JP7138800B2 - Control system, server, device, control method and program - Google Patents

Description

The present invention relates to control systems, servers, devices, control methods, and programs.

Based on the image or thermal image data captured by the air conditioner, an individual comfort range that each occupant in the room where the air conditioner is installed acquires, and based on the acquired individual comfort range, An air conditioner system has been proposed that learns the combination of room occupants and common comfort ranges by repeatedly extracting a common comfort range, which is a common range in the individual comfort ranges of all room occupants (for example, See Patent Document 1).

WO2018/163272

However, in the air conditioning system proposed in Patent Document 1, for example, when the air conditioner is updated with a new air conditioner, the new air conditioner is used to create a common comfort range with the occupants in the room. It is necessary to re-learn the combination of and, and it may take time to bring the new air conditioner into the optimal operating state for the person in the room.

The present invention has been made in view of the above reasons, and provides a control system, server, device, control method, and program that can continue the tendency of the operating state of the device before the device is updated even when the device is updated. With the goal.

In order to achieve the above object, the control system according to the present invention includes:
comprising a server and at least one device;
The server is
Coefficient information including information indicating neural network coefficients of a neural network for obtaining future device setting parameters of the device having a preset number of nodes and number of layers, and coefficient attribute information indicating attributes of the coefficient information, a first coefficient acquisition unit that acquires from the at least one device;
a first coefficient transmission unit configured to transmit the coefficient information and the coefficient attribute information acquired by the first coefficient acquisition unit to the at least one device when the coefficient request information is acquired from the at least one device. death,
Each of the at least one device comprises:
a second coefficient transmission unit configured to transmit, to the server, coefficient information indicating neural network coefficients of the neural networks used in each of the at least one devices, and coefficient attribute information;
a second coefficient acquisition unit configured to acquire coefficient information and coefficient attribute information from the server by transmitting coefficient request information requesting transmission of the coefficient information to the server;
a coefficient setting unit configured to set neural network coefficients of a neural network based on the coefficient information and the coefficient attribute information acquired by the second coefficient acquisition unit;
a neuro engine for obtaining future device setting parameters of the device from environmental parameters indicating the current environment using the neural network in which the neural network coefficients are set;
and a device control unit that controls the device based on the device setting parameters obtained by the neuro engine.

According to the present invention, in the server, the first coefficient acquisition unit stores the coefficient information including the information indicating the neural network coefficients of the neural network for obtaining the future device setting parameters of the device and the coefficient attribute information into at least one When the first coefficient transmission unit acquires the coefficient request information from at least one device, the coefficient information and the coefficient attribute information acquired by the first coefficient acquisition unit are transmitted to at least one device. On the other hand, in each of at least one device, a second coefficient transmission unit transmits coefficient information indicating neural network coefficients of a neural network used in each of at least one device and coefficient attribute information to a server, and obtains a second coefficient. A unit acquires coefficient information and coefficient attribute information from the server by transmitting coefficient request information to the server. Then, the coefficient setting unit sets the neural network coefficients of the neural network based on the coefficient information and the coefficient attribute information acquired by the second coefficient acquisition unit. As a result, for example, before updating the device, the coefficient information including the information indicating the neural network coefficients of the neural network used in the device and the coefficient attribute information are transmitted to the server, and after the device is updated, the coefficient attribute information is sent to the server. By transmitting the coefficient request information and acquiring the coefficient information from the server, it is possible to set the pre-update neural network coefficients for the neural network used in the updated device. Therefore, even when the equipment is updated, the trend of the operating state of the equipment before the update can be continued in the equipment after the update.

Schematic configuration diagram of a control system according to an embodiment of the present invention A block diagram showing a hardware configuration of an air conditioner according to an embodiment. 1 is a block diagram showing the functional configuration of an air conditioner according to an embodiment; FIG. 4 shows an example of information stored in a history information storage unit according to the embodiment; Explanatory diagram of a neural network used in the air conditioner according to the embodiment 1 is a block diagram showing the configuration of a neuro engine according to an embodiment; FIG. FIG. 2 is a block diagram showing the hardware configuration of the cloud server according to the embodiment; Block diagram showing the functional configuration of the cloud server according to the embodiment FIG. 2 is a block diagram showing the hardware configuration of the management device according to the embodiment; 1 is a block diagram showing the configuration of a management device according to an embodiment; FIG. 4 is a sequence diagram showing an example of the operation of the control system according to the embodiment; FIG. A diagram showing an example of history attribute information according to the embodiment A diagram showing an example of coefficient attribute information according to the embodiment 4 is a sequence diagram showing an example of the operation of the control system according to the embodiment; FIG. 4 is a sequence diagram showing an example of the operation of the control system according to the embodiment; FIG. FIG. 4 is a diagram showing an example of an operation screen image displayed on the display unit of the management device according to the embodiment; FIG. 4 is a diagram showing an example of an operation screen image displayed on the display unit of the management device according to the embodiment; FIG. 4 is a diagram showing an example of an operation screen image displayed on the display unit of the management device according to the embodiment; FIG. 4 is a diagram showing an example of an operation screen image displayed on the display unit of the management device according to the embodiment; 4 is a sequence diagram showing an example of the operation of the control system according to the embodiment; FIG. Flowchart showing an example of the flow of device control processing executed by the air conditioner according to the embodiment Flowchart showing an example of the flow of coefficient determination processing executed by the air conditioner according to the embodiment Flowchart showing an example of the flow of device control processing executed by the air conditioner according to the embodiment Flowchart showing an example of the flow of device control processing executed by the air conditioner according to the embodiment Flowchart showing an example of the flow of coefficient information generation processing executed by the cloud server according to the embodiment FIG. 4 is a diagram showing transition of information held by each of the air conditioner and the cloud server according to the embodiment; A diagram showing an example of an operation screen image displayed on the display unit of the management apparatus according to the modification. A diagram showing an example of an operation screen image displayed on the display unit of the management apparatus according to the modification. A diagram showing an example of an operation screen image displayed on the display unit of the management apparatus according to the modification. Sequence diagram showing an example of the operation of the control system according to the modification Sequence diagram showing an example of the operation of the control system according to the modification Sequence diagram showing an example of the operation of the control system according to the modification 1 is a block diagram showing the configuration of an air conditioner according to an embodiment; Block diagram showing configuration of cloud server according to embodiment

A control system according to an embodiment of the present invention will be described in detail below with reference to the drawings. The control system according to the present embodiment, based on user information about the user of the device, uses environmental parameters indicating the environment of the location where the device is installed and weather forecast information indicating future weather conditions to generate a neural network. is used to determine device configuration parameters for future devices. Further, in the control system according to the present embodiment, the device determines the weighting coefficient of the neural network for obtaining the future device setting parameters of the device having the preset number of nodes and the number of layers, and the weighting coefficient is Using the determined neural network, it has a function of obtaining future equipment setting parameters of its own equipment.

The control system according to the present embodiment, as shown in FIG. and a cloud server 2 that can communicate via NT1. The outside network NT1 is, for example, the Internet. Also, the outside network NT1 is connected to a weather server 3 that distributes actual weather information indicating current or past weather conditions and weather forecast information indicating future weather conditions. Furthermore, a customer server 9 that manages customers who have purchased air conditioners, for example, is connected to the outside network NT1. The customer server 9 associates the history information including the history of the device setting information of the air conditioner purchased by the customer, the environmental information indicating the environmental parameter including the temperature information, and the device identification information for identifying the air conditioner. A storage (not shown) is provided for storing. Every time the customer server 9 periodically receives history information from the air conditioner purchased by the customer, the customer server 9 associates the received history information with the device identification information and stores it in the storage.

Further, upon receiving the history request information from the cloud server 2, the customer server 9 generates history information corresponding to the history request information and history attribute information corresponding thereto. Here, the client server 9 generates history attribute information whose file format is, for example, a JSON schema file format, and also generates attribute information whose file format is a JSON file format. The customer server 9 then transmits the generated history information and history attribute information to the cloud server 2 . Here, the customer server 9 performs reversible information compression processing on the history information and the history attribute information before transmitting.

Operating devices 61, 62, and 63 for operating the air conditioners 41, 42, and 43 are arranged in the rooms R1, R2, and R3 in the dwelling unit H, respectively. Here, the rooms R1, R2, and R3 are, for example, a living room, a bedroom, and a child's room in the dwelling unit H, respectively. Also, in the dwelling unit H, a router 82 connected to the home network NT2 and a data line termination device 81 connected to the router 82 and the outside network NT1 are installed. The home network NT2 is, for example, a wired LAN (Local Area Network) or a wireless LAN. The data line terminating device 81 is a modem, gateway, or the like.

Each of the air conditioners 41, 42, and 43, as shown in FIG. have Note that the measuring device 461 is not limited to measuring the temperature of the room, and may measure other environmental parameters indicating the environment of the room, such as the humidity and brightness of the room. As the imaging device 481, for example, a camera that captures an image showing the surface temperature distribution of the user is employed. Air conditioners 41 , 42 and 43 each have a compressor (not shown) and a blower fan (not shown) that operate based on control signals input from control unit 400 .

The control unit 400 includes a CPU (Central Processing Unit) 401, a main storage unit 402, an auxiliary storage unit 403, a communication interface 405, a measuring device interface 406, an operation device interface 407, an imaging interface 408, and a neuro engine. 404 and a bus 409 interconnecting them. A main storage unit 402 is configured from a volatile memory and used as a work area for the CPU 401 . In addition, the main memory unit 402 has a storage area large enough to allow multiple types of neural network operations to be executed in parallel. Auxiliary storage unit 403 is composed of a nonvolatile memory such as a magnetic disk or semiconductor flash memory, and stores programs for realizing various functions of control unit 400 . The communication interface 405 is connected to the home network NT2, and transmits various information notified from the CPU 401 to the home network NT2, and notifies the CPU 401 of various information received from the home network NT2. The operating device interface 407 communicates with the operating devices 61, 62 and 63 on a one-to-one basis. The operation device interface 407 and the operation devices 61, 62, and 63 may be wirelessly connected by a communication method conforming to a communication standard such as an infrared communication standard or Bluetooth (registered trademark). Alternatively, the operating device interface 407 and the operating devices 61, 62, 63 may be connected via a dedicated communication line. Upon receiving operation information indicating the details of the operation performed by the user on the operation devices 61 , 62 and 63 from the operation devices 61 , 62 and 63 , the operation information is notified to the CPU 401 . When the measurement value signal is input from the measurement device 461 , the measurement device interface 406 generates temperature information corresponding thereto and notifies the CPU 401 of the temperature information. When an image signal is input from the imaging device 481 , the imaging interface 408 generates image information corresponding to the image signal and notifies the CPU 401 of the generated image information.

The CPU 401 reads out the program stored in the auxiliary storage unit 403 to the main storage unit 402 and executes it, so that as shown in FIG. 414, user identification unit 415, clock unit 416, weather information acquisition unit 417, coefficient acquisition unit 418, coefficient setting unit 419, coefficient determination unit 420, coefficient transmission unit 421, device setting update unit 422, operation mode setting unit 423, It functions as a coefficient information generation unit 424 , a history information generation unit 425 , a history information transmission unit 426 and a device identification information transmission unit 427 . 3, the auxiliary storage unit 403 shown in FIG. 2 includes a device setting storage unit 431, a user information storage unit 432, a neural network storage unit 433, a history information storage unit 434, and an operation mode storage unit 431. It has a storage unit 435 and a weather information storage unit 436 .

The device setting storage unit 431 stores device setting information indicating device setting parameters of the air conditioners 41 , 42 and 43 . The user information storage unit 432 stores user information about users of the air conditioners 41 , 42 and 43 . Specifically, the user information storage unit 432 stores, for example, the heat dissipation on the surface of the user calculated from the temperature distribution on the surface of the user indicated by the image information of each user imaged by the imaging device 481. The feature amount information indicating the positions of the large number of areas is stored in association with the user identification information for identifying the user.

The history information storage unit 434 stores the history of device setting information of the air conditioners 41, 42, and 43 and environment information indicating environment parameters including temperature information. For example, as shown in FIG. 4, the history information storage unit 434 stores operation history information indicating the history of the set temperatures and air volume levels of the air conditioners 41, 42, and 43, and environmental history information indicating the history of the room temperature. , is stored in association with date and time information. Returning to FIG. 3, operation mode storage unit 435 stores operation mode information indicating operation modes of air conditioners 41 , 42 , and 43 . The weather information storage unit 436 stores weather forecast information and actual weather information obtained from the cloud server 2 .

The neural network storage unit 433 stores information regarding the neural network implemented in the neural engine 404 . The neural network is a first neural network for obtaining future equipment setting parameters of the air conditioners 41 , 42 , 43 . This neural network has an input layer L10, a hidden layer L20 and an output layer L30 as shown in FIG. The input layer L10 inputs to the hidden layer L20 environmental parameters such as room temperature and hot water temperature, numerical values indicating the date and time, and information obtained by digitizing weather conditions. Here, as a method of quantifying the weather conditions, for example, if there are four types of weather conditions, "sunny", "cloudy", "rainy", and "snowy", numerical values NUM1, NUM2, and NUM3 corresponding to each weather condition are used. , NUM4 may be set so that the relationship NUM1<NUM2<NUM3<NUM4 is established. Specifically, numerical values corresponding to "sunny", "cloudy", "rainy" and "snowy" may be set to "10", "20", "30" and "40", respectively.

ここで、Ｗ［ｊ，ｉ，ｋ］は、重み係数を示し、ｆ（＊）は、活性化関数を示す。この重み係数Ｗ［ｊ，ｉ，ｋ］が、前述のニューラルネットワークの構造を決める第１ニューラルネットワーク係数に相当する。また、活性化関数としては、シグモイド関数、ランプ関数、ステップ関数、ソフトマックス関数等の非線形関数が用いられる。例えば、活性化関数が、シグモイド関数である場合、下記式（２）の関係式で表される。

ここで、ｙｉは、引数を示し、ｙｏは出力値を示す。
また、活性化関数がランプ関数である場合、下記式（３）の関係式で表される。

ここで、ｙｉは、引数を示し、ｙｏは出力値を示す。The hidden layer L20 includes N (N is a positive integer ) layers. That is, the hidden layer L20 has a structure in which each node string is connected. Here, the output y[j, i] of each node x[j, i] is represented by the following relational expression (1).

where W[j, i, k] denotes a weighting factor and f(*) denotes an activation function. This weighting coefficient W[j, i, k] corresponds to the first neural network coefficients that determine the structure of the aforementioned neural network. Non-linear functions such as a sigmoid function, a ramp function, a step function, and a softmax function are used as activation functions. For example, when the activation function is a sigmoid function, it is represented by the following relational expression (2).

where yi indicates an argument and yo indicates an output value.
Also, when the activation function is a ramp function, it is represented by the following relational expression (3).

where yi indicates an argument and yo indicates an output value.

In the hidden layer L20, the information input to the node is the sum of the output of each node in the previous layer multiplied by a weighting factor. The output of the activation function with the sum as an argument is then propagated to the next layer. The output layer L30 directly outputs the output y[j, i] from the final layer of the hidden layer L20.

Then, the neural network storage unit 433 stores information indicating the structure of the neural network represented by Equation (1) and information indicating the weighting factors W[j, i, k] of the neural network.

Returning to FIG. 3 , the environment information acquisition unit 411 acquires temperature information indicating the room temperature measured by the measurement device 461 from the measurement device interface 406 . If the measuring device 461 measures other room environmental parameters such as room humidity and brightness, the environment information acquisition unit 411 acquires environment information indicating these other environment parameters. The image acquisition unit 412 acquires image information of the user imaged by the imaging device 481 from the imaging interface 408 .

When operation device interface 407 notifies operation information transmitted from operation device 6, operation reception unit 413 receives the notified operation information. Then, when the operation information relates to updating the device setting parameters of the air conditioner 4, the operation reception unit 413 generates the device setting information indicating the device setting parameters corresponding to the operation information, be memorized. Further, when the operation information relates to changing the operation mode of the air conditioner 4, the operation reception unit 413 notifies the operation mode setting unit 423 of operation mode information indicating the operation mode corresponding to the operation information. The device control section 414 controls the operation of the compressor and the blower fan based on the device setting information stored in the device setting storage section 431 .

The user identification unit 415 identifies an area where heat is largely emitted from the surface of the user from the temperature distribution of the surface of the user indicated by the image information acquired by the image acquisition unit 412, and the user information storage unit 432 stores the area. The users of the air conditioners 41, 42, 43 are identified based on the feature amount information about the users and the positions of the identified regions. Further, the user identification unit 415 causes the user information storage unit 432 to store the user identification information of the users of the identified air conditioners 41 , 42 and 43 .

The clock unit 416 clocks the date and time when the environment information acquisition unit 411 acquires the environment information and the current time. The weather information acquisition unit 417 acquires weather information including actual weather information indicating current and past weather conditions and weather forecast information indicating future weather conditions from the cloud server 2 . The coefficient acquisition unit 418 is a second coefficient acquisition unit that acquires, from the cloud server 2 , coefficient information including information indicating weighting coefficients of the neural network implemented in the neuro engine 404 . Further, the coefficient acquisition unit 418 executes information decompression processing on coefficient information and coefficient attribute information that have been subjected to reversible information compression processing and are acquired from the cloud server 2 . Then, the coefficient obtaining unit 418 causes the neural network storage unit 433 to store the weighting coefficient information included in the coefficient information.

The coefficient determination unit 420 determines the weighting coefficients of the neural network based on the operation history information and environment history information included in the history information and the actual weather information. The coefficient determining unit 420 first acquires information indicating initial coefficients from the neural network storage unit 433, and sets the acquired initial coefficients as weight coefficients of the neural network. Next, the coefficient determination unit 420 determines whether the neuro engine 404 quantifies the current and past environmental parameters indicated by the environmental history information, the date and time indicated by the date and time information, and the current or past weather conditions indicated by the actual weather information. and obtain the device setting parameters calculated using the neural network. Subsequently, the coefficient determination unit 420 acquires the current and past device setting parameters indicated by the operation history information stored in the history information storage unit 434, and calculates the error from the device setting parameters calculated using the neural network. . Then, the coefficient determining unit 420 determines weighting coefficients of the neural network by error backpropagation based on the calculated error. Here, the coefficient determining unit 420 determines the weighting coefficients using, for example, an autoencoder.

ここで、Ｂｅｒｎｏｕｌｌｉ（＊）は、ベルヌ－イ分布に従った確率で「１」となる関数を表す。Also, the coefficient determining unit 420 uses the dropout information when determining the weighting coefficients of the neural network. The dropout information is defined for each node node in the hidden layer L20 described above, and deactivates the node when the coefficient determination unit 215 determines the weight coefficient of the neural network. This is information indicating whether or not the output is set to "0". Each node is activated with a preset probability P and deactivated with a probability (1−P). The probability P is set for each node and ranges from 0 to 1 or less. If the probability P is set to '1', the corresponding node will always be activated. For example, if the variable rY is “1” with probability P and “0” with probability (1−P), the output y[j,i] of each node x[j,i] is given by the following equation (4 ) and equation (5).

Here, Bernoulli(*) represents a function with a probability of "1" according to the Bernoulli distribution.

The coefficient setting unit 419 sets the weighting coefficient determined by the coefficient determining unit 420 as the weighting coefficient of the neural network. The device setting update unit 422 updates the device setting information stored in the device setting storage unit 431 with the device setting parameters calculated by the neuro engine 404 .

When operation receiving unit 413 receives an operation for setting the operation mode of air conditioners 41, 42, and 43, operation mode setting unit 423 sets operation mode information indicating an operation mode corresponding to the received operation content to the operation mode. Stored in the storage unit 435 . When the coefficient information generating unit 424 receives the destination notification information described later from the management device 7, the coefficient information including the weighting coefficient information indicating the weighting coefficient of the neural network stored in the neural network storage unit 433 and the coefficient attribute corresponding thereto. Generate information and The coefficient information generation unit 424 generates coefficient attribute information whose file format is, for example, a JSON schema file format, and also generates coefficient information whose file format is a JSON file format. The coefficient transmission unit 421 is a second coefficient transmission unit that transmits coefficient information including information indicating weighting coefficients stored in the neural network storage unit 433 and coefficient attribute information corresponding thereto to the cloud server 2 . The coefficient transmission unit 421 transmits the coefficient information and the coefficient attribute information after performing reversible information compression processing.

When the preset upload time comes, the history information generation unit 425 refers to the history information stored in the history information storage unit 434 to obtain the device setting parameters and History information indicating the history of environmental parameters and history attribute information corresponding to this are generated. The history information generation unit 425 generates history attribute information whose file format is, for example, a JSON schema file format, and also generates attribute information whose file format is a JSON file format. The history information transmission unit 426 transmits the history information and history attribute information generated by the history information generation unit 425 to the customer server 9 . The history information transmission unit 426 transmits the history information and the history attribute information after performing reversible information compression processing. The device identification information transmission unit 427 transmits device identification information for identifying the air conditioners 41 , 42 and 43 to the management device 7 .

The neuro engine 404 is hardware dedicated to arithmetic processing using a neural network having a preset number of nodes and layers. The neuro engine 404 has a processor 441, a work memory 442, an arithmetic accelerator 443, an input/output register 444, and a download buffer 445, as shown in FIG. Here, coefficient attribute information and coefficient information, which will be described later, are acquired from the cloud server 2 . Note that the coefficient attribute information has, for example, a JSON schema file format, and the coefficient information has a JSON file format. Here, the coefficient attribute information is stored once in the download buffer 445 and then stored in the work memory 442 used by the processor 441 . The processor 441 reads the coefficient attribute information DAZ2 from the work memory 442, and based on the information indicating the structure of the neural network and the information indicating the number of layers and nodes of the neural network included in the coefficient attribute information DAZ2, weighting factor information DAC2, A memory area necessary for storing the node calculation value information DAN21 and the input/output node value information DAN22 is secured. Then, the processor 441 associates the weight coefficients of the neural network with the nodes in each memory area.

Processor 441 also stores weighting factor information DAC2 in the corresponding portion of work memory 442 . Also, the processor 441 stores the input value information to the neural network input from the input/output register 444 in the memory area for storing the input/output node value information DAN22, and then sequentially reads the weighting factor information DAC2. Also, the processor 441 sets the activation function information included in the coefficient attribute information DAZ2 stored in the work memory 442 in the arithmetic program, and then sequentially executes arithmetic operations for each layer and each node of the neural network. Then, when the processor 441 completes the operation for each layer and each node of the neural network, it stores the obtained output value information in the memory area storing the input/output node value information DAN22, and then stores the input/output node value information DAN22. Transfer from the memory area storing the DAN 22 to the output portion of the input/output register 444 . Arithmetic processing by the processor 441 requires a large-capacity work memory 442 and frequently transfers numerical information between the processor 441 and the work memory 442 . Therefore, computation of the neural network using the processor 441 requires a certain amount of time. Therefore, in some cases, a GPU (Graphics Processing Unit) capable of high-speed calculation is adopted as the processor 441 in order to shorten the calculation time of the neural network.

On the other hand, the arithmetic accelerator 443 is a dedicated hardware-configured accelerator that specializes in processing specific to neural network arithmetic that executes a huge number of simple arithmetic operations required for each node of the neural network. be. The operation accelerator 443 has a plurality of node unit operation units 443a. Each node unit calculation unit 443a is provided for each node (eg, nodes X1 and Y1) of the neural network, and has a local register 443b, a sum-of-products calculation unit 443c, and a conversion table unit 443d. The number of node unit operation units 443a is the same as the number of nodes constituting the neural network. In addition, the local registers 443b corresponding to the sum-of-products operation unit 443c and the conversion table unit 443d are structured so that the required number of local registers 443b can be selected, considering that the number of registers varies depending on the scale of the neural network. there is Then, the operation accelerator 443 selects a required number of local registers 443b based on information indicating the number of layers and the number of nodes of the neural network included in the coefficient attribute information acquired from the cloud server 2. FIG.

Then, after the required number of local registers 443b are selected, the operation accelerator 443 stores the weighting coefficient information in each local register 443b and executes the operation of each node of the neural network. The conversion table section 443d is used to calculate the above-described activation function, and the content of the conversion table section 443d is set based on the information indicating the shape of the activation function included in the coefficient attribute information. be. Also, the coefficient attribute information includes structural information indicating the structure of the neural network, as will be described later. Then, the node unit calculation unit 443a refers to the information about the structure of the neural network included in the coefficient attribute information, and the position of the local register 443b that stores the weight coefficient information of the neural network and the node unit calculation unit 443a. Determine the connection relationship and acquire the coefficient information. By having such a hardware configuration, the operation accelerator 443 can perform calculation separately for each node of the neural network, or collectively execute calculations for a plurality of nodes. This operation accelerator 443 is capable of high-speed processing compared to operations using the work memory 442 and the processor 441 . Further, the operation accelerator 443 reads the result of the operation using the neural network from the local register 443b of the node unit operation unit 443a corresponding to the output node, and outputs it to the output part of the input/output register 444. FIG.

Note that the arithmetic accelerator 443 cannot change the hardware circuit scale regardless of the arithmetic scale. Therefore, the neuro engine 404 according to this embodiment has a configuration in which an arithmetic accelerator 443, a processor 441, and a work memory 442 are combined.

This neuro engine 404 uses a neural network in which weighting coefficients are set by the coefficient setting unit 419 to determine the air conditioners 41, 42, and 43 based on the environmental parameters indicating the current environment contained in the weather forecast information and the environmental history information. Calculate future device configuration parameters. Here, the weather forecast information is information that quantifies future weather conditions. The device setting parameters are information indicating the set temperatures of the air conditioners 41, 42, and 43, the air volume level, the set temperature of the water heater 51, and the like. Further, when the coefficient acquisition unit 418 acquires a plurality of pieces of coefficient information, the neural engine 404 acquires a plurality of pieces of coefficient information based on information indicating the number of nodes and the number of layers of the neural network included in the coefficient attribute information of each piece of pieces of coefficient information. A plurality of neural network calculators are generated that individually execute neural network calculations corresponding to respective pieces of coefficient information. Then, according to the user specified by the user specifying unit 415, the neuro engine 404 executes a neural network operation corresponding to the user to calculate device setting parameters suitable for the user.

The cloud server 2, as shown in FIG. 7, includes a CPU 201, a main storage unit 202, an auxiliary storage unit 203, a communication interface 205, and a bus 209 interconnecting them. The CPU 201 is, for example, a multicore processor. A main storage unit 202 is configured from a volatile memory and used as a work area for the CPU 201 . The auxiliary storage unit 203 is composed of a large-capacity nonvolatile memory and stores programs for realizing various functions of the cloud server 2 . The communication interface 205 is connected to the outside network NT1 and can communicate with the weather server 3 via the outside network NT1. The CPU 201 reads the program stored in the auxiliary storage unit 203 into the main storage unit 202 and executes it, thereby obtaining a history information acquisition unit 211, a weather performance acquisition unit 212, a neural network calculation unit 214, a coefficient It functions as a determination unit 215 , a coefficient information generation unit 218 , a coefficient transmission unit 219 and a coefficient acquisition unit 220 . Further, the auxiliary storage unit 203 shown in FIG. 7 has a history information storage unit 231, a weather information storage unit 232, and a neural network storage unit 233, as shown in FIG.

The neural network storage unit 233 stores information indicating initial coefficients, which are initial values of weight coefficients of the neural networks used in the air conditioners 41, 42, and 43 installed in the dwelling unit H. FIG. This initial coefficient is based on history information including operation history information and environmental history information of air conditioners in other dwelling units in which air conditioners of the same model as the air conditioners 41, 42, and 43 are installed, and actual weather information. It is the initial value of the weighting factor of the determined neural network. In addition, the neural network storage unit 233 stores information indicating the weighting coefficients of the neural networks included in the coefficient information acquired from the air conditioners 41, 42, and 43 as the weight coefficients of the air conditioners 41, 42, and 43 that have transmitted the coefficient information. It is stored in association with the device identification information.

The history information acquisition unit 211 obtains the operation history information of the air conditioners and water heaters of the same model as the air conditioners 41, 42, and 43 installed in the dwelling unit H, and the water heaters in other dwelling units. History information including environment history information is obtained from the customer server 9 . The history information acquisition unit 211 acquires history information from the customer server 9 that manages customers who have purchased air conditioners of the same model as the air conditioners 41, 42, and 43, for example. The history information acquisition unit 211 performs information decompression processing on the history information that has been subjected to the reversible information compression processing acquired from the customer server 9, and then acquires the operation history information, the environment history information, and the usage history information included in the history information. Get user information. Then, the history information acquisition unit 211 causes the history information storage unit 231 to store the operation history information, environment history information, and user information included in the acquired history information. The weather record acquisition unit 212 acquires weather record information indicating current and past weather conditions in the area where the dwelling unit of the household corresponding to the history information exists from the weather server 3 via the outside network NT1.

The neural network calculation unit 214 uses the neural network used in the air conditioners 41, 42, and 43 to calculate the air conditioner 41 from the weather forecast information and the environmental history information acquired by the history information acquisition unit 211 from the customer server 9. , 42, and 43 are obtained.

The coefficient determining unit 420 determines the weighting coefficients of the neural network based on the operation history information and environment history information included in the history information, and the actual weather information. The coefficient determination unit 420 first acquires information indicating initial coefficients from the neural network storage unit 433, and sets the acquired initial coefficients as weight coefficients of the neural network. Next, the coefficient determining unit 420 digitizes the current and past environmental parameters indicated by the environmental history information, the date and time indicated by the date and time information, and the current or past weather conditions indicated by the actual weather information. Based on the obtained information, the device setting parameters calculated using a neural network are acquired. Subsequently, the coefficient determination unit 420 acquires the current and past device setting parameters indicated by the operation history information stored in the history information storage unit 434, and calculates the error from the device setting parameters calculated using the neural network. . Then, the coefficient determination unit 420 determines the weight coefficients of the neural network by the error backpropagation method based on the calculated error.

Coefficient information generating section 218 generates coefficient information including information indicating the weighting factor determined by coefficient determining section 213 and information indicating that the weighting factor is an initial coefficient. In addition, the coefficient information generation unit 218 generates coefficient information including information indicating the weighting coefficients of the neural networks acquired from the air conditioners 41 , 42 and 43 . Coefficient transmission section 219 is a first coefficient transmission section that transmits the coefficient information generated by coefficient information generation section 218 to air conditioners 41, 42, and 43 via outside network NT1. Upon acquiring the coefficient information transmitted from the air conditioners 41, 42, and 43, the coefficient acquisition unit 220 stores information indicating the weighting coefficients of the neural networks included in the acquired coefficient information in the neural network storage unit 233. It is a first coefficient acquisition unit that stores the device identification information of 41, 42, and 43 in association with each other.

As shown in FIG. 9, the management device 7 has a CPU 701, a main storage section 702, an auxiliary storage section 703, a display section 704, an input section 705, a network communication section 706, and a bus 709 interconnecting each section. The auxiliary storage unit 703 stores programs for realizing various functions of the information management device 7 . A display unit 704 is, for example, a liquid crystal display, and displays various information input from the CPU 701 or an operation screen image. The input unit 705 is a touch pad, and outputs information to the CPU 701 according to the content of the operation when operated by the user. The input unit 705 is placed over the display unit 704 , and the user can operate the input unit 705 while viewing the operation screen image displayed on the display unit 704 . Network communication unit 706 is connected to home network NT2 and communicates with air conditioners 41, 42, and 43 via home network NT2.

The CPU 701 reads the program stored in the auxiliary storage unit 703 into the main storage unit 702 and executes it, thereby obtaining a reception unit 711, a specification unit 712, a display control unit 713, and a device identification information acquisition unit 716 as shown in FIG. , functions as a movement information generation unit 714 and a movement information transmission unit 715 . Further, the auxiliary storage unit 703 shown in FIG. 9 has a layout information storage unit 731, a movement information storage unit 732, and an image storage unit 733, as shown in FIG. The layout information storage unit 731 stores the device identification information of the air conditioners 41, 42, and 43 in association with the arrangement information indicating the positions of the air conditioners 41, 42, and 43 within the dwelling unit H. The layout information storage unit 731 also stores floor plan information indicating the layout of the rooms R1, R2, and R3 in the dwelling unit H. FIG. The movement information storage unit 732 stores the device identification information of the air conditioners 41, 42, and 43 that are movement sources of the coefficient information and the device identification information of the air conditioners 41, 42, and 43 that are destinations of the coefficient information. , is stored in association with the user identification information. The image storage unit 733 stores image information indicating floor plan images, mark images, and arrow images to be displayed on the display unit 704 .

When the input unit 705 is operated by the user, the receiving unit 711 receives operation information corresponding to the content of the operation. The device identification information acquisition unit 716 acquires device identification information from each of the air conditioners 41 , 42 and 43 . Based on the layout information stored in the layout information storage unit 731 and the operation information received by the reception unit 711, the identification unit 712 determines device identification information of the air conditioners 41, 42, and 43 from which the coefficient information is to be moved, and , and the device identification information of the air conditioners 41, 42, and 43 to be moved to. The display control unit 713 acquires image information indicating the floor plan image and image information indicating the mark image from the image storage unit 733, and displays the floor plan image based on the layout information and the floor plan information stored in the layout information storage unit 731. Mark images indicating the air conditioners 41 , 42 and 43 are combined and displayed on the display unit 704 . Further, when the user performs an operation of dragging the coefficient information from the source air conditioners 41, 42, and 43 to the destination air conditioners 41, 42, and 43 in the input unit 705, the display control unit 713 The corresponding arrow image is acquired from the image storage unit 733 and displayed on the display unit 704 .

The movement information generation unit 714 generates movement source notification information including device identification information of the air conditioners 41, 42, and 43 that are movement sources of the coefficient information identified by the identification unit 712, and air conditioners that are movement destinations of the coefficient information. Destination notification information including device identification information of the devices 41, 42, and 43 is generated. The movement information generation unit 714 causes the movement information storage unit 732 to store the generated movement source notification information and movement destination notification information. The movement information transmission unit 715 transmits the movement source notification information to the air conditioners 41, 42, and 43 that are the movement destinations of the coefficient information, and transmits the movement destination notification information to the air conditioners 41, 42 that are the movement sources of the coefficient information. 43.

Next, the operation of the control system according to this embodiment will be described with reference to FIGS. 11 to 17B. Here, it is assumed that the air conditioner 41 is newly installed in the dwelling unit H. As shown in FIG. First, as shown in FIG. 11, history request information requesting transmission of history information to the customer server 9 is transmitted from the cloud server 2 to the customer server 9 (step S1). Here, the history information includes operation history information and environmental history information of air conditioners in other dwelling units in which air conditioners of the same model as air conditioner 41 installed in dwelling unit H are installed. On the other hand, when the customer server 9 receives the history request information, it identifies another dwelling unit in which an air conditioner of the same model as the air conditioner 41 installed in the dwelling unit H is installed, and determines whether the air conditioner installed in the specified dwelling unit History information including operation history information and environmental history information of the harmonic machine and history attribute information corresponding thereto are generated (step S2).

The history information includes protocol information, history information identification information for identifying generated history information, operation history information, and environment history information. The protocol information includes various types of information regarding the communication protocol used when sending history information to the cloud server 2 . Note that the operation history information and the environment history information are, for example, the history of the average values of the device setting parameters and the average value of the environment parameters in a plurality of households in which air conditioners of the same model as the air conditioner 41 installed in the dwelling unit H are installed. It may also indicate a history of values. Also, the history attribute information includes protocol information and various history attribute information, as shown in FIG. 12, for example. The protocol information includes various types of information regarding the communication protocol used when sending history information to the cloud server 2 . The history attribute information includes history attribute information identification information for identifying the generated history attribute information, device identification information for identifying the air conditioner 4, 52 or water heater 51, the user identification information described above, and format information. , parameter acquisition condition information, device setting type information, environment type information, linked device identification information, linked target information, and operation device identification information. The history information identification information includes at least one of, for example, identification information given to the attribute information, identification information given to the history information, and identification information of the air conditioners 4 and 52 or the water heater 51 . The format information includes information indicating the data format or file format and information indicating the compression format of each of the attribute information and the history information. Here, the format information includes, for example, information indicating that the file format of the attribute information is the JSON schema file format and information indicating that the history information is the JSON file format. The format information also includes information indicating the number of history information files and the file size of each history information file. Here, if the flag information included in the format information is, for example, "0", it indicates the number of history information files; if it is "1", it indicates the size of the first history information file; The file size of the second history information, if "N", indicates the file size of the N-th history information, and if "N+1", the compression format of the history information may be indicated.

The parameter acquisition condition information includes information indicating acquisition conditions for various parameters, such as a period during which operation history information or environment history information is acquired, a time interval for acquiring device setting parameters, and environment parameters. The parameter acquisition condition information may also include information indicating whether or not there is a change history of the acquisition conditions of various parameters, and the date and time of change when the acquisition conditions of various parameters are changed. For example, if the flag information included in the parameter acquisition condition information is "0", it indicates the acquisition date and time of the parameter; if it is "1", it indicates the acquisition start date and time of the parameter; The acquisition end date and time may be indicated, and if "3", the parameter acquisition interval may be indicated. The device setting type information is information that supplements the content of the operation history information, and indicates the type of device setting parameters such as on/off, set temperature, set air volume, and set air direction for each of the air conditioners 4 and 52 and water heater 51. including. Here, if the flag information included in the device setting type information is, for example, "0", it indicates ON/OFF of each of the air conditioners 4 and 52 and the hot water heater 51, and if it is "1", it indicates the set temperature, and if it is "1", it indicates the set temperature. ' indicates the set air volume, and "3" indicates the set air direction. The operating device identification information indicates whether the operating device 6, 71, or 72 for which the device setting parameter is set is a remote controller in the dwelling unit H, a portable terminal such as a TV or a smart phone, or a remote control terminal via the cloud server 2. contains information indicating For example, the operating device identification information may be set to "0" for a remote controller, "1" for a portable terminal, and "2" for a remote information terminal. .

The environment type information is information that supplements the contents of the environmental history information, and includes the room temperature, the temperature outside the dwelling unit H, whether or not a person inside the dwelling unit H is detected, the surface temperature of the person living inside the dwelling unit H, and the detection by the odor sensor. Contains information indicating the type of environmental parameter, such as state, _CO2 concentration, concentration of particulates (eg PM2.5) in the air. Here, if the flag information included in the environment type information is "0", it indicates the room temperature, "1" indicates the humidity, "2" indicates the outside temperature, and "3" indicates the temperature. Alternatively, the presence or absence of human detection may be indicated. Also, the environment type information includes weather information. The linked device identification information includes, for example, identification information of devices operating in cooperation with the air conditioners 4 and 52 or the water heater 51 . The cooperation target information includes, for example, the identification information of the operating state of the target equipment to be cooperated with the air conditioners 4 and 52 or the water heater 51 . The linked device identification includes, for example, identification information of the ventilation fan that is linked with the water heater 51 . In this case, the cooperation target information includes, for example, information indicating that the operation of the ventilation fan that cooperates with the water heater 51 is an on/off operation.

Returning to FIG. 11, next, the generated history information and history attribute information are transmitted from the customer server 9 to the cloud server 2 (step S3). Subsequently, the cloud server 2 transmits the weather record request information requesting the weather server 3 to transmit the weather record information to the weather server 3 (step S4). On the other hand, when receiving the weather performance request information, the weather server 3 generates weather performance information in the area where the dwelling unit H is present (step S5). Here, the actual weather information is the actual weather information indicating the current and past weather conditions in the area where the dwelling unit of the household corresponding to the history information described above exists. After that, the generated weather record information is transmitted from the weather server 3 to the cloud server 2 (step S6). On the other hand, when receiving the actual weather information, the cloud server 2 stores the received actual weather information in the weather information storage unit 232 . After that, the cloud server 2 determines weighting coefficients of the neural network as initial coefficients based on the acquired operation history information, environmental history information, and weather performance information (step S7). The cloud server 2 causes the neural network storage unit 233 to store information indicating the determined initial coefficients.

Next, assume that a new air conditioner 41 is installed in the dwelling unit H and started. At this time, device identification information for identifying the new air conditioner 41 is transmitted from the air conditioner 41 to the management device 7 (step S8). On the other hand, when receiving the device identification information from the air conditioner 41, the management device 7 stores the received device identification information in the layout information storage unit 731 in association with the layout information indicating the position of the air conditioner 41 (step S9). Here, the management device 7 causes the display unit 704 to display an operation screen image for the user to register the position of the air conditioner 41 in the management device 7, and the user inputs the When an operation for registering the position of the air conditioner 41 is performed, arrangement information corresponding to the content of the operation is generated.

Subsequently, coefficient request information requesting transmission of initial coefficients from the cloud server 2 is transmitted from the air conditioner 41 to the cloud server 2 (step S10). On the other hand, when receiving the coefficient request information, the cloud server 2 generates coefficient information including information indicating initial coefficients stored in the neural network storage unit 233 and coefficient attribute information corresponding thereto (step S11). The coefficient information includes protocol information, coefficient information identification information for identifying generated coefficient information, and weighting coefficient information. The protocol information includes various information related to the communication protocol when transmitting the coefficient information to the air conditioners 41 , 42 and 43 . The coefficient attribute information includes protocol information and various attribute information, as shown in FIG. 13, for example. The details of this coefficient attribute information will be described later. Returning to FIG. 11, next, the generated coefficient information and coefficient attribute information are transmitted from the cloud server 2 to the air conditioner 41 (step S12). On the other hand, when receiving the coefficient information, the air conditioner 41 stores the received coefficient information in the neural network storage unit 433 .

After that, assume that the air conditioner 41 determines that it is time to update the preset weighting coefficients of the neural network. In this case, after the weather performance request information requesting the weather server 3 to transmit the weather performance information is transmitted from the air conditioner 41 to the cloud server 2 (step S13), the cloud server 2 transmits to the weather server 3 (step S14). On the other hand, when receiving the weather performance request information, the weather server 3 generates weather performance information in the area where the dwelling unit H is present (step S15). Next, as shown in FIG. 14, the generated weather record information is transmitted from the weather server 3 to the cloud server 2 (step S16), and then transmitted from the cloud server 2 to the air conditioner 41 (step S17). ). On the other hand, when receiving the actual weather information, the air conditioner 41 stores the received actual weather information in the weather information storage unit 436 . Thereafter, the air conditioner 41 performs a neural A network weighting factor is determined (step S18). Here, the air conditioner 41 causes the neural network storage unit 433 to store information indicating the determined weighting factor. Thereafter, the series of processes from steps S11 to S16 described above are repeatedly executed each time the time to update the weighting coefficients of the neural network arrives.

Next, assume that the user performs an operation to switch the operation mode of the air conditioner 41 from the manual mode to the automatic mode. In this case, the air conditioner 41 accepts the switching operation to the automatic mode by the user (step S19). At this time, the air conditioner 41 updates the operation mode information stored in the operation mode storage unit 435 to the operation mode information indicating the automatic mode. After that, assume that the air conditioner 41 determines that it is time to update the device setting information stored in the device setting storage unit 431 . In this case, after the weather information request information requesting the weather server 3 to transmit the weather information including the weather forecast information and the actual weather information is transmitted from the air conditioner 41 to the cloud server 2 (step S20), It is transmitted from the cloud server 2 to the weather server 3 (step S21). On the other hand, when the weather server 3 receives the weather information request information, it specifies the weather forecast information and the actual weather information in the area where the dwelling unit H exists, and sends the weather information including the specified weather forecast information and the actual weather information. Generate (step S22). Next, the generated weather information is transmitted from the weather server 3 to the cloud server 2 (step S23), and then transmitted from the cloud server 2 to the air conditioner 41 (step S24).

Subsequently, the air conditioner 41 uses a neural network to calculate future device setting parameters for the air conditioner 41 from the weather forecast information and environmental parameters indicating the current environment (step S25). After that, the air conditioner 41 uses the calculated device setting parameters to update the device setting information stored in the device setting storage unit 431 (step S26). Thereafter, the series of processes from steps S20 to S26 described above are repeatedly executed each time the time to update the device setting information arrives.

Assume that the upload time for uploading the coefficient information indicating the weighting coefficient of the neural network used in the air conditioner 41 to the cloud server 2 has come. The upload timing is set in advance so as to come at intervals of, for example, one month. In this case, the air conditioner 41 generates history information, history attribute information corresponding thereto, coefficient information, and coefficient attribute information corresponding thereto (step S27). Here, the air conditioner 41 refers to the history attribute information stored in the history information storage unit 434, and the history information indicating the history of the device setting parameters and environment parameters from the time of the previous upload to the time of the current upload. and history attribute information corresponding thereto. In addition, the air conditioner 41 acquires information indicating the weight coefficient of the neural network stored in the neural network storage unit 433, and generates coefficient information and coefficient attribute information corresponding thereto. Here, the structures of the history information and the history attribute information are the same as the structures of the history information and the history attribute information described with reference to FIG. 12 above. Further, the coefficient information includes protocol information, coefficient information identification information for identifying the generated coefficient information, and weighting coefficient information, as described above.

The coefficient attribute information includes protocol information and various attribute information, as shown in FIG. The attribute information includes coefficient attribute information identification information for identifying the generated coefficient attribute information, device identification information for identifying the air conditioners 41, 42, and 43 for which device setting parameters are to be calculated using the neural network, and The aforementioned user identification information, format information, neural network structure information, calculation information, learning method information, learning period information, coefficient update time information, realization function information, and device usage environment information are included. The coefficient information identification information includes, for example, at least one of identification information added to the attribute information, identification information added to the weighting coefficient information, and identification information of the air conditioners 41 , 42 and 43 . The format information includes information indicating the data format or file format and information indicating the compression format of each of the attribute information and the weighting factor information. Here, the format information includes, for example, information indicating that the file format of the attribute information is the JSON schema file format and information indicating that the weighting factor information is the JSON file format. Neural network structure information includes information indicating the number of neural network layers and the number of nodes in each layer, information indicating the order of the matrix used for computation using the neural network, and the shape of the activation function at each node of the neural network. information representing; In addition, the neural network structure information includes normalization processing or dropout information in calculation using the neural network, information about nodes connected to the input side and nodes connected to the output side of each node of the neural network, including. Here, the dropout information is information indicating whether or not any of the nodes forming the neural network is deactivated when determining the weighting coefficients of the neural network. The calculation information includes information indicating a processing method when performing calculations using a neural network such as multithread processing and pipeline processing. The learning method information includes information indicating a learning method such as error backpropagation using an autoencoder. The learning period information includes information indicating the current or past time when the operation history information, environment history information, and weather performance information used to determine the coefficients of the neural network were obtained. The coefficient update timing information includes information indicating when to update the weighting coefficients of the neural network. The realized function information includes information indicating the functions of the air conditioners 41, 42, 43 to be controlled by the device setting parameters calculated using the neural network. Further, the realization function information includes information indicating the details of operations to be performed on the operation devices 61, 62, and 63 when the user manually changes the device setting parameters calculated using the neural network. The equipment usage environment information includes information indicating the arrangement of each of the air conditioners 41, 42, and 43 in the dwelling unit H, and information indicating the composition of households living in the dwelling unit H.

Next, the generated history information, history attribute information, coefficient information, and coefficient attribute information are transmitted from the air conditioner 41 to the cloud server 2 (step S28). On the other hand, when the cloud server 2 receives history information, history attribute information, coefficient information, and coefficient attribute information from the air conditioner 41, it associates the received history information and history attribute information with the device identification information of the air conditioner 41. The received coefficient information and coefficient attribute information are stored in the neural network storage unit 233 in association with the device identification information of the air conditioner 41 (step S29).

Here, the cloud server 2 stores information specific to the air conditioner 41, such as device identification information and device setting type information included in the history information, or information related to the usage environment of the air conditioner 41, such as environment type information. information and so-called dynamic information such as device setting parameters and environmental parameters are stored in the history information storage unit 231 as IoT information. Further, the cloud server 2 associates the information indicating the weighting coefficient of the neural network included in the coefficient information with the learning period information included in the coefficient information, and stores it in the neural network storage unit 233 as AI information. During the period from the previous upload time to the current upload time, the cloud server 2 refers to the learning period information included in the coefficient information, and uses history information used to determine the weighting coefficients of the neural network as IoT information. The information is stored in the history information storage unit 231 in association with the coefficient identification information included in the information. The history information also includes information indicating the time when a failure or malfunction occurred in the air conditioner 41 . As a result, the cloud server 2 can manage the status of occurrence of problems in the air conditioner 41 .

Subsequently, the history information and the history attribute information received by the cloud server 2 from the air conditioner 41 are transmitted from the cloud server 2 to the customer server 9 (step S30). Here, when the customer server 9 receives the history information and history attribute information from the air conditioner 41, the received history information and history attribute information are associated with the device identification information of the air conditioner 41 and stored in the storage. .

Further, as shown in FIG. 15, the user, via the input unit 705 of the management device 7, updates the coefficient information of the neural networks used in the air conditioners 42 and 43 already set in the dwelling unit H. Assume that a coefficient information transfer operation is performed to transfer to the air conditioner 41 . In this case, the management device 7 receives the coefficient information transfer operation by the user (step S31).

Here, based on the layout information and floor plan information stored in the layout information storage unit 731, the management device 7 displays an operation screen image including a floor plan image GAH indicating the floor plan of the dwelling unit H as shown in FIG. 704 is displayed on the display screen 704a. Here, the image GAH includes areas A1, A2 and A3 corresponding to rooms R1, R2 and R3 in which the air conditioners 41, 42 and 43 are installed.

While the management device 7 displays the operation screen image on the display screen 704a of the display unit 704, the user points a portion of the input unit 705, which is a touch pad, corresponding to the mark image GA2 indicating the air conditioner 42. Suppose you touched with . After that, assume that the user drags the finger F that touches the input unit 705 to a portion of the input unit 705 corresponding to the mark image GA1 indicating the air conditioner 41 . At this time, the management device 7 causes the display screen 704a of the display unit 704 to display the arrow image GAAR1. Then, when the user removes the finger F from the input unit 705, the management device 7 causes the display screen 704a of the display unit 704 to display a window image PM1 as shown in FIG. 16B. The window image PM1 includes a button image BU11 that is displayed in a portion that is touched when ending the coefficient information moving operation. Window image PM1 also includes button image BU12 displayed in the portion touched when continuing the coefficient information moving operation after the portion corresponding to button image BU11 in input unit 705 is touched. A portion of the input unit 705 corresponding to the button image BU12 cannot be selected while the coefficient moving operation is being performed.

Further, if the user does not touch the portion corresponding to the button image BU11 in the input unit 705 and subsequently touches the portion corresponding to the mark image GA3 indicating the air conditioner 43 in the input unit 705 with the finger F. do. Assume that the user then drags the finger F touching the input unit 705 to a portion of the input unit 705 corresponding to the mark image GA1 indicating the air conditioner 41 . At this time, the management device 7 additionally displays an arrow image GAAR2 on the display screen 704a of the display unit 704, as shown in FIG. 17A.

Next, assume that the user touches a portion of the input unit 705 corresponding to the button image BU11 with a finger F, as shown in FIG. 17B. In this case, as shown in FIG. 15, the management device 7 identifies the air conditioners 42 and 43 that are the transfer sources of the coefficient information and the air conditioner 41 that is the transfer destination (step S32). Subsequently, the management device 7 sends destination notification information including device identification information of the destination air conditioner 41 of the specified coefficient information, and transfer destination notification information including device identification information of the specified source air conditioners 42 and 43 . The original notification information is generated (step S33). Thereafter, the generated movement source notification information is transmitted from the management device 7 to the air conditioner 41 (step S34), and the generated movement destination notification information is transmitted from the management device 7 to the air conditioners 42 and 43. (Step S35).

On the other hand, when the air conditioners 42 and 43 receive the destination notification information from the management device 7, the air conditioners 42 and 43 acquire the information indicating the weight coefficients of the neural networks used by themselves from the neural network storage unit 433, and store the coefficient information and the coefficient attribute information. and are generated (step S36). The structures of this coefficient information and coefficient attribute information are the same as the structures of the coefficient information and coefficient attribute information described above with reference to FIG. Next, the generated coefficient information and coefficient attribute information are transmitted from the air conditioners 42 and 43 to the cloud server 2 (step S37). On the other hand, when the cloud server 2 receives the coefficient information and the coefficient attribute information from the air conditioners 42 and 43, the cloud server 2 associates the received coefficient information and coefficient attribute information with the device identification information of the air conditioners 42 and 43 to create a neural network. It is stored in the network storage unit 233 (step S38).

Further, when the air conditioner 41 receives the movement source notification information from the management device 7, the coefficient request information including the device identification information of the air conditioners 42 and 43 included in the received movement source notification information is sent from the air conditioner 41. It is transmitted to the cloud server 2 (step S39). On the other hand, when receiving the coefficient request information, the cloud server 2 selects the coefficient information stored in the neural network storage unit 233 based on the device identification information included in the received coefficient request information. The corresponding coefficient information is identified (step S40). At this time, the cloud server 2 stores the learning period information included in the coefficient information corresponding to each of the air conditioners 42 and 43, and the information indicating the failure occurrence time included in the history information corresponding to each of the air conditioners 42 and 43. , are compared, and the coefficient information in which the failure occurrence time exists within the learning period indicated by the learning period information is excluded. As a result, coefficient information suitable for the air conditioner 41 is transmitted to the air conditioner 41, so that the air conditioner 41 can be automatically operated appropriately.

Subsequently, the specified coefficient information and the coefficient attribute information corresponding thereto are transmitted from the cloud server 2 to the air conditioner 41 (step S41). On the other hand, when receiving the coefficient information and coefficient attribute information from the cloud server 2, the air conditioner 41 stores the received coefficient information and coefficient attribute information in the neural network storage unit 433 (step S42).

Further, as shown in FIG. 18, the user can upload coefficient information including information indicating weight coefficients stored in the neural network storage unit 433 of the air conditioner 41 to the cloud server 2 via the operation device 61. Suppose you perform an upload operation. In this case, the air conditioner 41 receives an operation to upload coefficient information to the cloud server 2 (step S51), and generates coefficient information including weight coefficient information indicating the weight coefficient stored in the neural network storage unit 433 ( step S52). After that, the generated coefficient information and coefficient attribute information are transmitted from the air conditioner 41 to the cloud server 2 (step S53). On the other hand, when the cloud server 2 receives the coefficient information and the coefficient attribute information, the received coefficient information and coefficient attribute information are stored in the neural network storage unit 233 in association with the device identification information that identifies the air conditioner 41. (step S54).

It is also assumed that the user has performed a download operation for downloading the coefficient information from the cloud server 2 to the air conditioner 41 via the operation device 61 . In this case, the air conditioner 41 accepts an operation of downloading coefficient information from the cloud server 2 (step S55). Next, coefficient request information requesting the cloud server 2 to transmit coefficient information is transmitted from the air conditioner 41 to the cloud server 2 (step S56). This coefficient request information includes device identification information of the air conditioner 41 . On the other hand, when receiving the coefficient request information, the cloud server 2 identifies the coefficient information associated with the device identification information included in the received coefficient request information (step S57). Subsequently, the specified coefficient information and the coefficient attribute information corresponding thereto are transmitted from the cloud server 2 to the air conditioner 41 (step S58). On the other hand, when receiving the coefficient information and coefficient attribute information, the air conditioner 41 stores the received coefficient information and coefficient attribute information in the neural network storage unit 433 (step S59).

Next, device control processing executed by the air conditioners 41, 42, and 43 according to the present embodiment will be described with reference to FIGS. 19 to 22. FIG. This device control process is started, for example, when the air conditioners 41, 42, and 43 are newly installed in the dwelling unit H and then the air conditioners 41, 42, and 43 are started. First, as shown in FIG. 19, the device identification information transmitting unit 427 transmits device identification information for identifying the air conditioners 41, 42, and 43 to the management device 7 (step S101). Next, by transmitting coefficient request information to the cloud server 2 (step S102), the coefficient acquisition unit 418 obtains coefficient information including initial weighting coefficient information indicating initial coefficients of the neural network and coefficient attribute information from the cloud server 2. is obtained (step S103). The coefficient acquisition unit 418 causes the neural network storage unit 433 to store initial weighting coefficient information and coefficient attribute information included in the acquired coefficient information.

Next, the coefficient determination unit 420 determines whether or not it is time to update the coefficients of the neural network (step S104). When the coefficient determination unit 420 determines that the coefficient update timing has not yet arrived (step S104: No), the process of step S108, which will be described later, is executed as it is. On the other hand, it is assumed that the coefficient determination unit 420 determines that the coefficient update time has come (step S104: Yes). In this case, the weather information acquisition unit 417 acquires the actual weather information from the cloud server 2 by transmitting the actual weather request information to the cloud server 2 (step S105). The weather information acquisition unit 417 stores the acquired weather performance information in the weather information storage unit 436 . After that, a coefficient determination process is executed (step S107).

Details of the coefficient determination process will now be described in detail with reference to FIG. First, the neuro engine 404 acquires operation history information, environment history information, and date and time information of the air conditioner 41 from the history information storage unit 434, and acquires actual weather information from the weather information storage unit 436 (step S201). The operation history information, environment history information, and date/time information correspond to so-called teacher data for learning the neural network. Next, the coefficient setting unit 419 acquires information indicating the initial weighting coefficient, which is the initial value of the weighting coefficient, from the neural network storage unit 433, and sets the initial weighting coefficient indicated by the acquired information to the neural network for computing the initial weighting coefficient by the neural engine 404. A network weighting factor is set (step S202). Subsequently, the neuro engine 404 uses a neural network in which the initial weighting coefficients are set to digitize the environmental parameters included in the acquired environmental history information, the date and time indicated by the date and time information, and the weather conditions indicated by the actual weather information. , the device setting parameters for each of a plurality of preset time periods in the past and present are calculated (step S203). After that, the coefficient determining unit 420 calculates an error between the calculated device setting parameter and the device setting parameter included in the operation history information for each of the plurality of time periods (step S204). Next, the coefficient determining unit 420 newly determines each weighting coefficient of the neural network by error backpropagation based on the calculated error (step S205). Then, the coefficient determining unit 420 stores the determined weighting coefficients in the neural network storage unit 433 (step S206).

Returning to FIG. 19, next, the operation accepting unit 413 determines whether or not an operation for setting the operation mode has been accepted (step S108). When the operation reception unit 413 determines that the operation for setting the operation mode has not been received (step S108: No), the process of step S109, which will be described later, is executed as it is. On the other hand, when the operation accepting unit 413 determines that an operation for setting the operation mode has been accepted (step S108: Yes), the operation mode setting unit 423 generates operation mode information indicating an operation mode corresponding to the content of the accepted operation. It is stored in the operation mode storage unit 435 (step S109).

Subsequently, the device setting update unit 422 refers to the operation mode information stored in the operation mode storage unit 435 and determines whether or not the operation mode of the air conditioners 41, 42, 43 is the automatic mode (step S110). When the device setting update unit 422 determines that the operation mode of the air conditioners 41, 42, and 43 is the manual mode (step S110: No), the process of step S116, which will be described later, is executed. On the other hand, when the device setting update unit 422 determines that the operation mode of the air conditioners 41, 42, and 43 is the automatic mode (step S110: Yes), the device settings of the air conditioners 41, 42, and 43 set in advance are updated. It is determined whether or not it is time to update the information (step S111). When the device setting update unit 422 determines that the time to update the device settings of the air conditioners 41, 42, and 43 has not yet arrived (step S111: No), the process of step S116, which will be described later, is executed. On the other hand, it is assumed that the device setting update unit 422 determines that it is time to update the device setting information of the air conditioners 41, 42, and 43 (step S111: Yes). In this case, the weather information acquiring unit 417 acquires weather information from the cloud server 2 by transmitting weather information request information to the cloud server 2 (step S112) (step S113). Here, the weather information acquisition unit 417 causes the weather information storage unit 436 to store the weather forecast information and the actual weather information included in the acquired weather information.

After that, the neuro engine 404 uses the neural network in which the weight coefficients are set by the coefficient setting unit 419 based on the current environmental parameters and the weather forecast information included in the environmental history information, and the air conditioners 41, 42, 43 device setting parameters are calculated (step S114). Here, the coefficient setting unit 419 acquires weighting coefficient information indicating weighting coefficients stored in the neural network storage unit 433, and sets the weighting coefficients indicated by the acquired weighting coefficient information as weighting coefficients of the neural network. Next, the device setting update unit 422 updates the device setting information stored in the device setting storage unit 431 using the calculated device setting parameters (step S115).

Subsequently, as shown in FIG. 21, the history information generation unit 425 determines whether or not the preset upload timing has arrived (step S116). When the history information generation unit 425 determines that the upload time has not yet come (step S116: No), the process of step S118, which will be described later, is executed. On the other hand, it is assumed that the history information generation unit 425 determines that the upload time has come (step S116: Yes). In this case, the history information generation unit 425 generates history information and history attribute information corresponding thereto, and the coefficient information generation unit 424 generates coefficient information and coefficient attribute information corresponding thereto (step S117). . Here, the history information generation unit 425 refers to the history information stored in the history information storage unit 434, and the history information indicating the history of the device setting parameters and environment parameters from the time of the previous upload to the time of the current upload. to generate Further, the coefficient information generation unit 424 acquires information indicating the weighting coefficients of the neural network stored in the neural network storage unit 433 and generates coefficient information. Here, the structures of history information and history attribute information are the same as the structures of history information and history attribute information described with reference to FIG. It is the same as the structure of the coefficient information and the coefficient attribute information described using . After that, the history information transmission unit 426 transmits the history information generated by the history information generation unit 425 and the history attribute information to the customer server 9, and the coefficient transmission unit 421 transmits the coefficient information and the coefficient information generated by the coefficient information generation unit 424. coefficient attribute information to the customer server 9 (step S118).

Next, the coefficient information generating unit 424 notifies the destination of the coefficient information indicating the weighting coefficient of the neural network used by the own device from the management device 7 to the other air conditioners 41, 42, and 43. It is determined whether or not destination notification information has been received (step S119). When the coefficient information generation unit 424 determines that the destination notification information has not been received (step S119: No), the process of step S122, which will be described later, is executed. On the other hand, when the coefficient information generation unit 424 determines that destination notification information has been received (step S119: Yes), it generates coefficient information and coefficient attribute information corresponding thereto (step S120). The structures of this coefficient information and coefficient attribute information are the same as the structures of the coefficient information and coefficient attribute information described above with reference to FIG. The coefficient information includes weighting coefficient information indicating weighting coefficients stored in the neural network storage unit 433 . Further, the coefficient attribute information includes device identification information of its own device and device identification information of the air conditioner to which the coefficient information is transferred (for example, the air conditioner 41). Here, the device identification information of the air conditioner to which the coefficient information is transferred is included in the coefficient attribute information as device usage environment information, for example. Subsequently, the coefficient transmission unit 421 transmits the generated coefficient information and coefficient attribute information to the cloud server 2 (step S121).

After that, the coefficient acquisition unit 418 receives a movement source notification from the management device 7 to notify the movement source when the coefficient information indicating the weighting coefficient of the neural network used by the other air conditioners 41, 42, and 43 is moved to the air conditioner itself. It is determined whether information has been received (step S122). When the coefficient acquisition unit 418 determines that the movement source notification information has not been received (step S122: No), the process of step S125, which will be described later, is executed as it is. On the other hand, when the coefficient acquisition unit 418 determines that the movement source notification information has been received (step S122: Yes), the coefficient obtaining unit 418 requests the cloud server 2 to transmit the coefficient information of the other air conditioners 41, 42, and 43. The request information is transmitted to the cloud server 2 (step S123). This coefficient request information includes the device identification information of the air conditioner (for example, the air conditioners 42 and 43) that is the transfer source of the coefficient information included in the transfer source notification information. Thereby, the coefficient acquisition unit 418 acquires the coefficient information corresponding to the other air conditioner and the coefficient attribute information corresponding thereto from the cloud server 2 (step S124).

Next, as shown in FIG. 22, the operation accepting unit 413 determines whether or not an upload operation for uploading coefficient information to the cloud server 2 has been accepted (step S125). When the operation accepting unit 413 determines that the upload operation has not been accepted (step S125: No), the process of step S128, which will be described later, is executed. On the other hand, when the operation accepting unit 413 determines that the upload operation has been accepted (step S125: Yes), the coefficient information generating unit 424 generates coefficient information including information indicating weighting coefficients stored in the neural network storage unit 433 ( step S126). Subsequently, the coefficient transmission unit 421 transmits the generated coefficient information to the cloud server 2 (step S127). Thereafter, the operation reception unit 413 determines whether or not a download operation for downloading coefficient information from the cloud server 2 has been received (step S128). When the operation accepting unit 413 determines that the download operation has not been accepted (step S128: No), the process of step S103 is executed again. On the other hand, when the operation accepting unit 413 determines that the download operation has been accepted (step S128: Yes), the coefficient acquiring unit 418 transmits coefficient request information to the cloud server 2 (step S129), thereby obtaining coefficients from the cloud server 2. Information and coefficient attribute information corresponding thereto are obtained (step S130). The coefficient acquisition unit 418 causes the neural network storage unit 433 to store weighting coefficient information included in the acquired coefficient information. Here, when the coefficient acquisition unit 418 acquires a plurality of pieces of coefficient information and coefficient attribute information corresponding thereto, the neuro engine 404 calculates the number of nodes of the neural network included in each piece of coefficient attribute information corresponding to the plurality of pieces of coefficient information. And based on the information indicating the number of layers, neural network operations corresponding to each of the plurality of coefficient information are executed separately. After that, the process of step S103 is executed again.

Next, the coefficient information generation processing executed by the cloud server 2 according to this embodiment will be described with reference to FIG. 23 . This coefficient information generation process is started, for example, when the cloud server 2 is powered on. First, the history information acquisition unit 211 transmits history information including operation history information and environmental history information of air conditioners of the same model as the air conditioners 41, 42, and 43 installed in the dwelling unit H to the customer server 9. is sent to the customer server 9 (step S301), the history information and the history attribute information corresponding thereto are obtained from the customer server 9 (step S302). Next, the actual weather acquisition unit 212 acquires the actual weather information from the weather server 3 by transmitting the actual weather request information requesting transmission of the actual weather information to the weather server 3 (Step S303). S304). Subsequently, a coefficient determination process for determining coefficients of the neural network described above is executed based on the operation history information, environment history information, and weather performance information included in the history information (step S305). The content of the coefficient determination process is the same as the coefficient determination process described using FIG. The weighting factors calculated by this factor determination process are stored in the neural network storage unit 233 as initial weighting factors.

After that, the coefficient information generator 218 determines whether or not coefficient request information has been acquired from the air conditioners 41, 42, and 43 (step S306). When the coefficient information generator 218 determines that the coefficient request information has not been acquired (step S306: No), the process of step S301 is executed again. On the other hand, when the coefficient information generation unit 218 determines that the coefficient request information has been acquired (step S306: Yes), it generates coefficient information including information indicating initial weighting coefficients stored in the neural network storage unit 233 (step S307). . After that, the coefficient transmission unit 219 transmits the generated coefficient information to the air conditioners 41, 42, and 43 that have transmitted the coefficient request information (step S308). Next, the process of step S301 is performed again.

Here, an example of transition of information indicating weighting coefficients and history information held in each of the air conditioner 41 and the cloud server 2 in the control system according to the present embodiment will be described with reference to FIG. 24 . First, at time T0 when the air conditioner 41 is newly introduced, the cloud server 2 causes the neural network storage unit 233 to store information indicating the initial weighting coefficient to be transmitted to the air conditioner 41, and the air conditioner 41 Information indicating the initial weighting coefficients acquired from the cloud server 2 is stored in the neural network storage unit 433 . After that, it is assumed that the air conditioner 41 reaches the first renewal timing T1 due to the life of the air conditioner 41 . At this time, the air conditioner 41 causes the history information storage unit 434 to store the history information from time T0 to time T1, and the weighting factor determined based on the history information from time T0 to time T1. Information indicating (ver1) is stored in the neural network storage unit 433 . At this time, the cloud server 2 causes the history information storage unit 231 to store the history information received from the air conditioner 41 from time T0 to time T1 as so-called learning information, and also stores the information indicating the initial weighting coefficient and the air temperature. Information indicating the weighting coefficient (ver1) received from the harmonic machine 41 is stored in the neural network storage unit 233 .

After that, it is assumed that the air conditioner 41 reaches the second renewal time T2 due to the life of the air conditioner 41 . At this time, the air conditioner 41 causes the history information storage unit 434 to store the history information from time T1 to time T2, and the weighting factor determined based on the history information from time T1 to time T2. Information indicating (ver2) is stored in the neural network storage unit 433 . At this time, the cloud server 2 causes the history information storage unit 231 to store the history information received from the air conditioner 41 from time T1 to time T2 as learning information, and the weighting coefficient received from the air conditioner 41. (ver1) and information indicating the weighting factor (ver2) are stored in the neural network storage unit 233 . The cloud server 2 then deletes the initial weighting coefficients from the neural network storage unit 233 .

After that, at time T21, it is assumed that the air conditioner 41 malfunctions. Then, suppose that the third update timing T3 of the air conditioner 41 comes after the timing T21. At this time, the air conditioner 41 causes the history information storage unit 434 to store history information from time T2 to time T3. However, the air conditioner 41 stores, in the neural network storage unit 433, information indicating the weighting coefficient (ver3) newly determined based on the history information from time T1 to time T2 excluding the time when the problem occurred. I am letting At this time, the cloud server 2 leaves the history information received from the air conditioner 41 from time T1 to time T2 in the history information storage unit 231 as is. The history information stored in the history information storage unit 231 as learning information is avoided. Further, the cloud server 2 leaves the information indicating the weighting factor (ver1) in the neural network storage unit 233 as it is, and stores the information indicating the weighting factor (ver2) received from the air conditioner 41 in the neural network storage unit 233. To avoid. Then, the cloud server 2 causes the neural network storage unit 233 to store the newly determined weighting factor (ver3). This prevents history information about the time when a problem occurred in air conditioner 41 from being used to determine the weighting coefficients of the neural network used in air conditioner 41 . In addition, the cloud server 2 deletes the information indicating the weighting factor corresponding to the period before the preset period and the history information from the neural network storage unit 233 and the history information storage unit 231 . As a result, expansion of the capacity scale of the auxiliary storage 203 required for the cloud server 2 is suppressed.

As described above, in the control system according to the present embodiment, in the cloud server 2, the coefficient acquisition unit 220 acquires coefficient information from the air conditioner 41, and the coefficient transmission unit 219 transmits the coefficient information from the air conditioner 41. After obtaining the request information, the coefficient information obtained by the coefficient obtaining unit 220 is transmitted to the air conditioner 41 . On the other hand, in the air conditioner 41, the coefficient acquisition unit 418 acquires the coefficient information from the cloud server 2 by transmitting the coefficient request information to the cloud server 2, and the coefficient setting unit 419 acquires the coefficient information acquired by the coefficient acquisition unit 418. Based on the coefficient information, weight coefficients of the neural network are set. As a result, for example, before updating the air conditioner 41, the coefficient information including the information indicating the weight coefficient of the neural network used in the air conditioner 41 is uploaded to the cloud server 2, and the update of the air conditioner 41 is performed. By transmitting the coefficient request information to the cloud server 2 and downloading the coefficient information from the cloud server 2 after the update, it is possible to set the pre-update weighting coefficients for the neural network used in the air conditioner 41 after the update. . Therefore, even when the air conditioner 41 is updated, the tendency of the operation state before the update of the air conditioner 41 can be continued in the updated air conditioner 41 . That is, the operation tendency of the air conditioner 41 that has been used until then during automatic operation can be inherited by the new air conditioner 41 . In addition, in the control system according to the present embodiment, by downloading and updating the weight coefficients of the neural network of the air conditioner 41 from the cloud server 2, changes in the usage of the air conditioner 41 by the user, particularly, It is possible to flexibly cope with changes in the installation environment of the air conditioner 41 due to the moving of the user, changes in the family composition of the user, and the like.

Further, in the control system according to the present embodiment, by uploading information indicating the weight coefficients of the neural networks corresponding to a plurality of different users to the cloud server 2 in advance, coefficient information corresponding to each user can be downloaded and used. As a result, even when the number of users increases significantly, the air conditioners 41, 42, and 43 can be automatically operated with an operation tendency suitable for each user.

Furthermore, the air conditioners 41, 42, and 43 according to the present embodiment are provided with operation history information indicating the history of the device setting parameters of the air conditioners 41, 42, and 43 and the environment in which the air conditioners 41, 42, and 43 operate. A coefficient determining unit 420 for determining the weighting coefficient based on the environmental history information indicating the history of the environmental parameters, the information indicating the initial weighting coefficient, and the actual weather information. As a result, the air conditioners 41, 42, and 43 continue to operate while updating the weight coefficients of the neural networks, so that the air conditioners 41, 42, and 43 can be automatically operated with an operation tendency suitable for the user. can.

Further, when the coefficient acquisition unit 418 acquires a plurality of pieces of coefficient information, the neuro engine 404 according to the present embodiment obtains information indicating the number of nodes and the number of layers of the neural network included in the coefficient attribute information of each of the pieces of coefficient information. a plurality of neural network calculators that individually execute neural network calculations corresponding to a plurality of pieces of coefficient information, respectively. As a result, computations using a plurality of types of neural networks can be executed in parallel. Autonomous operation can be performed with an operation tendency suitable for each person.

Furthermore, in the present embodiment, the history information generation unit 425 generates operation history information indicating the history of the device setting parameters of the air conditioners 41, 42, and 43 and environment history information indicating the history of the environmental parameters of the environment in which the devices operate. , and the history information transmission unit 426 transmits the generated history information to the customer server 9 . Then, the coefficient determination unit 215 of the cloud server 2 acquires the history information and the actual weather information of the air conditioners 41, 42, and 43 and the air conditioners of the same model as the air conditioners 41, 42, and 43, which are acquired from the customer server 9. , the initial values of the weighting coefficients of the neural networks used in the air conditioners 41, 42, and 43 are determined. As a result, for example, when new air conditioners 41, 42, 43 are introduced, the initial weighting coefficients of the neural networks used in the air conditioners 41, 42, 43 can be set to appropriate values.

Although the embodiments of the present invention have been described above, the present invention is not limited to the configurations of the above-described embodiments. For example, the management device 7 may cause the display unit 704 to display an operation screen image for setting users for each of the air conditioners 41 , 42 , and 43 . When the management device 7 receives the device identification information from the newly installed air conditioner 41, the display control unit 713 of the management device 7 displays a list of devices that have not yet been registered in the management device 7, for example, as shown in FIG. 25A. is displayed on the display screen 704 a of the display unit 704 . The operation screen image shown in FIG. 25A includes a button image BU21 corresponding to the air conditioner 41 newly installed. Here, when the user touches a portion corresponding to the button image BU21 on the input unit 705, which is a touch pad, the display control unit 713 performs an operation for registering the user of the air conditioner 41 as shown in FIG. 25B, for example. The screen image is displayed on the display screen 704 a of the display unit 704 . The operation screen image shown in FIG. 25B includes a plurality of button images BU22 representing each user living in the dwelling unit H, for example. Here, when the user touches a portion corresponding to one of the button images BU22 on the input unit 705, the specifying unit 712 of the management device 7 sends the user identification information of the user indicated by the button image BU22 to the air. It is stored in the layout information storage unit 731 in association with the device identification information of the harmony machine 41 . When one user is registered in the management device 7 as a user of the air conditioner 41 in this manner, the display control unit 713 displays an operation screen image including a window image PM21 as shown in FIG. 26, for example. It is displayed on the display screen 704 a of the display unit 704 . Here, when the user touches the portion corresponding to the button image BU24 on the input unit 705, the display control unit 713 causes the display screen 704a of the display unit 704 to display the operation screen image as shown in FIG. 25B again.

On the other hand, when the user touches the portion corresponding to the button image BU23 in the input unit 705, the specifying unit 712 is associated with the device identification information of the other air conditioners 42 and 43 stored in the layout information storage unit 731. By referring to the stored user identification information, device identification information associated with the same user identification information as the user identification information associated with the air conditioner 41 is retrieved. Then, the specifying unit 712 identifies the air conditioners 42 and 43 to which the device identification information associated with the same user identification information as the user identification information associated with the air conditioner 41 is assigned with the coefficient information. They are identified as the air conditioners 42 and 43 of the movement source.

Note that, in this modified example, for example, the cloud server 2 may be configured to implement the function of the specifying unit 712 . In this case, the cloud server 2 includes a user information storage unit (not shown) that stores the user identification information of each air conditioner 42, 43 in association with the device identification information of each air conditioner 42, 43. It may be configured to include Then, the management device 7 transmits the device identification information received from the air conditioner 41 and the user identification information associated therewith to the cloud server 2, and the cloud server 2 stores the device identification information storage unit. By referring to the user identification information associated with the device identification information of the other air conditioners 42 and 43, the same user identification information as the user identification information associated with the air conditioner 41 is associated It is sufficient to search for the obtained device identification information.

According to this configuration, the management device 7 does not need to store the image information indicating the floor plan image GAH, so the storage capacity of the auxiliary storage unit 703 of the management device 7 can be reduced.

In the embodiment, an example has been described in which the user operates the input unit 705 of the management device 7 to move the coefficient information. It may specify the air conditioners 42 and 43 from which the information is transferred.

Here, the operation of the control system according to this modified example will be described with reference to FIG. In addition, in FIG. 27, the same reference numerals as in FIG. 14 are assigned to the same processing as in the embodiment. First, it is assumed that the user installs the air conditioner 41 in the dwelling unit H and then activates the air conditioner 41 . At this time, the air conditioner 41 images the user with the imaging device 481 (step S3031). Next, the air conditioner 41 acquires image information of the user imaged by the imaging device 481, and identifies the user of the air conditioner 41 from the surface temperature distribution of the user indicated by the acquired image information ( step S3032). Here, the air conditioner 41 causes the user information storage unit 432 to store the user identification information of the identified user of the air conditioner 41 . Subsequently, the device identification information of the air conditioner 41 and the user identification information stored in the user information storage unit 432 are transmitted from the air conditioner 41 to the management device 7 (step S3033).

On the other hand, when receiving the device identification information and the user identification information from the air conditioner 41, the management device 7 associates the device identification information with the device identification information of the other air conditioners 42 and 43 stored in the layout information storage unit 731. By referring to the user identification information, device identification information associated with the same user identification information as the user identification information associated with the air conditioner 41 is retrieved. Then, the specifying unit 712 specifies the air conditioner 41 to which the received device identification information is assigned as the coefficient information transfer destination, and the device identification information associated with the same user identification information as the air conditioner 41 is The assigned air conditioners 42 and 43 are specified as the transfer source of the coefficient information (step S3034). After that, the processing after step S33 is executed.

In this modification, for example, the cloud server 2 has a user information storage unit (not shown) that stores the user identification information in association with the device identification information of the air conditioners 42 and 43. There may be. In this case, as shown in FIG. 28, it is assumed that the user first installs the air conditioner 41 in the dwelling unit H and then activates the air conditioner 41 . At this time, the air conditioner 41 images the user with the imaging device 481 (step S4032). Next, the device identification information of the air conditioner 41 and the image information obtained by imaging the user are transmitted from the air conditioner 41 to the cloud server 2 (step S4033). Next, the cloud server 2 acquires the image information of the user imaged by the imaging device 481, and identifies the user of the air conditioner 41 from the surface temperature distribution of the user indicated by the acquired image information (step S4034). At this time, the cloud server 2 associates the user identification information of the identified user with the device identification information acquired from the air conditioner 41 and stores it in the user information storage unit.

Subsequently, the cloud server 2 refers to the user identification information associated with the device identification information of the other air conditioners 42 and 43 stored in the user information storage unit, and refers to the user identification information associated with the air conditioner 41. device identification information associated with the same user identification information as the user identification information received. Then, the cloud server 2 identifies the air conditioner 41 to which the received device identification information is assigned as the transfer destination of the coefficient information, and the device identification information associated with the same user identification information as the air conditioner 41 is The assigned air conditioners 42 and 43 are specified as the transfer source of the coefficient information (step S4035). After that, the cloud server 2 transmits destination notification information including device identification information of the destination air conditioner 41 of the specified coefficient information, and destination notification information including device identification information of the specified source air conditioners 42 and 43. Notification information is generated (step S4036). After that, the generated movement source notification information is transmitted from the cloud server 2 to the air conditioner 41 (step S4037), and the generated movement destination notification information is transmitted from the cloud server 2 to the air conditioners 42 and 43. (Step S4038). Next, the processing after step S36 is executed.

According to this configuration, the user does not need to operate the management device 7, so the user's convenience can be improved.

In addition, in the modification described with reference to FIG. 26, in the air conditioner 41, the main user of the air conditioner 41 is identified within a preset user identification period, The air conditioner from which the coefficient information is moved may be determined based on the user identification information. In this case, after the user installs the air conditioner 41 in the dwelling unit H and then activates the air conditioner 41, the air conditioner 41 captures the image of the user with the image pickup device 481 and obtains the image of the user. Acquisition of image information and identification of the user of the air conditioner 41 from the surface temperature distribution of the user indicated by the acquired image information are repeatedly executed until a preset user identification period elapses. do. The user identification period is set, for example, to a period of about one week after the air conditioner 41 is started. After that, when the user identification period elapses, the air conditioner 41 identifies the user who uses the air conditioner 41 most frequently among the users identified during the user identification period as the main user.

In addition, in this modification, for example, the cloud server 2 may identify the main user of the air conditioner 41 . In this case, after the user installs the air conditioner 41 in the dwelling unit H and activates the air conditioner 41, the image of the user obtained by imaging the user with the imaging device 481 is captured by the air conditioner 41. Obtaining information and transmitting the obtained image information to the cloud server 2 together with the device identification information of the air conditioner 41 are repeatedly executed until a preset user identification period elapses. On the other hand, the cloud server 2 identifies the user of the air conditioner 41 from the surface temperature distribution of the user indicated by each piece of image information obtained from the air conditioner 41 . Then, after the user identification period has passed, the cloud server 2 selects the user with the highest usage frequency of the air conditioner 41 among the users identified during the user identification period as the main user. Identify as

According to this configuration, the weighting factors to be set in the neural network used in the air conditioner 41 can be appropriately selected.

22, the management device 7 acquires the coefficient attribute information from the air conditioners 42 and 43, and based on the user identification information included in the acquired coefficient attribute information, the coefficient information The source air conditioner may be determined.

Here, the operation of the control system according to this modified example will be described with reference to FIG. In addition, in FIG. 29, the same reference numerals as in FIG. 27 denote the same processing as in the modification described using FIG. First, when the user installs the air conditioner 41 in the dwelling unit H and then activates the air conditioner 41, the air conditioner 41 takes an image of the user with the imaging device 481 (step S7031). Next, the air conditioner 41 acquires image information of the user imaged by the imaging device 481, and identifies the user of the air conditioner 41 from the surface temperature distribution of the user indicated by the acquired image information ( step S7032). Here, the air conditioner 41 causes the user information storage unit 432 to store the user identification information of the identified user of the air conditioner 41 . Subsequently, the device identification information of the air conditioner 41 and the user identification information stored in the user information storage unit 432 are transmitted from the air conditioner 41 to the management device 7 (step S7033).

On the other hand, when receiving the device identification information and the user identification information from the air conditioner 41, the management device 7 specifies the air conditioner 41 to which the received device identification information is assigned as the destination of the coefficient information (step S7034). ). Thereafter, coefficient attribute request information requesting transmission of coefficient attribute information to the air conditioners 42 and 43 is broadcast from the management device 7 to each of the air conditioners 42 and 43 (step S7035). On the other hand, when receiving the coefficient attribute request information, the air conditioners 42 and 43 generate coefficient attribute information (step 7036). This coefficient attribute information corresponds to information including only the coefficient attribute information DAZ2 in the coefficient information shown in FIG. Next, the generated coefficient attribute information is transmitted from each of the air conditioners 42 and 43 to the management device 7 (step S7037). When the management device 7 receives the coefficient attribute information, it extracts the user identification information included in the coefficient attribute information. Search for coefficient attribute information that includes . Then, the management device 7 identifies the air conditioners 42 and 43 that have transmitted the coefficient attribute information that includes the same user identification information as the user identification information received from the air conditioner 41, as coefficient information transfer sources (step S7038). . After that, a series of processes after step S33 are executed.

In this modification, when the management device 7 receives the coefficient attribute information, it extracts the learning period information together with the user identification information included in the coefficient attribute information. and the length of the learning period calculated from the extracted learning period information is equal to or greater than a preset reference period. Then, the management device 7 receives coefficient attribute information that includes the same user identification information as the user identification information received from the air conditioner 41 and that the length of the learning period calculated from the learning period information is equal to or longer than a preset reference period. The air conditioners 42 and 43 that are the transmission source of the coefficient information may be specified as the transfer source of the coefficient information. Alternatively, the user identification information extracted by the management device 7 includes the same user identification information as the user identification information received from the air conditioner 41, and the end of the learning period indicated by the learning period information is preset to be the latest. Coefficient attribute information included in the period may be searched. Then, the management device 7 transmits coefficient attribute information that includes the same user identification information as the user identification information received from the air conditioner 41 and that the end of the learning period indicated by the learning period information is included in the latest preset period. The original air conditioners 42 and 43 may be specified as the transfer source of the coefficient information.

Further, in this modified example, for example, the cloud server 2 may be configured to identify the air conditioners 42 and 43 from which the coefficient information is transferred. In this case, the cloud server 2 acquires the device identification information and the user identification information of the air conditioner 41 from the air conditioner 41 via the management device 7, and transmits coefficient attribute request information to the air conditioners 42 and 43. , the coefficient attribute information may be obtained from the air conditioners 42 and 43 by transmitting the . Here, when the cloud server 2 receives the coefficient attribute information from the air conditioners 42 and 43, it extracts the user identification information included in the coefficient attribute information. The source air conditioners 42 and 43 may be identified by searching for coefficient attribute information that includes the same user identification information received from the user identification information. The cloud server 2 also provides destination notification information including device identification information of the destination air conditioner 41 of the specified coefficient information, and destination notification information including device identification information of the specified source air conditioners 42 and 43 . generate notification information; The generated movement source notification information is transmitted from the cloud server 2 to the air conditioner 41 , and the generated movement destination notification information is transmitted from the cloud server 2 to the air conditioners 42 and 43 .

According to this configuration, the management device 7 or the cloud server 2 stores the user identification information of the users of the air conditioners 41, 42, and 43 in association with the device identification information of the air conditioners 41, 42, and 43. no need to keep Therefore, the storage capacity of the auxiliary storage unit 703 of the management device 7 or the auxiliary storage unit 203 of the cloud server 2 can be reduced.

In the embodiment, the history information is transmitted directly from the air conditioner 41 to the cloud server 2, and the coefficient information is transmitted directly from the air conditioner 41 to the cloud server 2 or directly from the cloud server 2 to the air conditioner 41. I explained the example that is done. However, the method of transmitting history information and coefficient information is not limited to this. For example, the history information may be relayed from the air conditioner 41 by a terminal device (not shown) having a so-called tethering function and transmitted to the cloud server 2 . Further, the coefficient information may be relayed from the air conditioner 41 by a terminal device and transmitted to the cloud server 2, or may be relayed from the cloud server 2 by a terminal device and transmitted to the air conditioner 41. can be anything. Further, in the embodiment, an example in which the weather information is directly transmitted from the cloud server 2 to the air conditioner 41 has been described, but the present invention is not limited to this. It may be transmitted to the air conditioner 41 as a result. Here, as the terminal device, for example, a mobile terminal such as a smart phone can be adopted.

According to this configuration, even when the air conditioner 41 is not directly connected to the home network NT2, the history information or coefficient information can be transmitted from the air conditioner 41 to the cloud server 2, or It is possible to transmit coefficient information or weather performance information to Therefore, convenience for the user of the air conditioner 41 can be enhanced.

In the embodiment, the cloud server 2 receives operation history information and environment history information about other air conditioners of the same model as the air conditioner 41 from the customer, and actual weather information at the place where the other air conditioners are installed. and the weighting coefficients of the neural network are determined as the initial coefficients. However, not limited to this, for example, the cloud server 2 acquires information indicating the weight coefficient of the neural network used in each of the other air conditioners of the same model as the air conditioner 41 stored in the neural network storage unit 233, The initial weighting factors may be determined based on the distribution of the weighting factors indicated by the acquired information.

In the embodiment, the cloud server 2 receives history information including operation history information and environmental history information of air conditioners in other dwelling units in which air conditioners of the same model as the air conditioner 41 are installed from the customer server 9. I explained the example to get. However, not limited to this, for example, the cloud server 2 manages history information including operation history information and environmental history information of air conditioners in other dwelling units in which air conditioners of the same model as the air conditioner 41 are installed. It may have a function to In this case, the cloud server 2 can be configured not to transmit history information to the customer server 9 .

In the embodiment, the air conditioners 41, 42, and 43 have explained the configuration in which the device setting parameters are calculated using the weather information acquired from the weather server 3. However, the present invention is not limited to this. 42 and 43 may calculate device setting parameters without using weather information. In this case, for example, as shown in FIG. 30, the air conditioners 41, 42, and 43 may be configured without the weather information acquisition unit 417 and the weather information storage unit 436. FIG. Further, the cloud server 2 may be configured without the weather performance acquisition unit 212 and the weather information storage unit 232, as shown in FIG.

Also, the method of providing the program to the computer is arbitrary. For example, the program may be uploaded to a bulletin board (BBS (Bulletin Board System)) of a communication line and delivered to a computer via the communication line. The computer then activates this program and executes it in the same way as other applications under the control of an OS (Operating System). Thereby, the computer functions as the control unit 400 and the cloud server 2 of the air conditioners 41, 42, and 43 that execute the above processes.

The present invention is capable of various embodiments and modifications without departing from the broader spirit and scope of the invention. Moreover, the above-described embodiments are for explaining the present invention, and do not limit the scope of the present invention. That is, the scope of the present invention is indicated by the claims rather than the embodiments. Various modifications made within the scope of the claims and within the meaning of equivalent inventions are considered to be within the scope of the present invention.

INDUSTRIAL APPLICABILITY The present invention is suitable for automatic operation control of home electric appliances installed in a dwelling unit.

2 cloud server, 3 weather server, 7 management device, 9 customer server, 41, 42, 43 air conditioner, 61, 62, 63 operation device, 81 data circuit termination device, 82 router, 201, 401, 701 CPU, 202 , 402, 702 main storage unit, 203, 403, 703 auxiliary storage unit, 205, 405, 705 communication interface, 209, 409, 709 bus, 212 weather performance acquisition unit, 214 neural network calculation unit, 215, 420 coefficient determination unit , 218, 424 coefficient information generation unit, 219, 421 coefficient transmission unit, 233, 433 neural network storage unit, 400 control unit, 404 neuro engine, 406 measurement device interface, 407 operation device interface, 408 imaging interface, 411 environment information acquisition Unit 412 Image acquisition unit 413 Operation reception unit 414 Device control unit 415 User identification unit 416 Clock unit 417 Weather information acquisition unit 418 Coefficient acquisition unit 419 Coefficient setting unit 422 Device setting update unit 423 operation mode setting unit 425 history information generation unit 426 history information transmission unit 427 device identification information transmission unit 431 device setting storage unit 432 user information storage unit 434 history information storage unit 435 operation mode storage unit 436 weather information storage unit 461 measurement device 481 imaging device 711 reception unit 712 identification unit 713 display control unit 714 movement information generation unit 715 movement information transmission unit 731 layout information storage unit 732 movement information storage unit; 733 image storage unit, H dwelling unit, L10 input layer, L20 hidden layer, L30 output layer, NT1 outside network, NT2 inside network

Claims (13)

comprising a server and at least one device;
The server is
Coefficient information including information indicating neural network coefficients of a neural network for obtaining future device setting parameters of the device having a preset number of nodes and number of layers, and coefficient attribute information indicating attributes of the coefficient information, a first coefficient acquisition unit that acquires from the at least one device;
a first coefficient transmission unit configured to transmit the coefficient information and the coefficient attribute information acquired by the first coefficient acquisition unit to the at least one device when the coefficient request information is acquired from the at least one device. death,
Each of the at least one device comprises:
a second coefficient transmission unit configured to transmit, to the server, coefficient information indicating neural network coefficients of the neural networks used in each of the at least one devices, and coefficient attribute information;
a second coefficient acquisition unit configured to acquire coefficient information and coefficient attribute information from the server by transmitting coefficient request information requesting transmission of the coefficient information to the server;
a coefficient setting unit configured to set neural network coefficients of a neural network based on the coefficient information and the coefficient attribute information acquired by the second coefficient acquisition unit;
a neuro engine for obtaining future device setting parameters of the device from environmental parameters indicating the current environment using the neural network in which the neural network coefficients are set;
a device control unit that controls the device based on the device setting parameters obtained by the neuro engine;
control system. Each of the at least one device comprises:
a coefficient determination unit configured to determine the neural network coefficient based on operation history information indicating a history of device setting parameters of the device, environment history information indicating a history of environmental parameters of an environment in which the device operates, and the coefficient information; further prepare,
A control system according to claim 1 . each of the at least one device further includes a weather information acquisition unit that acquires weather information including actual weather information indicating past weather conditions and weather forecast information indicating future weather conditions;
The neural engine uses the neural network in which the neural network coefficients are set to obtain future device setting parameters of the device from the weather forecast information and environmental parameters indicating the current environment,
The coefficient determination unit includes operation history information indicating a history of device setting parameters of the device, environmental history information indicating a history of environmental parameters of an environment in which the device operates, the coefficient information, the coefficient attribute information, and the actual weather information. determining the neural network coefficients based on
3. A control system according to claim 2. Each of the at least one device comprises:
history information including operation history information indicating a history of device setting parameters of the device and environmental history information indicating a history of environmental parameters of an environment in which the device operates; and history attribute information indicating attributes of the history information are generated. further comprising a history information generator;
The server further comprises a coefficient determination unit that determines initial values of the neural network coefficients of the device based on the history information and the history attribute information generated by the history information generation unit.
4. Control system according to claim 2 or 3. The history attribute information includes device identification information that identifies the device, device setting type information that indicates the type of the device setting parameter, and environment type information that indicates the type of the environment.
5. A control system according to claim 4. The coefficient attribute information includes structural information including information indicating the number of nodes and the number of layers of the neural network,
When the second coefficient acquisition unit acquires a plurality of coefficient information and a plurality of coefficient attribute information indicating attributes of each of the plurality of coefficient information, the neural engine includes a neural network included in each of the plurality of coefficient attribute information. Based on the information indicating the number of nodes and the number of layers of, each neural network operation corresponding to each of the plurality of coefficient information and the plurality of coefficient attribute information is executed,
A control system according to any one of claims 1 to 5. an imaging device that images a user of the device;
a user identification unit that identifies the user based on the image captured by the imaging device;
The coefficient attribute information further includes user identification information that identifies the user identified by the user identification unit.
A control system according to claim 6 . the at least one device includes at least one first device and a second device;
further comprising a management device for managing coefficient information including weighting coefficient information indicating weighting coefficients of neural networks used in each of the at least one first device and the second device;
The management device
a specifying unit that specifies the device identification information of the at least one first device from which the coefficient information is transferred and the device identification information of the second device to which the coefficient information is transferred;
a movement information generation unit that generates movement source notification information including device identification information of the at least one first device identified by the identification unit and movement destination notification information including device identification information of the second device; ,
a movement information transmission unit that transmits the movement source notification information to the second device and the movement destination notification information to the at least one first device;
When the destination notification information is acquired, the second coefficient transmission unit of the at least one first device transmits information indicating a weighting coefficient of a neural network used in the at least one first device and a device identification of the second device. transmitting coefficient information and coefficient attribute information including information to the server;
When acquiring the source notification information, the second coefficient acquisition unit of the second device acquires the at least one first coefficient based on the device identification information of the at least one first device included in the source notification information. Acquiring coefficient information including weighting coefficient information indicating weighting coefficients of a neural network used in the device and coefficient attribute information from the server;
Control system according to any one of claims 1 to 7. a history information acquisition unit for acquiring operation history information indicating a history of device setting parameters of a device and environmental history information indicating a history of environmental parameters of an environment in which the device operates;
Initial coefficients of neural network coefficients of a neural network for obtaining future device setting parameters for each of at least one device having a preset number of nodes and a preset number of layers , based on the operation history information and the environment history information a coefficient determining unit for determining;
a coefficient information generating unit that generates coefficient information indicating the initial coefficient;
a coefficient acquisition unit configured to acquire, from the at least one device, coefficient information including information indicating neural network coefficients of the neural network and coefficient attribute information indicating attributes of the coefficient information;
when the coefficient request information is obtained from the at least one device, the coefficient information indicating the initial coefficient generated by the coefficient information generating unit, or the coefficient information and coefficient attribute information obtained by the coefficient obtaining unit are transmitted to the at least one device; a coefficient transmitter that transmits to one device,
server. Coefficient information including information indicating neural network coefficients of a neural network for obtaining future device setting parameters of a device having a preset number of nodes and layers, and coefficient attribute information indicating attributes of the coefficient information to a server a coefficient transmission unit for transmission;
a coefficient acquisition unit configured to acquire the coefficient information and the coefficient attribute information from the server by transmitting to the server coefficient request information requesting transmission of coefficient information including information indicating the neural network coefficients of the neural network; ,
a coefficient setting unit for setting neural network coefficients of a neural network based on the coefficient information and the coefficient attribute information acquired by the coefficient acquisition unit;
a neuro engine for obtaining future device setting parameters of the device from environmental parameters indicating the current environment using the neural network in which the neural network coefficients are set;
a device control unit that controls the device based on the device setting parameters obtained by the neuro engine;
machine. Coefficient information including information indicating neural network coefficients of a neural network for determining future device setting parameters of a device in which at least one device has a preset number of nodes and a preset number of layers, and coefficients indicating attributes of the coefficient information sending the attribute information to the server;
a step of the server obtaining, from the at least one device, coefficient information indicating neural network coefficients of the neural network and coefficient attribute information indicating attributes of the coefficient information;
a step of, when the server acquires the coefficient request information from the at least one device, transmitting the acquired coefficient information and the coefficient attribute information to the at least one device;
obtaining coefficient information and coefficient attribute information from the server by the at least one device transmitting coefficient request information requesting transmission of the coefficient information to the server;
the at least one device setting neural network coefficients of a neural network based on the obtained coefficient information and the coefficient attribute information;
the at least one device determining future device configuration parameters of the device from environmental parameters indicative of a current environment using the neural network with the neural network coefficients;
said at least one device controlling said device based on said determined device configuration parameters;
control method. the computer,
a history information acquisition unit for acquiring operation history information indicating a history of device setting parameters of a device and environmental history information indicating a history of environmental parameters of an environment in which the device operates;
Initial coefficients of neural network coefficients of a neural network for obtaining future device setting parameters for each of at least one device having a preset number of nodes and a preset number of layers , based on the operation history information and the environment history information a coefficient determining unit for determining,
a coefficient information generating unit that generates coefficient information indicating the initial coefficient;
a coefficient acquisition unit configured to acquire, from the at least one device, coefficient information including information indicating neural network coefficients of the neural network and coefficient attribute information indicating attributes of the coefficient information;
when the coefficient request information is obtained from the at least one device, the coefficient information indicating the initial coefficient generated by the coefficient information generating unit, or the coefficient information and coefficient attribute information obtained by the coefficient obtaining unit are transmitted to the at least one device; a coefficient transmitter for transmitting to one device;
A program to function as the computer,
Coefficient information including information indicating neural network coefficients of a neural network for obtaining future device setting parameters of a device having a preset number of nodes and layers, and coefficient attribute information indicating attributes of the coefficient information to a server a coefficient transmitter for transmission;
a coefficient acquisition unit configured to acquire the coefficient information and the coefficient attribute information from the server by transmitting to the server coefficient request information requesting transmission of coefficient information including information indicating the neural network coefficients of the neural network;
a coefficient setting unit configured to set neural network coefficients of a neural network based on the coefficient information and the coefficient attribute information acquired by the coefficient acquisition unit;
A neuro engine that uses the neural network in which the neural network coefficients are set to obtain future device setting parameters of the device from environmental parameters that indicate the current environment;
a device control unit that controls the device based on the device setting parameters obtained by the neuro engine;
A program to function as

Images