JP2010133692A - Air conditioner - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To control air-conditioning according to a sleep state by determining user's sleep state. <P>SOLUTION: This air conditioner 10 includes an infrared ray sensor 1 detecting a person without contact, an image display section 2 representing a temperature distribution detected by the infrared ray sensor 1 as a thermal pixel image 8, a memorizing device 3 memorizing the information relating to the air-conditioning control including the thermal pixel image 8, a body motion analyzing section 4 specifying a head part, a trunk part, hand and leg parts of the detected person, and analyzing magnitude and number of times of body motion of at least one predetermined part and/or a surface temperature of the region, a sleep state determining section 5 for estimating and determining the sleep state of the person on the basis of a result of the analysis by the body motion analyzing section 4, and the air-conditioning control section 6 controlling the air-conditioning on the basis of the determination by the sleep state determining section 5. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an air conditioner that grasps the state of a user and provides air conditioning that is comfortable for the user.

Conventionally, technologies related to air conditioning during sleep have been developed. For example, the air conditioner with a sleep mode controls the air conditioning in the room according to the operation time of the sleep mode and the intake temperature of the air conditioner (the air temperature around the air conditioner body).
In addition, some thermometers are embedded in beds and pillows, and the temperature of a person is measured to control air conditioning (Patent Documents 1 and 2).

  Moreover, body motion information is acquired by a non-contact sensor such as an air conditioner that tracks the face with a pyroelectric sensor and an infrared sensor, and judges comfort from the surface temperature of the face (Patent Document 3), or a pyroelectric sensor, and sleeps. There is also a method of grasping the state and performing air conditioning control (Patent Documents 4, 5, and 6).

Japanese Patent Laid-Open No. 05-088758 Japanese Patent Application Laid-Open No. 07-042999 Japanese Patent Laid-Open No. 02-157551 JP 05-106899 A Japanese Patent Laid-Open No. 02-068439 JP 2007-285562 A

However, the conventional case has the following problems. In the air conditioner with sleep mode, the target temperature is gradually brought closer to room temperature according to the operating time of the same mode, or the target temperature is set closer to the room temperature than the user setting according to the suction temperature of the air conditioner. However, with these methods, individual differences until the user falls asleep was not taken into account, and the control was unambiguous.
Moreover, since the air conditioning control is performed so that the optimal body temperature history set in advance is approached in the technique disclosed in Patent Document 1, it is not possible to take into account the user's occasional state differences such as the user's clothing amount and detailed sleep depth.
Furthermore, regarding the technique shown in Patent Document 2, since the user needs to be in contact with the thermometer, it cannot respond to the user's turning over, and is affected by the temperature change due to the change of the contact part to the thermometer. was there. In addition, since the thermometer is in contact with the body, there is also a problem that disturbs the sleep state.

  The method shown in Patent Document 3 follows the face with a visible image and an infrared sensor and judges comfort from the surface temperature, so the user's clothes, activity, and end temperature cannot be judged. I can't judge whether it is in the state. Further, when the resolution is rough, the average temperature of the face and the background space is output, so that the face cannot be determined accurately. In order to accurately grasp the skin temperature, an expensive element having high resolution is required. In addition, when the signal is above the minimum level of the human skin surface temperature and a signal below the optimal level is obtained, the portion showing the maximum value is the face. It cannot be determined accurately when the background temperature is low. Furthermore, the ambient environmental temperature is not acquired by an infrared sensor or the like, and the surrounding environmental conditions cannot be taken into consideration.

  In Patent Documents 4, 5, and 6, body motion information is acquired by a non-contact sensor such as a pyroelectric sensor, and the air conditioning is controlled by grasping the sleep state. Since it cannot be judged, the sleep depth can be distinguished only about two stages. In addition, it is impossible to determine what the comfort level is, the sensor output depends on the amount of clothes including the futon, and further, the ambient temperature cannot be acquired, so that comfortable sleep cannot be escorted.

  The present invention has been made in consideration of the above-described problems, and an object of the present invention is to provide an air conditioner capable of air conditioning control according to each person's state (for example, sleep state) and providing comfortable sleep and the like. Is what you get.

The air conditioner according to the present invention specifies a human detection means for detecting a person in a non-contact manner, a head, a torso, and a limb of the detected person, and a body motion and a surface that are movements of the specified at least one part. Based on the analysis result of the body motion analysis unit that analyzes the temperature and the body motion analysis unit, the state of the person is identified, and the air conditioning control according to the state is performed.
Further, the analysis of the body motion analysis unit includes detection of the size and number of times of the body motion, and based on the size and number of times of the body motion analyzed by the body motion analysis unit and the surface temperature, A sleep state determination unit that predicts a sleep state and an air conditioning control unit that performs air conditioning control based on the determination of the sleep state determination unit.

  According to the air conditioner of the present invention, it is possible to grasp the state of a person relatively easily based on the movements of the person's head, trunk, limbs, and changes in surface temperature. Further, air conditioning control suitable for sleep is possible based on the information. This can also contribute to energy saving by eliminating useless air conditioning.

The figure which shows the thermal pixel image which shows the temperature distribution of the area | region detected with the structure of the air conditioner in Embodiment 1 of this invention, and the non-contact-type temperature detection apparatus. The block block diagram which shows the main functions of the air conditioner in Embodiment 1. FIG. Explanatory drawing which shows the outline | summary of the state change at the time of a person's sleep. 5 is a flowchart showing the operation of the air conditioner according to Embodiment 1. The network block diagram in Embodiment 2 of this invention. 5 is a flowchart illustrating an example of deep sleep determination operation performed by the sleep state determination unit according to the first embodiment.

Embodiment 1 FIG.
FIG. 1 is a diagram showing a configuration of an air conditioner according to Embodiment 1 of the present invention and a thermal pixel image showing a temperature distribution in a space detected by a non-contact temperature detecting device, and FIG. 2 is an air conditioner according to Embodiment 1. It is a block block diagram which shows the main functions.

The air conditioner (indoor unit) 10 according to the first embodiment includes a non-contact type temperature detection device 1 that detects a temperature distribution in a predetermined region (for example, a bed part of a bedroom or the entire living room), a temperature An image display unit 2 that displays the temperature distribution detected by the detection device 1 as a thermal pixel image 8, a storage device 3 that stores various types of information including the thermal pixel image 8, and the thermal pixel image 8 is within the above-described region. Based on the analysis output of the body motion analysis unit 4 and the body motion analysis unit 4 that identifies the human head, torso, and limbs and analyzes the size and number of body motions of each site and the surface temperature of each site A sleep state determination unit 5 that predicts the sleep state of the person and an air conditioning control unit 6 that performs air conditioning control of the air conditioning operation unit (air conditioner main body) 7 based on the determination output of the sleep state determination unit 5. Normally, the body motion analysis unit 4, the sleep state determination unit 5, and the air conditioning control unit 6 are configured by one or a plurality of microcomputers. Here, the air-conditioning control unit 6 also controls the operations of the parts 1 to 5 of the air conditioner (indoor unit) 10. Here, a part (human detection function) of the temperature detection device 1, the image display unit 2, the storage device 3, and the body motion analysis unit 4 functions as a human detection unit.
In addition, the air intake port of the air conditioner 10 is provided with a room temperature detection sensor 11 that detects the room temperature, and the air outlet of the air conditioner 10 is provided with a wind direction adjuster 12 that adjusts the direction of the blown air.

The non-contact type temperature detection device 1 is an infrared sensor such as a thermopile, for example, which acts as part of a human detection means. The detection of the temperature distribution in the predetermined region may be performed simultaneously using the temperature detection device 1 composed of a plurality of sensors that can cover the region, or the temperature detection device 1 composed of one or a small number of sensors. The temperature distribution may be detected by scanning. The temperature distribution detected by the temperature detection device 1 is color-coded for each temperature in the image display unit 2 like a thermo viewer and displayed as a thermal pixel image 8.
In addition, each said elements 1-6 may be comprised integrally with the air conditioning operation | movement part (air conditioner main body) 7, and those all or one part may be comprised separately from the air conditioning operation | movement part 7. Good.

Next, the operation of the air conditioner 10 will be described. The temperature detection device 1 repeatedly detects a temperature distribution in a predetermined region in a predetermined cycle. The temperature distribution detected by the temperature detection device 1 is displayed on the image display unit 2 as a thermal pixel image 8 that is color-coded according to the temperature. Each thermal pixel image 8 detected for each predetermined cycle is stored in the storage device 3.
Then, from the shape and position of the detected temperature distribution (or a specific temperature zone) of the thermal pixel image 8, the human head, torso, and limbs are specified, and the positions of the pixel aggregates corresponding to them are identified. From the change, the magnitude of the movement of those parts and the number of times are detected. In this specification, these head, torso, and limbs are also referred to as parts, and the movement of these parts is referred to as body movement.
When each part cannot be determined from the shape and position of the pixel aggregate, for example, the pixel aggregate part where motion is detected is determined to be a person, the temperature of that part is divided into three stages, The head, torso, and limbs may be considered. Further, based on the shape of each part and the temperature distribution, a pixel aggregate having a similar form is determined as the same person.
When the temperature of the air-conditioning target space is close to the radiant heat temperature of the human body, the surroundings are air-conditioned in advance until the temperature detection device 1 reaches a temperature at which a person can be detected.

  Further, it is determined in a stationary state (when the amount of activity is equal to or less than a predetermined value) from the center of gravity of the pixel aggregate determined to be one person or the total movement amount of the outline of the pixel aggregate. The state of the aggregate is monitored for a period of time. And if there is a relatively small change that is not a change of position or change of position, and it is determined that it is due to a person, it is determined to be the person's head and limbs, and the change is Judged to be dynamic. Then, it is determined that a person with a large number of body movements is a limb and a person with a small number of movements is a head. Further, when the resolution of the human detection sensor is high, the determination is made in consideration of the movable distance of those parts. In general, the movable distance of the head is short and that of the limb is large and long. In addition to these, if the body orientation calculated from the length of the pixel assembly and the temperature of the pixel assembly are taken into account, and the human head, torso, and limbs are judged, the parts are almost accurate. Can be specified.

If the body movement of the identified person's head, torso, and limbs falls below a predetermined threshold that is less than or equal to the number of times determined to have occurred in a certain period of time, the person falls asleep (becomes sleepy) It can be judged. For example, in the case of body movements of 1 to 2 times or less in 4 to 8 minutes can be set as falling asleep. Such a threshold learns the difference between motion during waking and motion during sleep and corrects every few cycles. In the case of the first time, there is a method of using average human body motion data and determining sleep when the number of body motions becomes a predetermined value or less.
Due to the scanning temperature detection device, if it takes time to make a decision to fall asleep, speeding up the sleep detection by decreasing the temperature resolution and shortening the acquisition time, or reducing the pixel resolution by thinning out the pixels Can be
When the temperature resolution or pixel resolution is lowered, the temperature detection device can be scanned at high speed, and the temperature acquisition time can be shortened. This mode can be changed depending on the model and situation, such as performing at high speed when waking up, normal operation after going to sleep, or vice versa, or constantly moving at high speed.

  Further, based on the body movements of the head, torso, and limbs specified by the above method, it is determined whether the body movement during sleep is a rollover or a simple movement of the limbs. For example, when there is body movement in the torso or when the change in the head region is large, it is determined that the face surface has changed so much as to change the orientation, and it is determined to turn over. Further, it is assumed that the higher temperature of the head is the front of the face, and the direction of turning over can be determined from the direction of the face. Similarly, when the movement of the limbs is large, it is determined to turn over. In this way, the size and the number of body movements of each part are calculated, and the sleep depth is determined from a preset general tendency of body movement change or user tendency. Since body motion varies from person to person, more accurate determination can be made by storing a top and bottom peak value of body motion change and providing a learning function for setting a threshold value for judging rollover based on the change ratio.

FIG. 3 is an explanatory diagram showing an outline of a state change during sleep of a person. Here, the thick solid line indicates the progress of sleep, the dotted line (A) indicates the human body temperature, the alternate long and short dash line (B) indicates the amount of sweating, and the alternate long and two short dashes line (C) indicates an example of air conditioning control. In addition to getting up, the progress of sleep depends on sleep depth, REM sleep (state where the brain is active in a state close to awakening), sleep depth 1, sleep depth 2, sleep depth 3, A distinction is made between six sleep states with a sleep depth of four.
Body movements decrease when sleep begins. There is a difference in the size and number of body movements when waking up and sleeping. It is thought that as the depth of sleep increases, body movements disappear and the brain rests. Once the sleep depth 4 is reached, the sleep depth becomes shallow and REM sleep starts. Such a repetition is periodically performed during sleep, and the duration of REM sleep gradually increases toward dawn. When the sleep depth changes, there is a body movement, but in many cases, the body movement is observed when the sleep depth is 1, during awakening. In addition, body movement is observed although it is not large at the time of REM sleep in the late stage of sleep. The length of the stationary duration without body movement also changes depending on the thermal environment. The number of REM sleeps is calculated from the information, and you can get up comfortably when you wake up immediately after REM sleep. As can be seen from sleep disorders such as night sweats and coldness, the influence of the thermal environment on sleep is great, and therefore sleep control by air conditioning control is effective.
If body movements during sleep cannot be detected and the sleep state cannot be determined, a cycle such as circadian rhythm (internal clock rhythm) is applied and the next detection opportunity is waited for. Further, even when body movements of the head, torso, and limbs are detected in the same amount even after a lapse of time, the same method is applied to the body movement of the cycle.
The sleep state can be determined by detecting the body movement of at least one part of the head, torso and limbs, but from the viewpoint of accuracy, the head, torso and limbs It is preferable to determine the sleep state in consideration of all body movements.

  In addition to body movement, changes in human body temperature are also important indicators for determining sleep depth. As shown in FIG. 3, during sleep, heat is released to lower the deep body temperature of the person, and the surface temperature at the end of the person rises higher than usual. Thereafter, heat radiation is gradually completed as the sleep depth 4 is approached, and the human deep body temperature rises linearly from the sleep depth 4 toward the wake-up. Heat is released before the depth of sleep is deepened, and while the sweating cycle is performed, the deep body temperature gradually rises as you get up. The surface temperature of the part related to heat dissipation such as the limbs has a temporal phase difference from the deep body temperature because of the heat dissipation relation. For a person with a large phase difference, the deep body temperature and the surface temperature may be 90 degrees shifted by a sine wave. As the temperature of the face is linked to the amount of brain activity, it decreases as the sleep depth increases, and increases as the sleep depth decreases, such as REM sleep. The amount of sweating has a certain degree of correlation with the depth of sleep. The surface temperature of the torso is similar to that of the deep body temperature, and the amplitude and phase are different but slightly change. The ratio between sleep depth 4 and REM sleep is inversely proportional. That is, deep sleep such as sleep depth 4 becomes longer after falling asleep, but the time for REM sleep becomes longer as it approaches dawn. The sleep depth 1 to 3 is sandwiched between the sleep depths 4 to REM, and the cycle is awakened in several cycles. In addition, sleep depth 2 is dominant except sleep depth 4 and REM sleep. It is important to adjust the body temperature in order to lengthen the time zone between sleep depth 4 and REM sleep, and adjusting the thermal environment by air conditioning control and assisting more optimal body temperature adjustment is a good sleep control Connected.

When entering sleep state with sleep depth determination, increase sensor accuracy, increase detection speed, change detection angle up / down, change detection width, human body detection by human body detection Judgment is made and the detection by the sensor is adjusted so that information is acquired in an optimal state, such as not detecting any part other than that part.
By optimizing the detection speed according to sleep, waste is eliminated, contributing to increased determination speed and power saving.
To obtain information in an optimal state, such as increasing / decreasing the sensor accuracy, increasing the detection speed, changing the vertical angle of the detection direction, changing the detection width (scanning range), specifically, infrared In the case of a sensor, in order to increase the speed of scanning and detection, the temperature detection accuracy is lowered and the speed is increased.
Moreover, at the time of sleep, sleep start determination can be performed automatically or manually by the interface or the sleep depth determination described above. From the result of automatically learning the sensor angle and the scanning range, it is possible to change the control to detect only the position where a person is present, thereby eliminating waste and increasing the detection speed. Even when a camera is used as a sensor, the dynamic range can be narrowed and the frame rate can be increased.
The sleep state can be judged by using an interface such as pressing a button at bedtime, or when various triggers such as human body detection, sleep mode, and human residence for a certain period of time are set as triggers and the trigger is turned on. May be executed. In addition, the detection change is automatically started when the sleep state is determined from the body movement or when the sleep state is moved from the learned result. Of course, the initial setting may be fixed and may be changed every time. It is also possible to set manually or automatically during installation.
Even in the case of judging a plurality of persons, the detection angle can be changed, so that an optimum angle can be designated.
Each of the above methods can be used for other purposes such as personal authentication in the living room besides sleeping.
In many cases, the sleeping state is grasped in the bedroom, and if the sensor is set for the bedroom, the detection can be automatically and optimally changed immediately after the human body is detected.
The position detected by the human body detection sensor may change for some reason. For example, when detecting by scanning an infrared sensor, the position changes depending on the accuracy of stepping. In addition, when data is acquired by scanning in a certain loop such as reciprocation, the same deviation may occur in the detection data every time. In addition, correction of data may be necessary due to deterioration in accuracy and changes in the environment to be detected. For example, when using a difference between results detected every time, the difference may be output greatly even though the detection target has not changed.
In the embodiment of the present invention, the correction is applied to items that can be measured in advance and the difference can be predicted, such as the accuracy of the stepping motor. The timing for applying the correction may be a timing, or a mechanism may be used in which a plurality of data is compared at a specified timing, and the correction timing can be automatically determined such that an average error such as a difference always exceeds a predetermined threshold. . The amount and value to be corrected can be automatically calculated. Any method may be used, such as a difference other than the data detection target, a difference when the same location is obtained by pattern matching, and an error from a reference mark.
Also, in the case of a method of detecting a thermal pixel image by reciprocating scanning to the left and right using an infrared sensor, there may be a shift in the left and right. The detection method of the deviation (error) can be performed by the number of pixels having a certain temperature error or more, or the average temperature error of each pixel of the thermal pixel image. However, when detecting the movement of a person in the living room, etc., if the radiation temperature of the wall and floor is constant (uniform), the difference between the thermal pixel images scanned back and forth by the sensor is smaller than the average temperature difference except for the human body detection part. Even when judging by the number of pixels, since there are few pixels having a temperature difference, it is difficult to set a threshold as a whole.
Therefore, when the shapes of the objects detected at the same position and in the same cycle are the same, or when it is determined that they are approximated by pattern matching, the difference in the number of pixels is taken with reference to the object, and correction is performed. In this case, a threshold value may be provided based on the number of objects or the number of pixels used as a reference, or correction may be performed only when the same position and cycle are detected several times. The pattern matching is used as a generic name of algorithms that can determine that a certain pixel of a certain thermal pixel image is the same as another thermal pixel image, and therefore, any of template matching, texture matching, and HMM (Hidden Markov Model) may be used.
A function that includes a non-contact sensor that detects a human body, compares detection data detected by the non-contact sensor with a predetermined value, and a function that corrects the detection data based on the comparison result or a predetermined reference value. The following are included. That is, in the case of a sensor that obtains a value by reciprocating the detection range (for example, left and right reciprocation), even if information is obtained by reciprocating scanning at the same place, a deviation ( If there is an error, in addition to the method described above, only information on one-way round trip (for example, sensor output scanned from left to right) is used, the information is compared in time series, and the difference is calculated. Also included is a method of avoiding round trip errors and outputting a difference.
Further, if it is desired to increase the acquisition speed of information from the sensor, the difference is similarly obtained using the other information of the reciprocating scan (for example, the output of the sensor scanned from the right to the left). In this way, information obtained by scanning the sensor from left to right and comparing them in time series and a differential output obtained by performing the same processing by scanning the sensor from right to left are alternately output. Thereby, a differential output with little error can be acquired efficiently.
Further explanation will be given by taking the first two reciprocations of the sensor as an example. In the case of extracting the image difference obtained by scanning in the left-right direction in the thermal pixel image by the infrared sensor, the human body is erroneously detected if there is a deviation. In such a case, the following is recommended. That is, even if a deviation occurs between the thermal pixel image 1 scanned in the left direction and the thermal pixel image 2 scanned in the right direction, the thermal pixel image 3 scanned in the left direction again and the thermal pixel image 4 scanned in the right direction again. The thermal pixel difference image 1 obtained by comparing the thermal pixel image 1 and the thermal pixel image 3 and the thermal pixel difference image 2 obtained by comparing the thermal pixel image 2 and the thermal pixel image 4 are output. And the sensor output with few errors can be utilized by using the thermal pixel difference image 1 and the thermal pixel difference image 2 as a continuous difference image.

Next, air conditioning control for realizing comfortable sleep will be described based on an operation flowchart of the air conditioner 10 shown in FIG. First, the temperature detection device 1 is operated in a predetermined cycle, and a temperature distribution in a predetermined region is represented as a thermal pixel image for each cycle and stored in the storage device 3 (S1). The body motion analysis unit 4 detects the presence of a person based on the shape of the temperature distribution of the acquired thermal pixel image and the difference between the thermal pixel images at different times (S2). For example, it is determined that there is a person in a temperature zone portion close to human body radiant heat or a temperature zone portion where motion is detected.
When the body motion analysis unit 4 detects a person, based on the thermal pixel image, the temporal position change of the temperature zone corresponding to the person, the pixel shape of the temperature zone, the surface temperature of the temperature zone, etc. The torso and limbs are specified (S3). Furthermore, the body motion analysis unit 4 determines the presence or absence of body motion in each part based on the acquired position and temperature change of the corresponding pixel portions of the human head, torso, and limbs (S4). When body movement is not detected, the time of the stationary state is stored in the storage device 3 (S5).
When body movement is detected in S4, body movement analysis unit 4 analyzes body movement based on the position and temperature change of each pixel portion of the human head, torso, and limbs. The size of the body motion, the number of times, and the surface temperature are calculated for each part (S6). Subsequently, the body motion analysis unit determines whether there are a plurality of persons (S7). If there are a plurality of persons, a separation process of those persons is performed (S8). Details of the person separation process will be described later. After performing human separation processing, or when multiple people are not detected in S7, the sleep state determination unit 5 calculates the size and number of body movements of the calculated human head, torso, and limbs. If necessary, the surface temperature is also taken into consideration to determine the sleep state (wake-up, REM sleep, sleep depth 1 to sleep depth 4) as shown in FIG. 3 (S9).
And the air-conditioning control part 6 controls the air-conditioning operation part 7 which is an air-conditioner main body based on the sleep state obtained by the sleep state determination part 5, and performs air-conditioning control (S10).
In the above, the relationship between the sleep state and air conditioning control for the detected person is stored in the storage device 3 by the learning function provided in the air conditioner 10 (S11).

When a plurality of people are detected at the same time, each person is basically detected from the shape of the pixel aggregate of the thermal pixel image. When people are close together such as a double bed and pixel groups of people overlap, identify each part of the body (head, torso, limbs), and separate multiple people based on that to accurately identify each person To detect. For example, when each part of a person who has entered the room first is detected, the orientation of the body and the shape of the person who has entered the room first can be stored, so that separation is easy. In addition, when the surface temperature of an overlapping person falls, each person can be detected separately based on the surface temperature.
In addition, when a plurality of people are detected at the same time, their faces are detected when they become stationary. In addition, body movements of limbs are also stored. Depending on the angle of the temperature detection device 1, there is a possibility that the person is the same person, but the ratio of how much overlap is obtained from the number of movable parts and the detected value of the number of persons at the time of entering the room. For example, when there are two places that are considered to be the right hand, the face can be divided into two according to the ratio of the distance from the left hand, and can be separated into a plurality of persons.
In the case of a configuration in which air conditioning control and sleep state determination are started with the selection button, there are cases where people in the futon are recognized by overlapping as viewed from the sensor. In that case, personal information such as the number of people, age, gender, and biological information is input to the air conditioner as an initial setting, and the determination is made on the assumption that people are overlapping. Alternatively, the biological information acquisition means is attached to the remote control and held before going to sleep, and the terminal temperature, blood pressure, pulse, body fat, etc. are acquired on the air conditioner side. By doing so, it can be determined whether it is a sleep state or not, and how many people are present. In addition, a distance sensor or the like is used in combination to detect a plurality of people using the temperature detection device 1 and the distance sensor, or another sensor such as a night vision camera is used to separate a plurality of people and detect each person. You can also.

Next, specific sleep states and examples of air conditioning control will be described. For example, if the magnitude and number of body movements decrease by a certain percentage from the number of body movements that occur when the person is still awake (in the same place) while waking up, it is determined that the person has fallen asleep. In this case, body movement may be analyzed for 3 to 20 minutes, and for example, when the body movement changes from 2 to 3 times in 1 minute or less, it may be regarded as falling asleep. Then, the same analysis is repeated about 2 to 10 times thereafter, and when a change in body motion occurs that is as large as when waking up, it is determined that it is awake and sleep is determined again. Since the end temperatures of the head, torso, limbs, etc. can be distinguished, the forehead temperature rises slightly and then begins to draw a descending line. Air-conditioning control is performed in the direction of lowering.
In addition, at the time of falling asleep, heat is dissipated at the end of the person, so the temperature rises. Therefore, when the surface temperature of the limbs rises, it may be determined as a sleep operation. Since the temperature of the face surface tends to decrease, the difference from the temperature at the time of entering the room is taken, and if the temperature of the face surface decreases, it can be determined to fall asleep.
When the radiation temperature of a person is low, it is preferable to control to raise the room temperature once, circulate the blood flow of the limbs smoothly, and then lower the room temperature. In addition, the air-conditioner is air-conditioned while limiting the lowering width so as not to hinder heat dissipation due to excessive reduction or judging the skin surface temperature.
In addition, since it is weak to irritation | stimulations, such as a wind at the time of falling asleep, it is preferable to make the display part of the air conditioning apparatus 10 gradually darken, to suppress noise, to make wind direction control, and to make it gentle to humans.

The prediction determination of the sleep state can be performed using the size and number of body movements. For example, a time zone in which there is almost no body movement (the number of body movements is, for example, 1 to 2 times / 4 to 8 minutes) is determined as the sleep depth 4. Since large body movements occur after sleep depths 3 and 4 and REM sleep, it is determined that the portion is a change in sleep depth. In the state of sleep depth 2, sleep depth 1 and wake-up are scattered and body movements are performed. The sleep state can be determined based on these characteristics. Regarding body movement, if the torso moves or the temperature of the face changes greatly, it is determined that the body is turned over, and if the limb moves, the amount of change is calculated as the body movement that is not turned over. . In addition, it is preferable to determine the sleep state by comprehensively judging all body movements of the human head, torso, and limbs including the size, number of times, and time.
As described above, the sleep depth 4 is determined in a state in which there is almost no body movement after falling asleep, and the end is determined by a large body movement and a plurality of minute movements within a certain time. By using the learning function described above, it is possible to grasp the average sleep depth 4 and predict the end. Further, the control that makes the sleep depth 4 the longest due to the temperature change can be stored in the same manner as at the time of falling asleep or at other sleep depths or at the time of REM sleep, and can be used for the control. The correlation between the number of body movements and the setting method of the air conditioning control may be stored, and the optimal air conditioning control may be automatically applied from the next time.
There are individual differences in body movements during REM sleep, and it may be indistinguishable from periodic movements of minute movements or body movements at sleep depth 2 in which sleep depth 1 is scattered. For this reason, the sleep cycle, sleep time, and the number of minute movements are compared with body movements at other sleep depths. The learning function is very effective because the person who has a lot of sampling data can easily determine the sleep state including REM sleep.

As described above, in the air conditioning control, when the sleep operation is determined, the temperature is lowered according to the ambient temperature of the person. For example, when the sleep state approaches the sleep depth 4 and the sleep becomes deep, heat radiation is performed, so the air conditioning temperature is lowered. Thereafter, the air conditioning temperature is raised while giving fluctuations to the temperature as shown in FIG. 3, taking into account the amount of sweat and the deep body temperature. That is, the air conditioning temperature is lowered when the sleep depth is deep, and the air conditioning temperature is raised when approaching REM sleep.
Overall, the temperature is gradually set closer to the comfortable temperature when getting up and the natural temperature obtained with the outside air thermometer, with fluctuations according to the sleep depth. If the surface temperature change of a person is small, the control may be performed without causing fluctuation.
In addition, when the deep body temperature changes depending on the sleep depth, there are individual differences such as the sweating and the period of heat dissipation from the limbs being delayed, so it is preferable to control with a phase according to the individual.
When waking up, the end of REM sleep is detected based on the wake-up time and the number of sleep cycles set by the timer, and air conditioning is controlled so that the body temperature rises after REM sleep at waking up. Moreover, you may assist the wake-up by air-conditioning control with a display part, illumination, an alarm, etc.
In the case of setting the wake-up time, the REM sleep cycle may not fit well around the set time. In this case, when entering REM sleep, air-conditioning control is performed so that the sleep depth 1 after the sleep depth 4 and the wake-up time are raised in this order. If any of these timings does not appear near the set-up time, the body temperature is raised, and an auxiliary means such as an alarm or lighting is used to get up.
At sleep depth 4, there are few things that happen due to changes in the thermal environment, so when the sleep depth becomes deeper, air conditioning can be stopped to save energy. In addition, when the sleep depth is deep, it is strong against irritation, so if the room temperature is close to the value that disturbs the sleep state, it operates comfortably without using air conditioning by controlling the wind direction and wind power, and without using cooling. Can create an environment.
In the air conditioning control of the first embodiment as described above, the energy saving can be improved by controlling the air conditioning only in the area where the person is present by detecting the person in the predetermined area.

The temperature change of the human head, torso, and limbs is judged when the slope of the surface temperature at each part changes toward the mountains and valleys and begins to incline, and the sleep situation is predicted based on that. Perform air conditioning control. The prediction method is based on the temperature change in FIG. 3, and individual variations are determined by learning.
Although it is common to the learning of air conditioning control using body movement, some data may be prone to appear as noise, such as staying up late at night, drunk, late-night wake-up. However, learning is performed so that data far from a predetermined average value is judged to be noise. Also, when the futon kicks the skin and changes in body movement or body temperature are recognized, it is determined that the body movement is large, and appropriate control such as prevention of overcooling is performed. In addition, you may make it make these learning functions act, without performing control with respect to the air-conditioning operation part 7. FIG. Thereby, even in the season when air conditioning is not used, the sleep state of a person can be learned, and the state of sleep depth and the sleep index can be displayed.

In the case of Embodiment 1, since the head, torso, and limbs can be identified, the surface temperature of each part is calculated to determine the exposure amount of the limbs. Can be judged as winter. Similarly, the pixel aggregate in a portion that can be regarded as a person is in a static state, and the amount of futon can be estimated from the temperature change of each part of the thermal pixel image. When the ambient temperature is high and low, the amount of futon can be detected more accurately by taking into account changes in skin temperature and clothing temperature.
At the same time, the amount of clothing can be determined by taking the temperature difference between the skin temperature at which the limbs are exposed and the nearby clothing part. If the skin part cannot be determined, the part considered to be the skin temperature may be averaged, and the amount of clothing may be determined from the temperature difference between it and the center of the torso. If the ambient temperature is obtained in advance and there is a temperature difference from a futon or the like, it becomes easier to distinguish between the clothing temperature and the skin temperature.
Also, it is determined that the surroundings are cold in situations such as when the limbs are hidden or there is little turn over, and the opposite situation is determined as being hot. Since this is also related to the difficulty in sleeping due to humidity or the like, the threshold value for the determination of “cold / hot” can be changed according to the situation.
In the case of late-night heat storage or whole building air conditioning, it is determined from the cooling / heating settings and the average radiation temperature. Since it converges in the late 20 ° C in summer and in the late 10 ° C to 20 ° C in winter, the judgment in this case is easy.

The calendar and time can be obtained to a certain degree from the acquired season information, air-conditioner off-season, room temperature, and sleep time. This information can be added as time concept and season information at the time of learning described above, and can be used for display as described later or used as communication information. In order to make these pieces of information more accurate, the date and time may be initialized, a radio clock may be used, or it may be used if it can be obtained through a network.
In addition, if there is a separate sensor, it can also be determined from there. For example, when the brightness can be determined, such as a camera or illuminance sensor, it is possible to determine daylight, direct light from lighting, indirect light, and to determine the bedtime and time.
The trigger for going to bed may be determined not by the temperature detection device 1 but by a human body detection sensor such as another pyroelectric sensor or camera. Alternatively, when the illuminance change is equal to or greater than a certain illuminance that is a camera or illuminance sensor, the sleep state determination and the air conditioning control may be started after that.

When a person is periodically detected at the same place, and when a person falls asleep at that place and spends all night, it can be determined that the person is lying at the place periodically. As a result, the head, torso, and limbs of the body that moves can be recognized more accurately. In addition, when the period of arrival at the place is constant, the air conditioning control can be started near the corresponding time, and until then, the energy saving state can be maintained in the standby state. Moreover, when starting the temperature detection apparatus 1, the form of starting at high speed as mentioned above etc. can be taken. Furthermore, since sleeping at the same position every day at bedtime, individual identification is possible based on body size, temperature change, body movement cycle, and the like.
In addition, when a distance sensor is provided, it is possible to recognize an individual more accurately by body size according to the distance, and when a camera is provided, face authentication or the like.
Even when only the information of the thermal pixel image 8 can be obtained, the gender, age, and body size can be estimated from the shape and size of the human-corresponding portion, the length of the limbs, the surface temperature of the skin portion, and the like.
Although the human skin temperature depends on the surrounding environment, the difference between the temperature calculated based on the detected value of the temperature detection device 1 and how much the temperature is calculated was calculated from the database related to the temperature detection device 1. Determine based on distance. In addition, since the size of the head is fixed to some extent, the ratio with the torso part is easy to understand.

In the first embodiment, it is assumed that a human body exists 3 to 6 m from the air conditioner, the pixel resolution of the infrared sensor is about 5 to 30 cm at 6 m, and the temperature resolution at 6 m is about 0.5 ° C. In the first embodiment, there are thermal environments such as individual differences and room temperature, but the detected temperature varies from 0.1 ° C. to 0.5 ° C., and varies by about 0.1 to 3.0 ° C. due to changes in sleep depth. When there is a body movement, the corresponding pixel aggregate (or temperature zone) moves, so there is a temperature difference from the space, and there is a slight temperature change. For this reason, it is easier to identify body movement. Therefore, it is preferable to determine the sleep state by weighting in order of body movement, face temperature, and limb temperature. Depending on the distance between the non-contact type temperature detection device 1 and a person, the pixel resolution, and the performance of the temperature resolution, it is possible to change the priority order, and it is possible to make a specification that cannot be determined unless a plurality of conditions are met.
If a person's skin temperature at the time of entering a room is recorded, a difference from that at the time of sleep is learned, and a threshold for sleep determination is provided, an auxiliary role can be expected when body movement cannot be detected by a futon or the like. For example, when the distance between the non-contact type temperature detection device 1 and a person is short, even if the body movement is unknown, the change in the temperature of the face surface can be known in detail, so that the sleep state can be estimated to some extent.

Here, the sensor that can be used in the first embodiment will be further described. A case where the already described infrared sensor is replaced with another sensor and a mode in which both the already described infrared sensor and the other sensor are provided are possible. In the case of another sensor, for example, a pyroelectric sensor, locations detected by a plurality of pyroelectric sensors are stored as image blocks in the same manner as in the case of an infrared sensor, and a part with a large number of body movements is a movable part (head, limbs) ), And a portion with a small number of body movements can be used as a fixed portion (body portion). If information including the torso is stored when turning over, recognition from the next time is easy. Even when a camera is used, it is possible to easily and inexpensively grasp body movements by the above-described respective part identification methods without using an advanced and complicated algorithm for identifying each part of a person by pattern matching or the like.
As described above, when the information acquisition of the infrared sensor is performed at a high speed, the amount of activity can be grasped in more detail, but the amount of activity can be known only by the frame speed in the case of a camera. The date and time, the amount of clothes, body movement, and multiple people can be judged more accurately by using other sensors together.
Sensors that can be used in the first embodiment include a distance sensor, a temperature / humidity, an atmospheric pressure, an illuminance sensor, a biological information acquisition sensor, and the like. By first sensing the room with an infrared sensor at first, and looking at the window position and the fluctuation of radiant heat from the temperature change that does not correspond to the human body, it can be used to obtain a more comfortable sleep state.

Moreover, it is also possible to install functions that are not directly related to air conditioning, such as sensors and display units such as the temperature detection device 1, separately from the air conditioning operation unit 7 (air conditioner main body). For example, by separating the temperature detection device 1 and installing it near the human body such as a ceiling, a bedside, or a tabletop, more accurate temperature detection becomes possible.
In the air conditioner 10, the learned content changes depending on the season, time, etc. as described above. In addition, the room needs to be changed by changing the room, growing people, changing the family structure, and increasing or decreasing the bedding. If there is a significant change in the information acquired by the sensor, refresh the learned content and return it to the initial state, or use the previous data until the information acquired from the sensor is satisfactory To do. The case where there is a significant change in information acquired by a sensor is a change in acquired information such as a change in the detection position of a person.
Also, if you take a bath when you wake up and come to the bedroom, your body temperature is higher than usual. For this reason, it is determined after exercise or after bathing as compared to the normal body temperature while waking up, and air conditioning control is performed taking this into consideration. In such a case in summer, there is a desire to cool the body that has been rapidly lit. In that case, after the cooling is pressed, the room temperature, season, and skin temperature are calculated, and the cooling is rapidly performed based on the calculated temperature. If the cooling capacity does not rise immediately, control the direction of the wind toward the person and take away the radiant heat of the skin. In this case, the above-described temperature detection can be performed at high speed. In addition, after controlling for a fixed time, a driving | operation is returned to normal so that it may not cool too much. The operation may be returned to normal by judging the skin temperature with a sensor.

If you roll over on a hot summer day, you can see the temperature change even where you were. This is because the hidden part where the pressure is applied to the futon and the temperature is rising is detected by turning over. Even in winter, if there are electric blankets or hot water bottles, a heat source near the human body temperature is detected when they are exposed. With respect to turning over, the temperature gradually decreases after a certain period of time, so that part is excluded to detect a person. Also, since the back part is hot and the temperature is maintained to some extent, if the pixel aggregate of the part where the person is detected becomes large, the person with the least temperature change is judged as a person and the other is recognized as a futon. is there. Electrical products such as hot water bottles do not change at a constant temperature, and their positions are fixed to some extent. Even if body movement is large, the temperature change of electrical appliances is small or the number of temperature changes is small, so if there is an abnormality in body movement or temperature change associated with the sleep cycle in winter, it is judged that it is caused by heating equipment can do.
It is also possible to determine whether or not a person is staying on the basis of the difference between the temperature at the detection position when the person is detected and the current temperature at that position. This is because the temperature change is small when a person is staying, and the temperature change is large otherwise.

If the threshold is increased so that only large body movements are judged and the sleep depth is judged only with large body movements, the number of times that deep sleep is judged is greater than methods that can be measured normally, such as brain waves, and actual sleep An error occurs in the determination of the depth and the sleep depth of the present invention. Conversely, if the threshold is reduced and the sleep depth can be determined from small body movements, noise will be picked up, so the number of times that it is determined to be shallow sleep will increase compared to methods that can be measured normally, such as brain waves, and this method is also actually An error occurs in the determination of the sleep depth and the sleep depth of the present invention.
Therefore, a plurality of thresholds are provided according to the size of body movement, and when large body movement is detected, it is determined that a certain time after the body movement is shallow sleep. Then, the sleep depth is determined by taking “and” when a small body motion is detected within a specific time after a large body motion.
If there is a lot of noise, it is judged as abnormal from the learning function, and instead of taking “and”, “or” is applied and individual judgments are made. When it is judged that there is a lot of noise, there are individual differences, but the initial value is when deep sleep is 10% or less, or 50% or more.
FIG. 6 is a flowchart showing a deep sleep determination operation example by the sleep state determination unit according to the first embodiment related to the above aspect. Here, it is assumed that the temperature difference of the thermal pixel which is a human body detection means detected by the body motion analysis unit 4 is used. Therefore, first, a human body is detected by the body motion analysis unit 4 (S21). Next, a large body motion is detected, and deep sleep (for example, sleep depth 3 or 4) is determined based on the detection (S22). Specifically, a temperature change difference of 1 ° C. or more is detected as a large body motion, and the number of thermal pixel blocks is calculated. And if the calculated thermal pixel block is a predetermined number (for example, 2) or less, it will determine with deep sleep. Subsequently, small body movement is detected, and deep sleep (for example, sleep depth 3 or 4) is determined based on the detection (S23). Specifically, a temperature change difference of less than a threshold value of 0.5 to 1 ° C. is detected as a small body motion, and the number of thermal pixel blocks is calculated. And if the calculated thermal pixel block is a predetermined number (for example, 5) or less, it will determine with deep sleep. Thereafter, “and” or “or” in the determinations of S22 and S23 is taken (S24), and it is determined whether or not deep sleep (S25). If deep sleep is determined, the result is output (S26). Otherwise, the process returns to the first step S21.
If the detected thermal pixel block is far from the assumed location and is small, it may be removed as noise. In order not to include noise, the human body may be detected only in a region where the human body is detected last (a region where a human is determined to be staying) or only in a region determined by learning.
In addition, the selection of “and” or “or” in S24 may be selected according to the sleep cycle or the ratio of deep sleep. Further, since the degree of skin exposure is high in summer, the accuracy of temperature detection increases, so “and” may be selected during cooling in summer and “or” may be selected during heating in winter.
Furthermore, in the case of a sensor with a fast response speed or a sensor whose detection threshold cannot be divided into two, the determination can be made based on the number of blocks related to body movement or the number of continuous detections (the number of detections within a certain time).
The threshold value should be changed according to the season. When body movements are acquired using an infrared sensor, there is a possibility that only a small change will occur in the futon in the winter, even with larger body movements. In that case, the threshold value is changed to a smaller value. However, since it is confused with noise, it is also effective to learn the place where the human body was detected and stayed last and to lower the threshold value. On the contrary, in summer, the futon area is small, so the threshold of body movement is increased.

  The air conditioning control of the first embodiment basically controls the temperature, but in addition, according to the function provided in the air conditioner 10, the wind direction, wind speed, air purification, dehumidification / humidification, ventilation, target set temperature, You can also control settings such as the perceived temperature target temperature, negative ions, supplements, odor control, oxygen control, atmospheric pressure setting, display / lighting, sound / buzzer, etc. In addition, it is preferable to correct | amend the threshold value used for those settings according to the detected season, time, a person, room temperature, the amount of clothes, a futon.

Embodiment 2. FIG.
The first embodiment relates to air conditioning control that enables comfortable sleep, and assumes an air conditioner that is applied to a bedroom. However, even in rooms other than the bedroom, such as the living room, it is possible to recognize the individual, that is, the user, by detecting the human head, torso and limbs and determining the size, number of times, and surface temperature of those movements. Therefore, it is possible to perform comfortable air conditioning control according to the user.
For example, in a living room, the number of people in the living room generally increases and the movements vary, but in the case of a general household, people are often located in a certain place (for example, a place with a sofa). Therefore, the size and number of body movements of each part of the person who came to the place, the surface temperature, the date and time, and the air conditioning setting at that time are stored. The next time, the number of body movements of the person who came to the position, the size, the surface temperature, the date and time are compared and personal identification is performed, and the stored air conditioning control corresponding to the identified individual is automatically performed. Select and apply. Furthermore, the other functions described in the first embodiment can be applied to the living room as they are. Thereby, for example, even when a person falls asleep in the living room, comfort can be ensured by avoiding over-cooling and over-warming, and energy-saving operation can be performed.

Moreover, the degree of relaxation can be determined from the body movement of each part of the person, and comfortable control can be performed. For example, when the body movements of the limbs increase, the degree of relaxation is considered low. In that case, by performing the air-conditioning control by shortening the detection cycle by the temperature detection device 1, linear air-conditioning control can be performed, and the uncomfortable time can be reduced.
In addition, it is judged whether sitting or lying down, and abnormal continuous body movements are detected and converted as discomfort index points. In the conversion of the body movement of the foot when sitting, it is lighter than the weight when lying down in the conversion. When lying down, it is lighter than standing, but heavier than sitting. The body movement at the discomfort index point is more frequent than the body movement during sleep and is clearly different. Furthermore, the surface temperature change of the limbs is also captured and compared with the temperature of the torso and face, and if the difference is large, it is determined that the temperature is too cold and too warm and is used for air conditioning control.

  The air conditioning control according to the first and second embodiments may be switched by a main body switch, a program, or an initial setting at the time of installation according to the use such as bedroom, living room, Japanese-style room, dining, etc. Good. Moreover, you may grasp | ascertain a sleep state from the information acquired with the infrared sensor, and may judge it automatically. Further, as described above, the determination may be made based on information obtained from other sensors. When the air conditioner is installed in both the living room and the bedroom, more comfortable control is possible by controlling the sleep control temperature in the bedroom based on the difference between the body temperature when waking up in the living room by providing communication functions. is there. If the temperature in the bedroom is higher than in the living room, the body temperature rises when you enter the bedroom. In that case, the temperature of the limbs is also rising, so there is no need to warm the limbs during sleep. In addition, the degree of relaxation learned in the living room can be automatically set to the optimum temperature in the bedroom from the personal settings. When reading a book in the bedroom, the surrounding environment such as the area is acquired by infrared rays, and the surrounding environment is taken into account. It can be controlled in the same way as the living room.

  Here, the learning function of the air conditioner will be further described. As described above, the air conditioner stores at least the correlation between the ambient temperature and the body movement, controls the body movement and the surface temperature of each part to be optimum, and learns and stores them. However, there are individual differences and the sleep depth cycle may remain disturbed. In that case, adjust the air conditioning temperature in 0.5 ° C increments, and if the sleep depth cycle worsens in the same time zone and the same ambient temperature, adjust the air conditioning temperature in 0.5 ° C increments in the opposite direction. Learn and remember it. In addition, by automatically adjusting the temperature of the air conditioning control, it is possible to automatically search for an optimal sleep thermal environment. If the learned information is transmitted to the server through the network as shown in FIG. 5 and the optimum air conditioning control information is sent from the server, the capacity of the storage device 3 of the air conditioner is increased. Even if you do not use it, you can download the latest optimal control settings sequentially.

If the air conditioner's information (pre-set information, information obtained by detection, judgment and analysis) can be sent to a personal computer or server via a network, etc. Health check can be done by checking. Of course, it is also possible to display with a display panel or to view on a personal computer via a portable memory.
If you can send information about people in the room by sending e-mail to the specified e-mail address or making a phone call via the network, you can check the safety of the elderly and children in remote areas and intrude It can also be used for home security, such as recording / browsing via a network and recording on a recorder, and it is also possible to develop sleep improvement proposals for hospitals and hotels as a business.
Furthermore, date and time information, regional information, latest control information, and the like can be acquired via the network, and information for more optimal sleep control can be obtained. Also, these functions can be automatically given to a server connected to the network.
As described above, it is also possible to use the air conditioner 10 without using air conditioning. In this case, the air conditioner 10 can be used for the personal recognition device and the sleep state determination.

When a human body is detected in a certain area, there is a method of comparing with the next time using the image, but if the heat source does not move, it is necessary to determine absence and to determine the presence / absence of a person . Therefore, the human detection means may have the following functions. When the human body moves from the range where the human body is detected, if there is no significant temperature difference due to changes in ambient temperature, etc. during the time of movement, and the absence cannot be determined, the moment when the human body is detected By taking the current data and the current difference, the difference from the human body temperature becomes clear, and human body detection can be continued. This makes it possible to accurately determine absence.
Similarly, when a human body is detected, a thermal pixel before a human body is detected only in the human body detection range (thermal pixel) is overwritten as a background image. Also, the background temperature thermal pixels outside the human body detection range are overwritten as they are. As a result, when the time advances to the next cycle (frame), the difference between the background image and the heat source image is generated even when the person does not move, so that the presence / absence determination is less likely to be mistaken. Thus, it is possible to easily determine how long a person is in the area without performing absence determination.
Also, analyze the temperature change of the data up to the present with the data (thermal image) at the moment when the human body was detected, and if the temperature is changing as well as around the area where the human body was detected, determine that it is absent Also good.

When the background temperature other than the area where the human body is detected is updated, when the human body is newly moved, the threshold is determined from the average ambient temperature of the previous thermal pixel, and the accuracy of the detection result is increased. Even in the human body detection area, the average ambient temperature taking into account the ambient temperature and the last background temperature at the human body detection position is calculated and the threshold value is optimally set, so that the detection error can be reduced.
In addition, when a human body stays at the same position and the surrounding background temperature fluctuates, when a human body is detected, the thermal pixel of the previous cycle in which the human body was detected is displayed in the human body detection range (thermal pixel). Since the image is overwritten, the ambient temperature changes before the person moves, and there is a possibility that the absence determination is erroneous, for example, when the background image before human body detection is the same as the current background temperature. On the other hand, when the average value of the ambient temperature rises while the person is staying, it is preferable to correct the thermal pixel of the portion where the person is staying according to the ambient environment temperature. For example, if the background temperature before a person moves is 10 ° C. and a person at 32 ° C. moves, then the ambient temperature rises to 20 ° C. If the person moves, 20 ° C. and 10 ° C. The difference is detected at 10 ° C. Therefore, based on the thermal pixels around the human body detection area or the human body, temperature correction is added to the human body staying background temperature thermal pixel portion. As a result, even if a person moves, it can be accurately determined whether the person is staying without being detected erroneously.

There is also a method of acquiring minute body movement by changing a threshold value of human body determination in an area where a human body is detected. In addition, the threshold value can be set in multiple stages in order to distinguish the head, torso, and limbs. Conversely, since the stay determination is performed at the background temperature, the threshold can be set to a value that allows detection of only large body movements.
In the background, noise is generated due to a temperature change. Therefore, when the ambient temperature is close to the human body temperature, erroneous detection frequently occurs at a threshold that is too small. Further, when it is desired to acquire a small movement such as a movement of only the hand, it cannot be detected if the threshold value is large. However, by changing the threshold value only in the area where the human body is detected separately from the change in the background, it is possible to detect minute movements such as when only the hand is moved so as not to generate noise in other parts.
By doing so, minute body movements can be acquired more accurately, so that the amount of activity during sleep or living can be calculated accurately.
Even when the ambient temperature of the staying part changes, the threshold of the part detected by the human body is lowered so that the minute body movement can be acquired even when the ambient temperature of the staying part changes. Therefore, the same effect can be obtained by the presence / absence determination. In addition, the threshold corresponding to the temperature correction is determined by simultaneously performing both the background temperature correction and the threshold change of the human body detection area. If the background temperature approaches or is close to the human body temperature, the human body surface temperature may approach the background temperature before the human body is detected due to the influence of the ambient temperature. By setting a threshold value so that even minute body movements can be acquired, accurate human body detection can be performed.

Since it is possible to discriminate by a change in ambient temperature or a clock function during sleep, it is possible to change a threshold other than the human body detection area in the same manner, and the same effect can be obtained.
In addition to areas where human bodies are detected, areas where human bodies are detected within a certain period of time, and areas where there is a large amount of activity during learning can be processed by changing the threshold value in the same manner.

  This technology can be applied not only to the human body but also to others. The engine area and the tire area of the car become hot, but the other parts are not so hot. It is also possible to accurately detect the contour of the entire vehicle by using a multi-stage threshold for the detected part. If the threshold value is changed according to the detected temperature change of the engine portion, minute engine temperature changes and other noises can be detected separately.

  DESCRIPTION OF SYMBOLS 1 Temperature detection apparatus (infrared sensor), 2 Image display part, 3 Storage device, 4 Body movement analysis part, 5 Sleep state determination part, 6 Air-conditioning control part, 7 Air-conditioning operation part (air-conditioner main body), 8 Thermal pixel image, DESCRIPTION OF SYMBOLS 10 Air conditioner (indoor unit), 11 Room temperature detection sensor, 12 Wind direction adjustment tool.

Claims (17)

A human detection means for detecting a person without contact;
A body motion analysis unit that identifies the detected human head, torso, and limbs, and analyzes body motion and surface temperature, which are movements of at least one identified site;
An air conditioner characterized in that the state of a person is identified based on the analysis result of the body motion analysis unit and air conditioning control is performed according to the state. As the human detection means, an infrared sensor, and an image display unit representing a temperature distribution detected by the infrared sensor as a thermal pixel image,
The body motion analysis unit performs the analysis based on a shape of a specific temperature zone of a thermal pixel image representing a temperature distribution detected at different times by the infrared sensor and a change in position of the temperature zone. The air conditioner according to claim 1. The analysis of the body motion analysis unit includes detection of the size and number of times of the body motion,
A sleep state determination unit that predicts and predicts a person's sleep state based on the size and number of the body movements analyzed by the body movement analysis unit, and the surface temperature;
An air conditioning control unit that performs air conditioning control based on the determination of the sleep state determination unit;
The air conditioner according to claim 1 or 2, further comprising:   When the sleep state determination unit determines that a person is asleep, the air-conditioning control unit gradually decreases the air-conditioning temperature as the sleep depth increases, and then causes the temperature to fluctuate according to changes in the sleep depth. The air conditioner according to claim 3, wherein the air conditioner is controlled so as to gradually approach a temperature set at the time of rising or an outside air temperature.   When a person is detected at the same position in the same time zone, the air conditioning control information in that case is patterned and stored, and the air conditioning control is performed based on the patterned information. Item 5. The air conditioner according to any one of Items 1 to 4.   The air conditioner according to any one of claims 1 to 5, wherein the body motion analysis unit determines a person's body posture and whether or not he / she has turned over based on a state of the part.   The air conditioner according to any one of claims 1 to 6, wherein the body motion analysis unit estimates a season and an amount of clothes, and uses the estimated information for air conditioning control.   8. A learning function for learning and storing optimal control information while correlating information of a determination output from the sleep state determination unit, ambient temperature, and air conditioning control. The air conditioner described in the paragraph.   The learning data stored by the learning function is refreshed according to a change in the environment or periodically, and the refreshing is performed by newly acquiring data while retaining past data or by returning to the initial state. The air conditioner according to claim 8, wherein:   The air conditioner according to any one of claims 1 to 9, wherein the human detection means is configured by a sensor that is performed by a method other than temperature detection.   The air according to any one of claims 2 to 9, further comprising a second sensor that assists the infrared sensor, and analyzing a human body movement based on the infrared sensor and the second sensor. Harmony machine.   A plurality of air conditioning control modes corresponding to a room to be used are stored, and a mode selection device for selecting one of the stored air conditioning control modes is provided. The air conditioner according to one item.   The air conditioner according to any one of claims 1 to 12, wherein the air conditioner is connected to a network, and information can be shared, analyzed, and updated via the network.   The human detection means detects a person in a non-contact manner, and can change a threshold value used for determining the sleep depth of the human body in the detected human region or other regions. Any one item is an air conditioner. It has a non-contact sensor that detects the human body,
The human body motion detected by the non-contact sensor is divided into a plurality of body motions including a large body motion and a small body motion, and an individual threshold value related to sleep depth is given to each body motion,
A sleep state determination unit that predicts a sleep state from AND of a threshold value related to the plurality of measured body movements, or or;
The air conditioner according to claim 1, further comprising an air conditioning control unit that performs air conditioning control based on the determination of the sleep state determination unit. It has a non-contact sensor that detects the human body,
The detection result of the non-contact sensor or the setting of the non-contact sensor changes one or more of the detection range, detection speed, and detection accuracy of the non-contact sensor. The air conditioner described. It has a non-contact sensor that detects the human body,
A function of comparing detection data detected by the non-contact sensor with a predetermined value;
The air conditioner according to claim 1, further comprising a function of correcting the detection data based on the comparison result or a predetermined reference value.

Images