JPH0544972A - Controller for air conditioner - Google Patents

Abstract

PURPOSE:To effect comfortable air-conditioning operation, harmonized with the environment of a room or private taste, automatically and easily by a method wherein the title controller is provided with an estimating means, estimating the contents of operation of an user, a learning means, learning the content of operation, and a control signal producing means, for controlling an air conditioner. CONSTITUTION:A contents of operation estimating means 15, into which signals 111, 121, 131, 132, 141 outputted from a sensor 11, a storage means 12, an operating means 13 and a time detecting means 14 are inputted, estimates the contents of operation of the user of an operating means 13 using a neural net and outputs the estimating value 151 of the contents of operation, which is obtained by the result of said estimation, into a learning means 16. The estimated value 151 of the contents of operation, which is outputted from the operating content estimating means 15, is weighed by a weighing signal 161, obtained by learning whether the estimated value is right or not by the learning means 16 through a neural net. An operating contents estimating value 152 is inputted into a control signal producing means 17 and a control signal 171 is produced here to control an air conditioner 18. Accordingly, air-conditioning operation, harmonized with the environment of a room and private taste, can be effected.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an air conditioner control device for increasing indoor comfort by controlling indoor temperature, air volume and air direction. In an air conditioner, it is used for an air conditioner control device that automatically performs comfortable air conditioning operation.

[0002]

2. Description of the Related Art An air conditioner controller is disclosed in, for example, Japanese Patent Laid-Open No.
The example described in Japanese Patent Laid-Open No. 3-108145 is generally known. Hereinafter, a conventional air conditioner control device will be described with reference to the drawings.

FIG. 6 is a block diagram showing the structure of inference control software in a conventional air conditioner controller.
In FIG. 6, reference numeral 61 is an environment detection program that senses an air-conditioning environment such as temperature, humidity, and outside light from each sensor (not shown) and accumulates data, and 62 is to derive a specific fact from the above data. 63 is a condition setting program for setting conditions, 63 is a processing program for converting the accumulated knowledge database for air-conditioning into element syntactic meaning symbolization, and 64 is a predicate for the meaning symbolized knowledge. 65 is an inference program that infers by a logical programming language to obtain an optimal solution, and 65 is a drive interface control program that distributes the digital signal of the optimal solution to a control output display, an actuator, a motor, etc., and outputs a control output. Yes, 66 is an air conditioner.

Next, the operation of the above conventional example will be described with reference to FIG. First, in step 71, the facts from the environment, which are the inputs from the sensor group, are accumulated as data. Next, at step 72, a knowledge having a new meaning is created from the above data by using a control knowledge rule according to a condition or the like, for example, the IF to THEN format. For example, if a user has a 24-hour clock device electronically, and a day when the temperature exceeds 25 degrees at a specific time continues for 5 days, the knowledge rule of the judgment condition such as "summer" is digitally converted to a microcomputer. To be held in memory.

Next, in step 73, the knowledge generated by the control knowledge rule based on the above conditions etc. blows cold air to the lower direction for a short time because the person is lightly dressed in summer, for example, and then directly to the person. It is inferred that a pattern that changes the direction upward so that it does not hit is good and is controlled by the wind direction switching means. In addition, it outputs control outputs such as operation start time, operation mode switching, and temperature setting.

Next, in step 74, when the user performs an operation input for changing the wind direction to the inferred output, the information is stored. Then, in the next step 75, the accumulated data is analyzed to determine the user's habits and preferences, and the result is fed back to step 72,
It is configured so that the reasoning assumption can be changed.

[0007]

However, in the above-mentioned conventional device, knowledge is detected from the input of the sensor group,
In order to generate and infer new knowledge from the detected knowledge, it is necessary to generate a control knowledge rule based on the experience already obtained by an expert in advance, and it takes a lot of time and effort to create the control knowledge rule. There was a problem that it took.

The present invention solves such a conventional problem, and an air conditioner control device capable of automatically and easily performing comfortable air-conditioning operation according to the environment of the room and personal preference. The purpose is to provide.

[0009]

In order to achieve the above object, the present invention provides a plurality of sensor means for detecting indoor and outdoor environmental conditions, a storage means for holding the previous state of these sensor means, and a set temperature. Operation means for changing, time detecting means for knowing the change time of the set temperature, indoor / outdoor environmental conditions and time which are output values from the sensor means and the storage means, and the user's operation content from the temperature value set by the user And a means for learning the operation content by referring to the result obtained from each means, and a control signal for controlling the air conditioner based on the signal output from the operation content estimation means. And means.

[0010]

According to the present invention, with the above-described configuration, the operation contents of the user are estimated from the indoor and outdoor environmental conditions and the time output by the sensor means and the storage means and the temperature value set by the user, and the set temperature is set. By learning the operation content based on the result obtained from the operation means for changing the air conditioner, it is possible to automatically and easily perform the air-conditioning operation that suits the environment of the room and personal preference.

[0011]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic block diagram of an air conditioner controller according to an embodiment of the present invention. In FIG. 1, 1
1 is a sensor, 111 is a sensor signal, 12 is a storage means for storing the sensor signal, 121 is a stored sensor value, 13
Is an operation means such as a remote controller or an operation panel, 131 is a signal value of the operation means 13, 132 is an operation signal indicating whether or not the operation means 13 is operated, and 14 is the operation means 13.
The time detection means for detecting the time when the set temperature, the air volume, the wind direction, etc. are changed, 141 is the time signal, 15 is the operation content estimation means, 151 is the operation content estimated value, and 152 is the weight output from the operation content estimation means 15. The estimated operation content value, 16 is a learning unit, 161 is a weight signal obtained by the learning unit 16, and 17 is a control signal for generating a control signal for controlling the air conditioner by the operation content operation estimated value 152. Generating means, 171 is a control signal, and 18 is an air conditioner.

The operation of the above arrangement will be described below. Many sensors 11 (outside temperature sensor,
Suction temperature sensor, humidity sensor, etc.) are air conditioners 18
It is provided inside. When the signal 111 is output from the sensor 11, the contents such as the outdoor temperature, the suction temperature, and the humidity are output to the operation content estimation unit 15 and the storage unit 12. The storage means 12 stores the input sensor signal 1
The history of 11 in the past N seconds (N is a positive real number) is stored, and the stored state before the sensor, for example, the inclination sensor value 121 of the room temperature at intervals of N seconds is output. On the other hand, the operation means 13 outputs a signal value 131 such as a temperature, an air volume, and a wind direction set by the user. An operation signal 132 indicating that the set temperature or the like has been changed by the operation means 13 is output to the time detection means 14. When the signal 132 is input to the time detection means 14, the time detection means 14 calculates the time t1 from the time when the setting of the operation means 13 is changed to the present, and the operation content estimation means 15 is calculated.
Output to.

T1 = tn−t0 (1) tn: current time t0: time when the remote controller is changed

The operation content estimating means 15 which receives the signals 111, 121, 131, 132, 141 output from the respective means 11, 12, 13, 14 estimates the operation content of the operating means 13 using a neural network. Then, the resulting operation content estimated value 151 is output to the learning means 16. For example, "the set temperature is increased by 1 degree", "the air volume is increased by one step", and the like. The operation content estimation value 151 output from the operation content estimation unit 15 is weighted by the weight signal 161 from the learning unit 16 and the operation content estimation value 152 is weighted.
Is input to the control signal generation means 17, and the control signal 171 is generated here to control the air conditioner 18.

On the other hand, the output signals 111, 121, 131, 132, 141 from the respective means 11, 12, 13, 14 are also input to the learning means 16. The learning unit 16 learns whether or not the operation content estimation value 151 output by the operation content estimation unit 15 is correct by using the neural network, using the input value of each unit and the actual operation of the operation unit 13 as a teacher signal. After learning, the weight signal 16 of the neural network
1 is output to the operation content estimation unit 15, the operation content estimation unit 15 updates the operation content estimation value 151 using the weight signal 16, and outputs it as the learned operation content estimation value 152 to the control signal generation unit 17.

Next, referring to FIG. 2, operation content estimating means 1
5 will be described in detail. FIG. 2 shows the configuration when a neural network is used. The neural network in the present invention can use various pattern classification type models, but in the present invention, LVQ (Learning Vector Quan) is used.
tum: Bibliography T. Kohonen, "Self-Organization and Ass
ociative Memory ", 2ed, Springer Verlag.1988). In FIG. 2, 21 is an input signal normalization unit,
A reference vector unit 22 has N categories (events) and m reference vectors (corresponding to the weight signal 161 shown in FIG. 1) for each category. 23
Is a distance calculator, and 24 is a category calculator.

The operation of the above arrangement will be described below. The input signal 210 is normalized by the input signal normalization unit 21. That is, all input signals x1 are processed according to the following equation.

Sx1 = (xmax−x1) / (xmax−xmin) ... (1) Sx1: normalized signal value xmax: maximum value of input signal xmin: minimum value of input signal

The input signal 210 is the ambient temperature, the suction temperature, the slope of the suction temperature (suction temperature before T0 seconds),
The air volume, the set temperature, the time, etc. Input signal normalizer 21
The signal normalized by is output to the distance calculation unit 23 as an input vector 211. The distance calculator 23 calculates the distance between the input vector and all the reference vectors of each category of the reference vector unit 22. For example, the distance (d
A _j ) is obtained from the following equation (2), and then the directory 221 of the reference vector unit 22 is referenced to calculate the number of categories N and the number of reference vectors of each category. The contents of the directory 221 of the reference vector section 22 are shown in FIG.

DA _j = Σ (x _i − Ra _ji ) ² (2) dA _j : distance from j-th reference vector x _i : i-th input vector value Ra _ji : j-th reference vector the i th vector value

Next, the distance between the reference vector and the input vector is calculated for each category, and the same process is performed for all categories to calculate the distance. The total number (dnum) of distances calculated by the distance calculation unit 23 is
It becomes like the following formula (3).

Dnum = m1 + m2 ... + mN (3) m1: total number of references in A category m2: total number of references in B category mN: total number of references in N category

Next, the distance dN _j obtained by the distance calculation unit 23.
And dnum are output to the category calculator 24. The category calculation unit calculates the shortest distance dmin among the dnum distances, and further calculates the category to which the corresponding reference vector belongs. A general algorithm for outputting the shortest distance is shown in FIG. The output of the category calculation unit 24, which has calculated the category to which the reference vector having the shortest distance from the input vector belongs, is the operation content estimation unit 15 of FIG.
Is output. Then, the event of the air conditioner is associated with each category. For example, for the A category, "Set temperature is 1
"Increase temperature" or "Decrease set temperature by 1 degree" for category B.

Next, the learning means 16 will be described in detail with reference to FIG. In FIG. 5, reference numeral 51 is an input signal normalization unit, which normalizes an input signal common to the input signal normalization unit 21 of FIG. 2 and converts it into an input vector. Reference numeral 52 is also a reference vector part common to FIG. 53 is a reference vector search means, 54 is a reference vector generation unit,
Reference numeral 55 is a reference vector learning unit.

The operation of the above arrangement will be described below. The input signal normalized by the input signal normalization unit 51 is input to the reference vector search unit 53 as an input vector. Further, the correct category (teacher data) of the input vector is also input to the reference vector search unit 53 at the same time. The reference vector search unit 53 refers to the directory of the reference vector unit 52, and when the correct category (TC) of the input vector is the A category, for example, when the A category reference number has the following relationship, Perform processing.

If ARNUM <ARNUMmax, the processing of the reference vector generation unit 54 is performed. If ARNUM ≧ ARNUMmax, the processing of the reference vector learning unit 55 is performed. ARNUM: Reference number of A category ARNUMmax: Maximum reference number of A category

The reference vector generator 54 generates the input vector as a reference vector of the correct category (TC). Here, a reference vector of category A is generated and ARNUM is incremented by 1.

The reference vector learning unit 55 compares the output value of the operating means 13 with the correct category of the input vector, and if they match, the reference vector with the shortest distance is brought closer. If the output value of the operating means 13 does not match the correct category of the input vector, the reference vector with the shortest distance is moved away.
That is, if OC = TC, then RV = RV + α (RV−Sx) If OC ≠ TC, then RV = RV−α (RV−Sx) RV: Reference vector with the shortest distance from the input vector OC: Output category (RV Category: TC: correct category of input vector Sx: input vector α: learning rate

Therefore, the reference vector learning unit 55
Then, until the reference number exceeds the maximum value of the reference, the input vector is generated as the reference vector of the correct category of the input vector. Further, after the maximum value is exceeded, when the output category and the input category match, the reference vector is moved a little closer to the direction of the input vector, and when they do not match, the reference vector is moved a little away from the input vector. The distance to approach or move away is set by α.

As described above, the fixed knowledge representation for estimating the standard environment desired by the user collects data on how a person controls comfort in various environments, and collects environmental data and Can be obtained by training the neural network to learn the correlation with the control data.

As described above, according to the above-described embodiment, by using the neural network, the rule of relating the input and the output from the sensor group is not necessary, the operation content of the user is learned, and the operation is pre-fetched. By making a prediction,
It is possible to realize air conditioning in automatic mode operation that suits the environment of the room and personal preference, and to relieve the user from the hassle of setting remote controls by learning.

[0032]

As is apparent from the above embodiments, the present invention is based on the indoor and outdoor environmental conditions, which are the output values from the respective sensor means and storage means, the time, and the user's operation, and the operation content estimating means and operation. Since the content learning means estimates and learns the content operated by the user, it is possible to automatically and easily perform the air-conditioning operation that suits the environment of the room and personal preference.

[Brief description of drawings]

FIG. 1 is a schematic block diagram showing a configuration of an air conditioner control device according to an embodiment of the present invention.

FIG. 2 is a detailed block diagram of an operation content estimation unit in the device.

FIG. 3 is a schematic diagram showing the contents of a directory of a reference vector part in the same device.

FIG. 4 is a flowchart showing an example of an algorithm for outputting the shortest distance in the same device.

FIG. 5 is a detailed block diagram of learning means in the device.

FIG. 6 is a block diagram showing the configuration of inference control software in a conventional air conditioner controller.

FIG. 7 is a block diagram showing the operation of inference control software in the same device.

[Explanation of symbols]

 11 sensor 111 sensor signal 12 storage means 121 stored sensor value 13 operating means 131 signal value of operating means 132 operation signal indicating whether or not operated 14 time detection means 141 time signal 15 operation content estimation means 151 operation content estimated value 152 Weighted operation content estimation value 16 Learning means 161 Weight signal 17 Control signal generating means 171 Control signal 18 Air conditioner

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor ▲ Yoshi ▼ T ▲ Kuni ▼ husband 3-10-1 Higashisanda, Tama-ku, Kawasaki City, Kanagawa Matsushita Giken Co., Ltd. (72) Inventor Ikuo Akamine Osaka Prefecture 1006 Kadoma, Kadoma-shi, Matsushita Electric Industrial Co., Ltd. (72) Inventor: Shin Shimizu, 1006 Kadoma, Kadoma, Osaka (72) Inventor: Matsuhiko Electric Industry Co., Ltd. (72) Katsuhiko Fujiwara 1006, Kadoma, Kadoma, Osaka Address: Matsushita Electric Industrial Co., Ltd. (72) Inventor: Akira Yokouchi 1006 Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Industrial Co., Ltd.

Claims (4)

[Claims] 1. A plurality of sensor means for detecting environmental conditions indoors and outdoors, a storage means for holding a previous state of the sensor means, an operation means for changing a set temperature, and a time detection for knowing a change time of the set temperature. And means for estimating the operation content of the user from the indoor and outdoor environmental conditions and the time and the temperature value set by the user, which are output values from the sensor means and the storage means, and means obtained from the respective means. An air conditioner comprising means for learning the operation content with reference to the result, and means for generating a control signal for controlling the air conditioner based on the signal output from the operation content estimation means. Control device. 2. The air conditioner control apparatus according to claim 1, wherein the means for estimating and learning the operation content uses a neural network. 3. The air conditioner control apparatus according to claim 1, wherein the sensor means for detecting indoor and outdoor environmental conditions includes means for detecting indoor and outdoor temperature, air volume of the air conditioner, and humidity. 4. The storage means for holding the previous state of the sensor is N
The air conditioner control device according to any one of claims 1 to 3, further comprising storing intake air temperature gradients of the air conditioner at intervals of seconds (N is a positive real value).