JP3096507B2 - Operation content learning device and air conditioner control device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an operation content learning device used for home electric appliances and the like and an air conditioner control device using the same.

[0002]

2. Description of the Related Art In recent years, home electric appliances such as air conditioners are equipped with a microcomputer, and an operation according to a user's preference can be obtained by making the device memorize operation contents. . FIG. 3 shows a configuration of a device for learning operation contents of a remote controller (hereinafter, abbreviated as a remote controller) in a conventional air conditioner.
In FIG. 3, 31 is input data, 32 is a remote controller, 3
3 is a remote control operation estimating means, 34 is a learning means, and 35 is a device control means.

Next, the operation of the above-mentioned conventional device will be described. First, the signal flow during learning will be described. The user operates the remote controller 32 to inform the learning means 34 of the user's intention by the signal value 32a. Remote control signal value 32a
Indicates one of two types of remote control operations, for example, one of "hot" and "cold" in an air conditioner. When the signal value 32a of the remote control operation is transmitted to the learning means 34, the input data 3 at the time when the remote control operation is performed is
The learning means 34 learns the user's intention on condition of 1.

On the other hand, in the flow of estimating remote control operation,
When the input data 31 is input, the remote control operation estimating means 33 uses the learning means 3 based on the condition of the signal value 31a.
4, the user's intention is predicted in advance based on the knowledge 34a obtained in the past learning obtained from 4 and transmitted to the control means 35 of the device as an estimated value 33a. In this case, the estimated value 33a is transmitted as two types of signals as in the case of the remote control operation. For example, for an air conditioner, "hot" or "cold"
The estimated value 33a is transmitted to the device control unit 35, and is converted into a device control signal by the device control unit 35.

[0005]

However, in the above-mentioned conventional operation content learning apparatus, there is a problem that even if a user operates a remote controller by mistake, the apparatus learns all of them. In addition, the air conditioner has a problem that only two types of operations, "hot" and "cold", can be learned, so that the degree of "hot" and "cold" cannot be learned.

The present invention has been made in view of such a conventional problem,
It is an object of the present invention to provide an excellent operation content learning device capable of learning by judging the degree of operation content without learning an erroneous operation and an air conditioner control device using the same.

[0007]

In order to achieve the above object, the present invention provides a storage means for storing input data, and a value + M or -M for each operation of two kinds of instructions whose operation contents conflict with each other. An operation unit that outputs M, and a value + M output from the operation unit for N seconds from the first operation.
And a counter for accumulating -M (M is an integer), the accumulated output of the counter and the condition of the input data at the time of operation stored in the storage means are learned, and the operation content of the user is estimated according to the condition of the input data. Means for performing the operation.

[0008]

According to the present invention, with the above configuration, the initial value of the counter is set to, for example, 0, and the values of two types of conflicting instructions are set to +1,-.
By integrating the operation content for N seconds from the first operation as 1, for example, if the operation is redone with an incorrect input, the total of the counter becomes 0, so that the erroneous operation is not learned, and By accumulating the operation contents, it is possible to quickly transmit the degree of instruction desired by the user to the device, thereby improving the operability and communicating the user's accurate intention to the device. .

[0009]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic block diagram of an operation content learning apparatus showing one embodiment of the present invention. In FIG. 1, 1 is input data, 2 is input data storage means, 3 is a remote controller (operation means), 4 is a counter, 5 is N seconds detection means, 6
Is a remote control operation estimating means, 7 is a learning means, and 8 is a device control means.

Next, the operation of the above embodiment will be described.
First, the operation at the time of learning will be described. When the user outputs his / her own intention through the signal 3a using the remote controller 3, the input data 1 at the time of operating the remote controller becomes the signal 1
The data is stored in the input data storage means 2 by a. The initial value of the counter 4 before the first remote control operation is performed is 0
Is set, and when two types of remote control operations are A and B, + M = + 1 for A and −M = − for B
1 is counted by the counter 4. When the first remote control operation is performed, the detection means 5 starts detecting time for N seconds, and outputs signals 5a and 5b when N seconds have elapsed, and the counter 4 and the input data storage means 2 indicate that N seconds have elapsed. Inform means 7. Upon receiving the signal 5b, the learning means 7 extracts the input data N seconds before stored in the input data storage means 2 using the signal 2a. Further, the user's intention and the degree thereof are determined based on the accumulated output value 4a of the counter 4, and the conditions under which the user operates the remote controller are learned based on the input data signal 2a, and stored as knowledge in the learning means 7. . There are various algorithms for learning, but a method using a neural network (neural network model means) is recommended. The counter 4 is initialized based on the elapsed signal 5a for N seconds, and 0 is set.

Next, the processing operation at the time of remote controller operation estimation will be described. The input data 1 is, as an input signal 1b,
It is transmitted to the remote control operation estimating means 6. Based on the output signal 7a of the learning means 7, the remote control operation estimating means 6 determines whether or not the input signal 1b of the input data 1 is close to the input data condition learned by the remote control operation of the past user. If the input data is close to the past remote control operation input data, the remote control operation estimating means 6 informs the user of the degree of intention by a signal 6.
The signal is output as a and transmitted to the device control means 8, where it is converted into a device control signal.

Next, an embodiment of an air conditioner control device incorporating the operation content learning device will be described with reference to FIG. 2, reference numeral 21 denotes a plurality of sensors for sensing indoor and outdoor environment information; 22, a storage unit for storing a sensor signal value 21a; 23, an operation unit which is a remote control or an operation panel; and 24, an operation unit 23 for N seconds. A counter for counting the operation, 25 is a detecting means for N seconds, 26 is an operation estimating means, 27 is a learning means, 28 is a device controlling means, and a control signal generating means 29 for generating a control signal based on the operation estimating value; And an air conditioner 30 controlled by the control signal.

The operation of the air conditioner control device configured as described above will be described below. First, the flow at the time of learning will be described. From a number of sensors 21 (outside air temperature sensor, suction temperature sensor, humidity sensor) provided in the air conditioner 30, signals 21a and 21b such as outdoor temperature, suction temperature, and humidity are stored in the storage unit 22 and the operation content estimating unit. 26 is output. Storage means 22
Detects the input sensor signal 21a for N seconds
Is stored for N seconds under the signal 25a from When a request for the user's room temperature, air volume, wind direction, and the like is input by the operation means 23 which is a remote control or an operation panel, the operation means 23 outputs a signal value 23a as the intention of the user. Here, the signal value 23a is output to the counter 24 as two kinds of contradictory signal values + M = + 1 and -M = -1. The meaning of each signal is that +1 is "cold" and -1
Is set to "hot".

Signal value 23a input to counter 24
Are integrated for N seconds based on a signal 25b from the detection means 25 for N seconds. The counter 24 has an upper limit of +3 and a lower limit of -3.
The value of +3 or more is set to +3, and the value of -3 or less is set to -3. The N second detecting means 25 detects the time of N seconds from the first operation of the operating means 23, and when N seconds have elapsed, the signals 25a, 25b, 25c cause the storing means 22, the counter 24 and the learning means 27 to transmit N seconds. Tell them what you did. Counter 24 transmitted to learning means 27
Is limited to ± 3, so an integer value in the range of ± 3 is the output value 24a. Output value 24a
Has the following meaning to indicate the degree of dissatisfaction of the user with the air conditioner. +3: very cold +2: cold +1 ... slightly cold 0: no dissatisfaction -1: slightly hot -2: hot -3: very hot

Here, if the user has operated the operation means 23 by mistake, if the correction operation is executed within N seconds, the incorrect operation will not be a learning target. For example, in the first operation, regardless of the cold condition, even if -1 is pressed,
If +1 is continuously pressed within N seconds, it becomes 0, and no learning is performed.

The learning means 27 and the operation content estimating means 26
A neural network is used, and the operation contents of the operation means 23 are learned and estimated. Signal 22 output from storage means 22 for storing signal value 21a from sensor 21
a and the output value 24a of the counter 24
The learning means 27 learns the neural network using the signal 22a from the storage means 11 as an input value and the output value 24a of the counter 24 as a teacher signal.

Next, the flow of signals when estimating the operation contents will be described. After the learning is completed, when the signal value 21b of the sensor 21 is input to the operation content estimating unit 26, the learned knowledge is transmitted from the learning unit 27 in the neural network to the reference information 2.
7a is output to the operation content estimating means 26. This reference information 27a is compared with the signal value 21b from the sensor 21 to determine whether or not the input condition has been operated by the user, and is output as a guess signal 26a if the input condition is close to the operated condition. You. The guess signal 26a takes an integer value up to ± 3 and represents the above-mentioned meaning as a signal, and is transmitted to the control signal generating means 29, where it is converted into an actual control signal 29a and transmitted to the air conditioner 30. Then, the temperature, the air volume, and the wind direction are controlled.

[0018]

As described above, according to the present invention, the storage means for storing the input data and the operation for outputting the value + M or -M per one operation for each of two kinds of instructions whose operation contents are inconsistent. Means, a counter for integrating the values + M and -M output from the operation means for N seconds from the first operation, and the conditions of the output of the counter and the input data at the time of operation stored in the storage means. Means for estimating the content of the user's operation according to the conditions of the input data, so that the user does not learn erroneous operations, and by accumulating the contents of the operation, the instruction desired by the user can be obtained. It is possible to quickly transmit the degree of operation to the device, thereby improving the operability and transmitting the user's accurate intention to the device.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram of an operation content learning device according to an embodiment of the present invention.

FIG. 2 is a schematic block diagram showing an embodiment of an air conditioner control device incorporating the operation content learning device of the present invention.

FIG. 3 is a schematic block diagram showing a conventional operation content learning device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input data 2 Input data storage means 3 Remote control 4 Counter 5 N second detection means 6 Remote control operation estimation means 7 Learning means 8 Device control means 21 Sensor 22 Storage means 23 Operation means 24 Counter 25 N second detection means 26 Operation content estimation means 27 Learning means 28 Equipment control means 29 Control signal generation means 30 Air conditioner

Continuing from the front page (72) Inventor ▲ Yoshi ▼ Kunio Tada 3-10-1, Higashimita, Tama-ku, Kawasaki City, Kanagawa Prefecture Inside Matsushita Giken Co., Ltd. (72) Inventor Ikuo Akamine 1006 Kadoma Kadoma, Kadoma City, Osaka Prefecture Address Matsushita Electric Industrial Co., Ltd. (72) Inventor Makoto Shimizu 1006 Odoma, Kadoma City, Osaka Prefecture Inside Matsushita Electric Industrial Co., Ltd. (72) Inventor Katsuhiko Fujiwara 1006 Odakadoma, Kadoma City, Osaka Matsushita Electric (58) Fields surveyed (Int.Cl. ⁷ , DB name) F24F 11/02 G05B 13/02 G06F 15/18

Claims (4)

(57) [Claims] 1. A storage means for storing input data, an operation means for outputting a value + M or -M for each operation for two kinds of instructions whose operation contents are inconsistent, and N seconds after the first operation. A counter for accumulating the values + M and -M output from the operation means, and learning the accumulated output of the counter and the condition of the input data at the time of operation stored in the storage means, according to the condition of the input data. Operation learning device comprising: means for estimating the operation content of a user by using the method. 2. The operation content learning device according to claim 1, wherein the means for learning and estimating the operation content is constituted by a neural network (neural network model means). 3. An upper limit value and a lower limit value of a counter are provided,
The operation content learning device according to claim 1 or 2, wherein the degree of intention of the user is represented by an integration result of the values + M and -M. 4. An air conditioner control apparatus using the operation content learning device according to claim 1, wherein the values + M and −M correspond to cold and hot of the user, respectively.