operation information indicating the operation content of the device 9 intended by the user. Then, the change request generation unit 516 generates change request information including the operation information notified from the reception unit 511 and notifies the generated change request information to the change request transmission unit 517. The change request transmission unit 517 transmits the change request information notified from the change request generation unit 516 to the cloud server 1 via the relay server 3.

[0020] Returning to FIG. 2, the cloud server 1 includes a CPU 101, a main memory unit 102, an auxiliary storage unit 103, a wide area communication unit 106, a timing unit 107, and a bus 109 that interconnects them. The CPU 101 is, for example, a multi-core processor. The main memory unit 102 is composed of a volatile memory and is used as a working area for the CPU 101. The auxiliary storage unit 103 is composed of a large-capacity non-volatile memory and stores programs for realizing various functions of the cloud server 1. The wide area communication unit 106 is connected to the wide area network NW1. The timing unit 107 is, for example, a real-time clock.

[0021] The CPU 101 functions as a position information acquisition unit 111, a movement status determination unit 112, a homecoming notification unit 113, an estimation unit 114, a device control unit 115, an operation information acquisition unit 116, a model generation unit 117, an operation content notification unit 118, a change request acquisition unit 119, and a function update unit 120, as shown in FIG. 4, by reading out and executing the programs stored in the auxiliary storage unit 103 into the main memory unit 102. Further, the auxiliary storage unit 103 has a movement status storage unit 131, a model storage unit 132, an operation history storage unit 133, a location storage unit 134, and a function storage unit 135.

[0022] As shown in, for example, FIG. 5, the movement status memory unit 131 stores movement status information including movement route information indicating the user's movement route and stay time information indicating the user's stay time at at least one preset location included in the movement route, in association with movement status identification information for identifying the movement status. The movement route information is information in which the position coordinates of at least one location included in the user's movement route are arranged in time series based on the time when the user stopped by. The stay time information includes arrival time information indicating the arrival time of the user at each of at least one location included in the movement route and departure time information indicating the departure time of the user from that location. For example, the movement status information identified by the movement status identification information "a2" indicates that the user left home at 8:00, arrived at the workplace at 9:00, stayed there until 17:00, and then stayed at a nearby supermarket from 17:40 to 18:10, and then returned home at 18:30.

[0023] Returning to FIG. 4, the model memory unit 132 stores a learned model for estimating the operation content of each device 9 from the user's movement route and the user's stay time at at least one location. The learned model is, for example, a feedforward neural network. In this case, the model memory unit 132 stores information indicating the structure of the neural network and information indicating the weight coefficients in the neural network. Here, the information indicating the structure of the neural network includes information indicating the number of nodes, the number of layers, the weight coefficients corresponding to each node, and the activation function of the neural network. This neural network has an input layer L10, a hidden layer L20, and an output layer L30 as shown in FIG. 6. The position coordinates of at least one location included in the user's movement route and the stay time at each of at least one location are input to the input layer L10.

[0024] The hidden layer L20 is composed of N (N is a positive integer) layers including a preset number M[j] of nodes x[j,i] (1 ≤ i ≤ M[j], and M[j] is a positive integer). That is, the hidden layer L20 has a structure in which each node sequence is connected. Here, the output y[j,i] of each node x[j,i] is represented by the relational expression of the following formula (1).

Number

[0025] Here, W[j,i,k] represents a weight coefficient, and f(*) represents an activation function. As the activation function, a non-linear function such as a sigmoid function, a ramp function, or a step function is used. In the hidden layer L20, the information input to the node is the sum of the outputs of each node in the previous layer multiplied by their respective weight coefficients. Then, the output of the activation function with the sum as an argument is transmitted to the next layer. The output layer L30 outputs a Q-value vector (hereinafter referred to as "Q-value") having as elements the expected values of the action values for each combination of a preset plurality of types of devices 9, the operation content for the device 9, and the time when the operation content is executed (hereinafter, "Q(s0,a0)Q(s0,a1)···Q(s t ,a t )Q(s t ,a t+1 )···). Here, the output layer L30 executes a process using, for example, a softmax function. In this case, the model storage unit 132 stores information indicating the structure of the hidden layer L20 of the neural network represented by formula (1) and information indicating the weight coefficient W[j,i,k], and information indicating the function used in the output layer L30.

[0026] Returning to FIG. 4, the operation history storage unit 133 stores operation information indicating a combination of the type of the device 9 to be operated, the operation content, and the operation time within a preset period after the user once left home and then arrived home in the past. As shown in FIG. 7, for example, the operation history storage unit 133 stores the operation information in association with operation information identification information for identifying the operation information, and movement situation identification information for identifying the movement situation of the user before arriving home immediately before the operation related to the combination of the type of the device 9 to be operated, the operation content, and the operation time indicated by the operation information. Similar to the operation information transmitted from the device 9 described above, the operation information includes operation device information indicating a number for identifying the device 9 to be operated, operation time information indicating the time when the device 9 to be operated is operated, and operation information indicating a number for identifying the operation content for the device 9. Further, the operation history storage unit 133 stores each operation information in association with season information indicating the month in which the corresponding operation was performed. For example, movement situation identification information "a2" is assigned to operation information having the content that the user performed an "opening / closing operation of the door" on the "refrigerator" at "18:40" in "August". As shown by the movement situation information corresponding to the movement situation identification information "a2" in FIG. 5, this indicates that the user left home at 8:00, arrived at the workplace at 9:00 and stayed there until 17:00, then stayed at a nearby supermarket from 17:40 to 18:10, and then returned home at 18:30, and the user performed an opening / closing operation on the refrigerator at 18:40.

[0027] Returning to FIG. 4, the location storage unit 134 stores information indicating the range of position coordinates at each of a plurality of preset locations. Here, the information indicating the range of position coordinates at each of the plurality of locations indicates the range of position coordinates inside an area of a preset size for each location. When the position information acquisition unit 111 acquires the position information of the terminal device 5 transmitted from the position management server 2, the acquired position information is notified to the movement situation determination unit 112.

[0028] The function memory unit 135 stores function information indicating an update function for updating Q values for each combination of a preset plurality of types of devices 9 calculated using a learned model, the operation content for the device 9, and the time at which the operation content is executed. Here, the update function is represented by, for example, the relational expression shown in the following formula (2).

[0029]

Number

[0030] Here, Q(s t ,a t ) represents the Q value before update by the function update unit 120, and Q(s t ,a t ) new represents the Q value after update by the function update unit 120. Also, s t is operation information identification information indicating a combination of the type of device 9, the operation content, and the operation time determined based on the learned model before update, and s t+1 is operation information identification information for identifying the operation information included in the aforementioned change request information. Further, at is movement situation identification information representing the movement situation of the user. R t+1 (s t ,a t ,s t+1 ) is a value indicating the reward when changing the combination represented by st to the combination represented by st+1 for the combination of the type of device 9, the operation content, and the operation time. γ represents a discount rate, and α represents a learning coefficient. γ is set in the range of 0 < γ ≤ 1, and α is set in the range of 0 < α ≤ 1. Also, maxQ(s t+1 ,a) represents the Q value when the movement situation of the user becomes a movement situation that maximizes the Q value of the combination of the type of device 9, the operation content, and the operation time indicated by the operation information included in the change request information.

[0031] The movement status determination unit 112 determines the movement route of the user and the stay time period of each location included in the movement route based on the location information of the terminal device 5 notified from the location information acquisition unit 111 and the time information notified from the time measurement unit 107. Here, the movement status determination unit 112 refers to information indicating the range of the position coordinates of each of the plurality of locations stored in the location storage unit 134, and when the position indicated by the location information of the terminal device 5 transitions from a state where it exists outside the range of the position coordinates of any one of the plurality of locations to a state where it exists within the range of the position coordinates, it determines that the user has arrived at that location. Further, the movement status determination unit 112 determines the time of arrival at that location based on the time information notified from the time measurement unit 107. On the other hand, the movement status determination unit 112 determines that the user has departed from a location when the position indicated by the location information of the terminal device 5 transitions from a state where it exists within the range of the position coordinates of any one of the plurality of locations to a state where it exists outside the range of the position coordinates. Further, the movement status determination unit 112 determines the time of departure from that location based on the time information notified from the time measurement unit 107. Then, every time the user departs from or arrives at any one of the plurality of locations where the information indicating the range of the position coordinates is stored in the location storage unit 134, the movement status determination unit 112 associates the information indicating the position coordinates of the center of the range of the position coordinates of that location and the time information indicating the time when the user departed or arrived with the movement route identification information and stores them in the movement status storage unit 131 in chronological order. Further, when the movement status determination unit 112 determines that the user has arrived home, it notifies the home arrival notification information to the home arrival notification unit 113. Then, when the home arrival notification unit 113 is notified of the home arrival notification information from the movement status determination unit 112, it notifies the home arrival notification information to the estimation unit 114 and transmits it to the location information acquisition unit 111.

[0032] The estimation unit 114 estimates the operation content of each device 9 by using a learned model that estimates the operation content of each device 9 from the moving route of the user determined by the movement situation determination unit 112 and the staying time of the user at the locations included in the moving route. Specifically, the estimation unit 114 uses the learned model stored in the model storage unit 132, and from the position coordinates of each location included in the moving route of the user stored in the movement situation storage unit 131, the departure time and arrival time of the user at each location, and the month to which the estimation time belongs, calculates the Q value for each combination of the operation content and operation time for each device 9 that can be an operation target. Then, the estimation unit 114 estimates the combination of the device 9 with the highest calculated Q value, the operation content for the device 9, and the operation time. The estimation unit 114 notifies the device control unit 115 and the operation content notification unit 118 of the operation information indicating the combination of the estimated operation target device 9, the operation content for the device 9, and the operation time. Here, when the device 9 is, for example, an air conditioner, the operation content includes the set temperature of the space where the air conditioner is installed.

[0033] The device control unit 115 controls each device 9 based on the operation content of each device 9 estimated by the estimation unit 114. Specifically, the device control unit 115 generates control information for controlling the operation of the device 9 based on the operation information notified from the estimation unit 114, and controls the operation of the device 9 by transmitting the generated control information to the device 9 via the wide area network NW1 and the local network NW2. Also, after transmitting the control information to the device 9, the device control unit 115 generates operation information based on the operation information notified from the estimation unit 114 and stores it in the operation history storage unit 133. Furthermore, when estimating the operation content by the estimation unit 114, if there is at least one operating device 9, the device control unit 115 avoids the control based on the estimated operation content of the at least one operating device 9. Specifically, when the operation information is notified from the estimation unit 114, the device control unit 115 refers to the operation information stored in the operation history storage unit 133, and if it is determined that the device 9 that is the operation target indicated by the notified operation information is operating, the generation of the control information for the device 9 based on the notified operation information is avoided.

[0034] When the operation information acquisition unit 116 acquires the operation information transmitted from the device 9, each time it does so, it stores the acquired operation information in the operation history storage unit 133. Further, the operation information acquisition unit 116 identifies the movement route identification information of the movement route determined to have arrived home between the time point when the operation information was acquired and the time point before the preset time from the time point when the operation information was acquired. Then, when the operation information acquisition unit 116 can identify the movement route identification information, it stores the acquired operation information in the operation history storage unit 133 in association with the identified movement route identification information. For example, as shown in FIG. 8, when the user returns home at 19:00, turns on the cooling of the air conditioner 9A, opens and closes the door of the refrigerator 9D, starts cooking rice with the rice cooker 9E, and then starts filling the hot water tank with the hot water supply device 9F every 21:00. In this case, the operation information acquisition unit 116 stores the operation information related to the operations of these series of devices 9 (9A, 9D, 9E, 9F) in the operation history storage unit 133 in association with the movement route information that identifies the movement route immediately before the user returns home at 19:00.

[0035] Returning to FIG. 4, the model generation unit 117 generates a learned model using movement route history information indicating the past movement routes of users and stay period history information indicating the stay periods at locations included in the past movement routes of users, and operation information of each of the past devices 9 corresponding to combinations of the movement route history information and the stay period history information. Specifically, the model generation unit 117 first calculates, using a neural network as shown in FIG. 6 stored in the model storage unit 132, Q values for each combination of a plurality of preset types of device 9 types, operation contents for the device 9, and the time at which the operation contents are executed, from the past movement routes of users and the stay periods at locations included in the past movement routes of users. Next, for each movement situation identified by a plurality of movement situation identification information stored in the operation history storage unit 133, the model generation unit 117 sets the Q value of the combination of the types of a plurality of device 9 types indicated by the operation device information stored in the operation history storage unit 133, the operation contents for the device 9, and the time at which the operation contents are executed, so that the Q value of the combination is higher than the Q values of other combinations. For example, for the movement situation identified by the movement situation identification information "a2", the model generation unit 117 sets the Q value of the combination of the operation target device "refrigerator", the operation content "door opening / closing", and the operation time "18:40" to a number greater than 0, and sets the Q values of other combinations to "0". Then, the model generation unit 117 calculates the error between the Q value calculated using the neural network and the Q value set based on the operation device information stored in the operation history storage unit 133. Then, the model generation unit 117 determines the weight coefficients of the neural network by the error backpropagation method based on the calculated error, and stores information indicating the determined weight coefficients in the model storage unit 132. Further, each time the model generation unit 117 estimates the operation content of the device 9, it calculates the Q value of the combination of the corresponding device 9 type, operation content, and operation time using the update function indicated by the function information stored in the function storage unit 135, and updates the neural network based on the error between the calculated Q value and the Q value of the combination calculated by the learned model before the update.

[0036] The operation content notification unit 118 transmits operation information indicating the operation content estimated by the estimation unit 114, which is notified from the estimation unit 114, to the terminal device 5 held by the user.

[0037] When the change request acquisition unit 119 acquires the above-described change request information from the terminal device 5, it extracts the operation information included in the acquired change request information, and updates the operation information stored in the operation history storage unit 133 based on the extracted operation information. Here, the change request information includes operation information indicating an operation content different from the operation content estimated by the estimation unit 114. Further, the change request acquisition unit 119 notifies the model generation unit 117 that the change request information has been acquired.

[0038] When the function update unit 120 acquires the change request information transmitted from the terminal device 5 by the change request acquisition unit 119, the reward R of the update function for updating the Q value of the combination of the device 9, operation content, and operation time estimated by the estimation unit 114 t+1 is a negative value, for example, "-1", and updates the function information stored in the function storage unit 135. On the other hand, after the function update unit 120 transmits the operation information to the terminal device 5, if the change request acquisition unit 119 does not acquire the change request information within a preset waiting time, the reward R of the update function for updating the Q value of the combination of the device 9, operation content, and operation time estimated by the estimation unit 114 t+1 is a positive value, for example, "1", and updates the function information stored in the function storage unit 135. That is, when the combination of the device 9, operation content, and operation time estimated by the estimation unit 114 is appropriate, the function update unit 120 sets the value of the reward R so that the Q value of the combination becomes larger, and when the combination of the device 9, operation content, and operation time estimated by the estimation unit 114 is not appropriate, the function update unit 120 sets the value of the reward R so that the Q value of the combination becomes smaller. t+1 The value is set, and when the combination of the device 9, operation content, and operation time estimated by the estimation unit 114 is not appropriate, the value of the reward R t+1 is set.

[0039] Next, the operation of the control system according to the present embodiment will be described with reference to FIGS. 9 to 11. In FIGS. 9 to 11, the relay processing at the relay server 3 is omitted for the communication performed between the terminal device 5 and the cloud server 1 via the relay server 3. First, when the preset position information acquisition timing arrives, position information request information for requesting the transmission of the position information of the terminal device 5 is transmitted from the position management server 2 to the terminal device 5 (step S1). On the other hand, when the terminal device 5 acquires the position information request information, the terminal device 5 generates the position information of the terminal device 5 (step S2). Next, the generated position information is transmitted from the terminal device 5 to the position management server 2 (step S3). Here, the position management server 2 manages the position information acquired from the terminal device 5 in association with the terminal device identification information for identifying the terminal device 5.

[0040] Here, it is assumed that the position of the terminal device 5 transitions from a state where it exists inside the area including the user's home set in advance to a state where it exists outside, and the position management server 2 determines that the user has gone out with the terminal device 5 (step S4). In this case, when the preset position information transmission timing arrives, the position information of the terminal device 5 is transmitted from the position management server 2 to the cloud server 1 (step S5). On the other hand, when the cloud server 1 acquires the position information from the position management server 2, it discriminates the movement status of the user based on the acquired position information. Then, when the cloud server 1 discriminates that the user has arrived at or departed from a location included in the movement route, it generates movement status information based on the acquired position information and the time information notified from the timer unit 107 and stores it in the movement status storage unit 131 (step S6). Thereafter, a series of processes of steps S5 and S6 are executed every time the position information transmission timing arrives.

[0041] Subsequently, when the preset position information transmission timing arrives, the position information of the terminal device 5 is transmitted from the position management server 2 to the cloud server 1 (step S7). Based on the position information obtained from the position management server 2, the cloud server 1 determines the movement status of the user, generates movement status information, and stores it in the movement status storage unit 131. After that (step S8), it is assumed that the cloud server 1 determines that the user has returned home (step S9). In this case, home arrival notification information for notifying that the user has returned home is transmitted from the cloud server 1 to the position management server 2 (step S10). Here, when the position management server 2 acquires the home arrival notification information, it stops the periodic transmission of the position information of the terminal device 5 to the cloud server 1.

[0042] Thereafter, the cloud server 1 estimates the operation content of each device 9 using a learned model that estimates the operation content of each device 9 from the movement route of the user and the stay time of the user at the locations included in the movement route (step S11). Next, operation information indicating the estimated operation content is transmitted from the cloud server 1 to the terminal device 5 (step S12). On the other hand, when the terminal device 5 acquires the operation information, it generates an operation screen image including the acquired operation information and displays it on the display unit 504 (step S13).

[0043] Next, as shown in FIG. 10, after the cloud server 1 transmits the operation information to the terminal device 5, if the change request acquisition unit 119 does not acquire the change request information within the preset waiting time, the reward R of the update function for updating the Q value of the combination of the device 9, the operation content, and the operation time estimated by the estimation unit 114 t+1The function information stored in the function storage unit 135 is updated so that it becomes a positive value (step S14). Subsequently, the cloud server 1 calculates the Q value of the combination of the type of the device 9, the operation content, and the operation time using the updated function indicated by the updated function information, and updates the learned model based on the calculated Q value (step S15). After that, when the device operation timing indicated by the operation information arrives, the cloud server 1 generates control information for controlling the operation of the device 9 based on the operation information (step S16). Next, the generated control information is transmitted from the cloud server 1 to the device 9 (step S17). On the other hand, when the device 9 acquires the control information, it operates based on the acquired control information (step S18).

[0044] Also, when the preset model update timing arrives, the cloud server 1 uses the past movement route information and stay period information of the user stored in the movement status storage unit 131 and the past operation information stored in the operation history storage unit 133 to newly generate a learned model and store it in the model storage unit 132 (step S19).

[0045] After that, assuming that the preset position information transmission timing has arrived and the series of processes from steps S5 to S13 described above have been executed. Here, it is determined that the operation information displayed on the display unit 504 of the terminal device 5 is different from the operation content intended by the user, and as shown in FIG. 11, it is assumed that the user has performed an operation content change request operation for requesting a change in the operation content of the device 9 via the input unit 505. In this case, the terminal device 5 accepts the change request operation (step S20) and generates the above-described change request information (step S21). Next, the generated change request information is transmitted from the terminal device 5 to the cloud server 1 (step S22). On the other hand, when the cloud server 1 acquires the change request information, the reward R in the update function of the update function for updating the Q value of the combination of the device 9, the operation content, and the operation time estimated by the estimator 114 t+1The function information stored by the function storage unit 135 is updated so that it becomes a negative value (step S23). Subsequently, the cloud server 1 extracts operation information indicating the operation content intended by the user from the acquired change request information (step S24). Thereafter, the cloud server 1 calculates the Q value of the combination of the type, operation content, and operation time of the device 9 indicated by the operation information extracted from the change request information using the updated function indicated by the updated function information, and updates the learned model based on the calculated Q value (step S25). Next, the cloud server 1 generates control information based on the updated learned model (step S26). Next, the generated control information is transmitted from the cloud server 1 to the device 9 (step S15), and the device 9 operates based on the control information acquired from the cloud server 1 (step S16).

[0046] Next, the device control process executed by the cloud server 1 according to the present embodiment will be described with reference to FIG. 12. This device control process is started, for example, when a program for executing the device control process is started in the cloud server 1. First, the model generation unit 117 determines whether or not a preset model update time has arrived based on the time information notified from the timer unit 107 (step S101). If the model generation unit 117 determines that the model update time has not yet arrived (step S101: No), the process of step S103 described later is executed. On the other hand, when the model generation unit 117 determines that the model update time has arrived (step S101: Yes), a new learned model is generated using the past movement route information and stay time information of the user stored in the movement status storage unit 131 and the past operation information stored in the operation history storage unit 133, and the generated learned model is stored in the model storage unit 132 (step S102).

[0047] Next, the position information acquisition unit 111 determines whether it has acquired the position information transmitted from the position management server 2 (step S103). If the position information acquisition unit 111 determines that it has not yet acquired the position information (step S103: No), the process of step S105 described below is executed. On the other hand, if the position information acquisition unit 111 determines that it has acquired the position information (step S103: Yes), the movement status determination unit 112 determines the movement status of the user based on the position information and generates movement status information. Then, the movement status determination unit 112 stores the generated movement status information in the movement status storage unit 131 (step S104). Subsequently, the movement status determination unit 112 determines whether the user has returned home based on the generated movement status information (step S105). If the movement status determination unit 112 determines that the user has not yet returned home (step S105: No), the process of step S103 is executed again.

[0048] On the other hand, if the movement status determination unit 112 determines that the user has returned home (step S105: Yes), the home arrival notification unit 113 transmits home arrival notification information to the position management server 2 (step S106). Thereafter, the estimation unit 114 estimates the operation content of each device 9 from the movement route of the user and the stay time of the user at the locations included in the movement route using the learned model stored in the model storage unit 132 (step S107). Next, the operation content notification unit 118 transmits operation information indicating the operation content estimated by the estimation unit 114 to the terminal device 5 (step S108).

[0049] Subsequently, after transmitting the operation information to the terminal device 5, the change request acquisition unit 119 determines whether it has acquired the change request information transmitted from the terminal device 5 within a preset waiting time (step S109). If, after transmitting the operation information to the terminal device 5, the change request acquisition unit 119 does not acquire the change request information and the above-mentioned waiting time elapses (step S109: No), the function update unit 120 calculates the reward R of the update function for updating the Q value of the combination of the device 9, the operation content, and the operation time estimated by the estimation unit 114. t+1The function information stored in the function storage unit 135 is updated so that it becomes a positive value (step S110). After that, the process of step S113 described later is executed.

[0050] On the other hand, assume that after the change request acquisition unit 119 transmits the operation information to the terminal device 5, it determines that it has acquired the change request information transmitted from the terminal device 5 within the aforementioned waiting time (step S109: Yes). In this case, the function update unit 120 updates the reward R of the update function for updating the Q value of the combination of the device 9, the operation content, and the operation time estimated by the estimation unit 114 t+1 The function information stored in the function storage unit 135 is updated so that it becomes a negative value (step S111). Next, the change request acquisition unit 119 extracts the operation information from the acquired change request information (step S112). Subsequently, the model generation unit 117 calculates the Q value of the combination of the type of the device 9, the operation content, and the operation time using the update function indicated by the function information stored in the function storage unit 135, and updates the learned model stored in the model storage unit 132 based on the error between the calculated Q value and the Q value of the combination calculated by the learned model before the update (step S113).

[0051] After that, the device control unit 115 refers to the operation information stored in the operation history storage unit 133 and determines whether the device 9 to be operated, estimated by the estimation unit 114, is in operation (step S114). Here, when the device control unit 115 determines that the device 9 to be operated, estimated by the estimation unit 114, is in operation (step S114: Yes), the process of step S101 is executed again as it is. On the other hand, when the device control unit 115 determines that the device 9 to be operated, estimated by the estimation unit 114, is stopped (step S114: No), control information for controlling the device 9 to be operated is generated based on the operation content indicated by the operation content estimated by the estimation unit 114 or the operation information extracted by the change request acquisition unit 119, and is transmitted to the device 9 (step S115). Next, the process of step S101 is executed again.

[0052] As described above, according to the control device according to the present embodiment, the estimation unit 114 estimates the operation content of each of the devices 9 from the movement route of the user and the staying time of the user at at least one location included in the movement route of the user using the above-described learned model. Then, the device control unit 115 controls the operations of the devices 9 based on the operation content of each of the devices 9 estimated by the estimation unit 114. Thereby, the devices 9 can be appropriately controlled based on the positions and action patterns of the users of the devices 9.

[0053] Incidentally, in recent years, the IoT (Internet of Things) of devices used in houses has advanced, and a device cooperation system that provides a service for controlling devices in a home via a wide area network from a cloud server is being provided. Some of these types of services control devices according to scenarios pre-embedded on a scenario platform triggered by unconscious daily actions (unconscious actions) of the user or changes in the living environment (environmental changes, physical changes). In this case, it is necessary to prepare a plurality of scenarios in advance for each user preference or living environment. When the user preferences or living environments are diverse, the number of scenarios to be prepared becomes enormous, and there is a risk that the burden on the user for setting scenarios will become extremely large. Also, when a new device is added to the house, the user may need to update the scenario each time, increasing the burden on the user. On the other hand, according to the control device according to the present embodiment, since the devices 9 to be operated, the operation content of the devices 9, and the operation timing can be specified automatically according to the action pattern of the user, the burden on the user can be significantly reduced.

[0054] In addition, each time the preset model update time arrives, the model generation unit 117 according to the present embodiment generates a learned model by using movement route history information indicating the movement routes of past users and stay time history information indicating the stay times of past users at at least one location, and operation information indicating the operation content of each of at least one past device corresponding to the combination of the movement route history information and the stay time history information. As a result, since the learned model is periodically updated based on the most recent behavior pattern of the user, it becomes possible to control the operation of the device 9 in a form that conforms to the most recent behavior pattern of the user.

[0055] Furthermore, the operation content notification unit 118 according to the present embodiment transmits operation information indicating the operation content of the device 9 estimated by the estimation unit 114 to the terminal device 5 held by the user. As a result, the user can recognize in advance the operation content of each of the devices 9 that the cloud server 1 attempts to control, so that it is possible to suppress the device 9 from being operated with an operation content unintended by the user.

[0056] Also, when the change request acquisition unit 119 according to the present embodiment acquires change request information from the terminal device 5, the model generation unit 117 updates the learned model by using the operation information included in the acquired change request information. As a result, it becomes possible to control the operation of the device 9 in a form that conforms to the most recent behavior pattern of the user.

[0057] Furthermore, when estimating the operation content by the estimation unit 114, if there is at least one operating device 9, the device control unit 115 according to the present embodiment avoids control based on the estimated operation content for the at least one operating device 9. As a result, for example, it is possible to prevent the operation content of the device 9 that is already in use in the building H from being changed when the user arrives at the building H, so that the convenience of the user can be improved particularly when a plurality of users use the building H.

[0058] (Embodiment 2) The control device according to the present embodiment is different from the control device according to Embodiment 1 in that it estimates the operation content of the device using, in addition to the user's movement route and the staying time of the user at each at least one location included in the movement route, the map information of at least one location included in the user's movement route and the weather information indicating the weather conditions at the user's destination.

[0059] For example, as shown in FIG. 13, the cloud server 2001 according to the present embodiment is communicable via a map server 2004 that manages map information and a weather server 2005 that manages weather information and a wide area network NW1. In FIG. 13, the same components as those in Embodiment 1 are denoted by the same reference numerals as those in FIG. 1. The map server 2004 manages map information including information indicating the type of each location where buildings such as houses and companies are arranged in each range when the earth is divided into a preset number of ranges, in association with the position information indicating the range. When the map server 2004 acquires map request information including the position information of at least one location included in the user's movement route from the cloud server 2001, it selects map information including all the positions indicated by the position information included in the acquired map request information and transmits it to the cloud server 2001.

[0060] The weather server 2005 manages weather information indicating the weather conditions at each location on the earth, in association with the position information of that location. When the weather server 2005 acquires weather information request information including the position information of at least one location included in the user's movement route from the cloud server 2001, it selects the weather information corresponding to the position information included in the acquired weather information request information and transmits it to the cloud server 2001.

[0061] The hardware configuration of the cloud server 2001 according to this embodiment is the same as that of the cloud server 1 according to Embodiment 1 shown in FIG. 2. Hereinafter, the hardware configuration according to this embodiment will be described using the same reference numerals as those used in the description of Embodiment 1. The CPU 101 of the cloud server 1 reads the program stored in the auxiliary storage unit 103 into the main storage unit 102 and executes it, so that, as shown in FIG. 14, a position information acquisition unit 111, a movement situation determination unit 112, a home arrival notification unit 113, an estimation unit 2114, a device control unit 115, an operation information acquisition unit 116, a model generation unit 2117, an operation content notification unit 118, a change request acquisition unit 119, a location type specification unit 2120, and a weather information acquisition unit 2121 function. In FIG. 14, the same components as those in Embodiment 1 are denoted by the same reference numerals as those in FIG. 4. Further, the auxiliary storage unit 103 includes a movement situation storage unit 2131, a model storage unit 132, an operation history storage unit 133, a location storage unit 134, and a weather history storage unit 2136.

[0062] The movement situation storage unit 2131 stores, for example, as shown in FIG. 15, movement situation information including the movement route information of the user, the stay time information of the user at at least one location, and location type information indicating the type of each of at least one location, in association with movement situation identification information. The location type information is information indicating the type when each of at least one location is classified into types such as a company and a commercial facility. For example, to the position information Pos[0], Pos[1], Pos[2] constituting the movement route information identified by the movement situation identification information "a2", location type information CP[0] indicating the type "residence", location type information CP[1] indicating the type "company", and location type information CP[2] indicating the type "commercial facility" are associated respectively. The weather history storage unit 2136 stores, in time series, weather information indicating the weather conditions at each of at least one preset location, in association with the position information of each location.

[0063] The location type specifying unit 2120 generates map request information including the location information of at least one location included in the user's movement route, and acquires map information from the map server 2004 by transmitting the generated map request information to the map server 2004. Then, the location type specifying unit 2120 specifies the type of each of the at least one location based on the acquired map information. The location type specifying unit 2120 stores the location type information indicating the specified type in the movement status storage unit 2131 in association with the location information of each of the at least one location included in the user's movement route.

[0064] The weather information acquisition unit 2121 generates weather information request information including the location information of at least one location included in the user's movement route, and acquires weather information from the weather server 2005 by transmitting the generated weather information request information to the weather server 2005. Then, the weather information acquisition unit 2121 stores the acquired weather information in the weather history storage unit 2136 in time series in association with the location information of each of the at least one location.

[0065] The estimation unit 2114 estimates the operation content of each of the devices 9 using a learned model for estimating the operation content of each of the devices 9 from the location type and weather conditions of each of the at least one location, together with the stay period at the at least one location indicated by the movement route information and the stay period information.

[0066] The model generation unit 2117 generates a learned model using the past location type information and weather history information together with the past movement route history information, stay time history information, and operation information of the users described in the first embodiment. Specifically, the model generation unit 2117 first uses the neural network described in the first embodiment to identify, from the past movement route, stay time, location type, and weather conditions of the users, the number for identifying the device 9 to be operated, the time for operating the device 9 to be operated, and the number for identifying the operation content for the device 9. Next, the model generation unit 2117 calculates the error between the calculated various values, the number for identifying the device 9 indicated by the operation device information stored in the operation history storage unit 133, the time for operating the device 9 indicated by the operation time information, and the number for identifying the operation content for the device 9 indicated by the operation information. Then, the model generation unit 2117 determines the weight coefficients of the neural network by the error backpropagation method based on the calculated error.

[0067] Next, the operation of the control system according to the present embodiment will be described with reference to FIG. 16. In FIG. 16, the same reference numerals as those in FIGS. 9 to 11 are given to the same processes as those in the first embodiment. First, after a series of processes from step S7 to step S10 are executed, the cloud server 2001 generates map request information including the position information of at least one location included in the movement route of the user (step S2001). Next, the generated map request information is transmitted from the cloud server 2001 to the map server 2004 (step S2002). On the other hand, when the map server 2004 acquires the map request information, the map server 2004 selects map information including all the positions indicated by the position information included in the acquired map request information from the map information managed by the map server 2004 (step S2003). Subsequently, the selected map information is transmitted from the map server 2004 to the cloud server 2001 (step S2004). On the other hand, when the cloud server 2001 acquires the map information, based on the acquired map information, the cloud server 2001 specifies the type of each of at least one location included in the movement route of the user, and stores the location type information indicating the specified type in the movement status storage unit 2131 (step S2005).

[0068] Next, the cloud server 2001 generates weather information request information including the location information of at least one location included in the user's movement route (step S2006). Subsequently, the generated weather information request information is transmitted from the cloud server 2001 to the weather server 2005 (step S2007). On the other hand, when the weather server 2005 acquires the weather information request information, it selects the weather information corresponding to the location information included in the acquired weather information request information (step S2008). Subsequently, the selected weather information is transmitted from the weather server 2005 to the cloud server 2001 (step S2009). Here, when the cloud server 2001 acquires the weather information, it stores the acquired weather information in the weather history storage unit 2136.

[0069] Thereafter, the cloud server 2001 estimates the operation content of each device 9 using a learned model that estimates the operation content of each device 9 from the user's movement route, stay period, location type, and weather conditions (step S2010). Next, the processes after step S12 described in Embodiment 1 are executed.

[0070] Next, the device control process executed by the cloud server 2001 according to this embodiment will be described with reference to FIG. 17. In FIG. 17, the same reference numerals as those in FIG. 12 are assigned to the same processes as those in Embodiment 1. First, it is assumed that a series of processes from step S101 to S106 have been executed. In this case, after the process of step S106, the location type specifying unit 2120 generates map request information including the location information of at least one location included in the user's movement route, and transmits the generated map request information to the map server 2004 (step S2101), thereby acquiring the map information transmitted from the map server 2004 (step S2102). Next, the location type specifying unit 2120 specifies the type of each at least one location based on the acquired map information, and associates the location type information indicating the specified type with the location information of each at least one location included in the user's movement route and stores it in the movement status storage unit 2131 (step S2103).

[0071] Subsequently, the weather information acquisition unit 2121 generates weather information request information including the location information of at least one location included in the user's travel route, and transmits the generated weather information request information to the weather server 2005 (step S2104) to obtain weather information from the weather server 2005. Then, the weather information acquisition unit 2121 stores the acquired weather information in the weather history storage unit 2136 in chronological order in association with the location information of each of the at least one location (step S2105).

[0072] After that, the estimation unit 2114 estimates the operation content of each device 9 from the user's travel route, stay period, location type, and weather conditions using the learned model stored in the model storage unit 132 (step S2106). Next, the processes of steps S108 and S109 are executed. In the process of step S109, it is assumed that the change request acquisition unit 119 determines that the change request information has been acquired (step S109: Yes). In this case, the function update unit 120 updates the reward R of the update function for updating the Q value of the combination of the device 9, operation content, and operation time estimated by the estimation unit 114. t+1 The function information stored in the function storage unit 135 is updated so that it becomes a negative value (step S111). Next, the change request acquisition unit 119 extracts operation information from the acquired change request information (step S112). Subsequently, the model generation unit 117 calculates the Q value of the combination of the device 9 type, operation content, and operation time using the update function indicated by the function information stored in the function storage unit 135, and updates the learned model stored in the model storage unit 132 based on the error between the calculated Q value and the Q value of the combination calculated by the learned model before the update (step S2107). Next, the processes after step S114 are executed.

[0073] As described above, for the cloud server 2001 according to the present embodiment, the operation content of the device 9 is estimated based on at least one type of location included in the user's movement route and the weather conditions of each location. Thereby, since the influence of the type of location included in the user's movement route and the weather conditions of each location on the operation content of the user's device 9 is reflected, the device 9 can be controlled in a form adapted to the user's behavior pattern.

[0074] As described above, the embodiments of the present disclosure have been described. However, the present disclosure is not limited to the above-described embodiments. For example, the cloud server may be able to change the operation content of the device 9 according to the movement routes of a plurality of users respectively. The control system according to this modification example may be such that, as shown in FIG. 18 for example, the cloud server 3001 identifies the movement routes of the users respectively possessing the terminal devices 5A and 5B based on the position information of the plurality of terminal devices 5 (5A, 5B). Here, when the cloud server 3001 controls a plurality of devices 9 installed in a building H as shown in FIG. 19(A) for example, the operation content may be estimated using separate learned models for the movement route information of each of the terminal devices 5A and 5B. In this case, as shown in FIG. 19(B) for example, the model storage unit 3131 stores, as the learned model corresponding to the terminal device 5A, a learned model for estimating the operation content of the air conditioners 9A, 9B, the refrigerator 9D, and the rice cooker 9E, and stores, as the learned model corresponding to the terminal device 5B, a learned model for estimating the operation content of the air conditioners 9A, 9C, the refrigerator 9D, and the water heater 9F.

[0075] According to this configuration, the operation of the device 9 can be appropriately controlled for each user based on each of the movement routes of a plurality of users living together in the same building H.

[0076] In each embodiment, the model generation units 117 and 2117 may generate a learned model based on the movement routes of the terminal devices held by the residents of each of the plurality of buildings H, the staying times at at least one location included in the movement routes, and the operation contents of each of the devices 9 by the residents. Further, the model generation units 117 and 2117 may generate a learned model based on the movement routes of the terminal devices held by the residents of each of the plurality of buildings H located within a preset area, the staying times at at least one location included in the movement routes, and the operation contents of each of the devices 9 by the residents. Furthermore, the model generation units 117 and 2117 may generate a learned model for determining the operation contents of the devices 9 installed in the building H located within a preset area, based on the movement routes of the terminal devices held by the residents of the buildings located in other areas, the staying times at at least one location included in the movement routes, and the operation contents of each of the devices 9 by the residents.

[0077] In each embodiment, the model generation units 117 and 2117 may exclude a part of the past movement route information and staying time information, and the past operation information of each of the devices 9, from the teacher data used for generating the learned model, or may add the past movement route information and staying time information, and the past operation information of each of the devices 9, which have not been used as teacher data, to the teacher data. For example, the model generation units 117 and 2117 may generate a learned model without using the operation information when the user was located at any of at least one location in the past.

[0078] In each embodiment, the model generation units 117 and 2117 may regenerate a learned model based on the movement routes of the terminal devices held by the residents of the building H where the device 9 for determining the operation contents is installed, the staying times at at least one location included in the movement routes, and the operation contents of each of the devices 9 by the residents, using a learned model for determining the operation contents of the devices 9 in another building different from the building H where the device 9 for determining the operation contents is installed.

[0079] In each embodiment, an example has been described in which the estimation unit 114, the model generation unit 117, and the model storage unit 132 are realized in the same cloud server 1. However, the present invention is not limited to this. For example, the model generation unit 117 and the model storage unit 132 may be realized in a device different from the estimation unit 114.

[0080] In each embodiment, an example has been described in which the estimation unit 114 estimates the operation content using a feedforward neural network. However, the present invention is not limited to this. For example, a recurrent neural network may be used to estimate a combination of the device 9 to be operated, the operation content of the device 9, and the operation time. In this case, the user can estimate a combination of the device 9 to be operated, the operation content of the device 9, and the operation time at a stage before arriving home. Further, the estimation unit 114 may estimate a combination of the device 9 to be operated, the operation content of the device 9, and the operation time using, for example, a multi-class classification type support vector machine.

[0081] Also, various functions of the cloud servers 1, 2001, 3001 and the terminal device 5 according to the present disclosure may be realized by software, firmware, or a combination of software and firmware. In this case, the software or firmware is described as a program, and the program is stored and distributed in a computer-readable recording medium such as a flexible disk, a CD-ROM (Compact Disc Read Only Memory), a DVD (Digital Versatile Disc), and an MO (Magneto-Optical Disc). By reading and installing the program in a computer, a computer capable of realizing the above-described respective functions may be configured. When each function is realized by sharing between an OS (Operating System) and an application, or by cooperation between the OS and the application, only a part other than the OS may be stored in the recording medium.

[0082] Furthermore, it is also possible to superimpose each program on a carrier wave and distribute it via a network. For example, the program may be posted on a bulletin board system (BBS) on the network and distributed via the network. Then, these programs are started and executed under the control of the OS in the same manner as other application programs so that the above-described processing can be executed.

Industrial Applicability

[0083] The present disclosure is suitable as a control system that controls the operation of a device according to a movement route from when a user goes out of the house until returning home.

Explanation of Signs

[0084] 1,2001,3001 Cloud server, 2 Location management server, 3 Relay server, 4,4004 Substation equipment management device, 5 Terminal device, 101,501 CPU, 102,502 Main memory, 103,503 Auxiliary memory, 106,506 Wide area communication unit, 107 Timing unit, 109,509 Bus, 111 Location information acquisition unit, 112 Movement status determination unit, 113 Return home notification unit, 114,2114 Estimation unit, 115 Device control unit, 116 Operation information acquisition unit, 117,2117 Model generation unit, 118 Operation content notification unit, 119 Change request acquisition unit, 120 Function update unit, 131,2131 Movement status memory unit, 132 Model memory unit, 133 Operation history memory unit, 134 Location memory unit, 135 Function memory unit, 504 Display unit, 505 Input unit, 508 GNSS receiver, 511 Reception unit, 512 Display control unit, 513 Positioning unit, 514 Location information notification unit, 515 Operation content acquisition unit, 516 Change request generation unit, 517 Change request transmission unit, 521 Location memory unit, 531 Image memory unit, 2004 Map server, 2005 Weather server, 2120 Location type identification unit, 2121 Weather information acquisition unit, 2136 Weather history memory unit, NW1 Wide area network, NW2 Local network

Claims

1. A movement status determination unit that determines the movement route of a user and the stay time of the user at at least one preset location included in the movement route, An estimation unit that estimates the operation content and operation time of each of at least one device installed at a location different from the at least one location from the movement route and the stay time at the at least one location determined by the movement status determination unit, using a learned model, A device control unit that controls each of the at least one device based on the operation content and operation time of each of the at least one device estimated by the estimation unit. A control device comprising: Control device.

2. Each time a preset model update time arrives, using movement route history information indicating the movement routes of past users and stay time history information indicating the stay time of the past users at the at least one location included in the movement routes of the past users, and operation information indicating the operation content and operation time of each of the at least one device in the past corresponding to the combination of the movement route history information and the stay time history information, further comprising a model generation unit that generates the learned model, The control device according to claim 1.

3. Further comprising an operation content notification unit that transmits operation information indicating the operation content estimated by the estimation unit to a terminal device held by the user, The control device according to claim 2.

4. Further comprising a change request acquisition unit that acquires change request information including operation information indicating an operation content different from the operation content indicated by the operation information from the terminal device and requesting the control device to change the operation content of the at least one device, When the change request acquisition unit acquires the change request information, the model generation unit updates the learned model using the operation information included in the acquired change request information. The control device according to claim 3.

5. A location type identification unit that acquires map information of the at least one location and identifies the type of the at least one location based on the acquired map information, Further comprising a weather information acquisition unit that acquires weather information indicating weather conditions at the at least one location. The estimation unit uses a learned model for estimating the operation content and operation time of each of the at least one device from the type of the at least one location specified based on the map information, the weather conditions indicated by the weather information, together with the movement route and the stay time of the user at the at least one location included in the movement route, to estimate the operation content and operation time of each of the at least one device. The control device according to claim 1.

6. The model generation unit further includes a movement route history information indicating a past movement route of a user, a stay time history information indicating a stay time at the at least one location included in the past movement route of the user, a weather history information indicating weather conditions at the past at least one location, map information of the at least one location included in the past movement route of the user, and operation information indicating the operation content and operation time of each of the at least one device in the past corresponding to a combination of the movement route history information, the stay time history information, the weather history information, and the map information, and generates the learned model using the above. The control device according to claim 5.

7. The estimation unit uses the learned model different for each user to estimate the operation content and operation time. The control device according to any one of claims 1 to 6.

8. The estimation unit uses the learned model for estimating the operation content and operation time of each of the at least one device associated with the user to estimate the operation content and operation time of each of the at least one device associated with the user. The control device according to claim 7.

9. When there is at least one operating device at the time of estimating the operation content by the estimation unit, the device control unit avoids controlling the at least one operating device based on the operation content of the at least one operating device estimated by the estimation unit. The control device according to any one of claims 1 to 8.

10. A movement situation determination unit that determines a movement route of a user and a stay time of the user at at least one preset location included in the movement route. A model generation unit that generates a learned model for estimating the operation content and operation time of each of at least one device installed at a location different from the at least one location, using movement route history information indicating the movement routes of past users and stay time history information indicating the stay times of past users at the at least one location of the past users, and operation information indicating the operation content and operation time of each of at least one device corresponding to the combination of the movement route history information and the stay time history information. Learning device.

11. A terminal device possessed by a user, A movement situation determination unit that determines the movement route of the user and the stay time of the user at at least one preset location included in the movement route, An estimation unit that estimates the operation content and operation time of each of at least one device installed at a location different from the at least one location, using a learned model that estimates the operation content and operation time of each of at least one device installed at a location different from the at least one location, based on the movement route and the stay time at the at least one location determined by the movement situation determination unit. A device control unit that controls each of the at least one device based on the operation content and operation time of each of the at least one device estimated by the estimation unit. An operation content notification unit that transmits operation information indicating the operation content estimated by the estimation unit to the terminal device. Control system.

12. A step in which a control device determines the movement route of a user and the stay time of the user at at least one preset location included in the movement route. A step in which the control device estimates the operation content and operation time of each of at least one device installed at a location different from the at least one location, using a learned model that estimates the operation content and operation time of each of at least one device installed at a location different from the at least one location, based on the determined movement route and the stay time at the at least one location. A step in which the control device controls each of the at least one device based on the estimated operation content and operation time of each of the at least one device. Device control method.

13. A computer, a movement situation determination unit that determines a user's movement route and the user's staying time at at least one preset location included in the movement route, an estimation unit that estimates the operation content and operation time of each of at least one device installed at a location different from the at least one location, using a learned model that estimates the operation content and operation time of each of at least one device installed at a location different from the at least one location from the movement route and the staying time at the at least one location determined by the movement situation determination unit, a device control unit that controls each of the at least one device based on the operation content and operation time of each of the at least one device estimated by the estimation unit, a program for causing the computer to function as such.

Images