Thermal comfort estimation system, thermal comfort estimation method, and thermal comfort estimation device

Abstract

A thermal comfort estimation system (1) includes a thermal comfort estimation device (40), a first environment sensor for sensing the environment of a first space in which a subject is present, and a first imaging device, wherein the thermal comfort estimation device (40) includes: an acquisition unit (411) for acquiring environment information indicating the environment of the first space that is output from the first environment sensor, attribute information indicating attributes of the subject, and moving images in which the subject is imaged by the first imaging device; an estimation unit (42) that estimates comfort information indicating thermal comfort by using a trained machine learning model (421) that takes the acquired environment information, the acquired attribute information, and subject information relating to the subject based on the acquired moving images, as input, to output the thermal comfort of the subject; and an output unit (412) that outputs the estimated comfort information.

Description

Thermal comfort estimation system, thermal comfort estimation method, and thermal comfort estimation device

[0001] The present invention relates to a thermal comfort estimation system, a thermal comfort estimation method, and a thermal comfort estimation device.

[0002] There are known techniques for estimating the thermal comfort of a subject. For example, Patent Literature 1 discloses a technique for estimating the thermal comfort of a subject using image data of the subject and a temperature sensor.

[0003] U.S. Pat. No. 1,037,2990

[0004] However, there is a demand for more accurate estimation of thermal comfort.

[0005] The present invention provides a thermal comfort estimation system that estimates thermal comfort with higher accuracy.

[0006] A thermal comfort estimation system according to one aspect of the present invention comprises a thermal comfort estimation device, a first environmental sensor that senses the environment of a first space in which a subject is present, and a first imaging device. The thermal comfort estimation device comprises an acquisition unit that acquires environmental information indicating the environment of the first space output from the first environmental sensor, attribute information indicating the attributes of the subject, and a video image of the subject captured by the first imaging device, an estimation unit that estimates comfort information indicating the thermal comfort using a trained machine learning model that inputs the acquired environmental information, the acquired attribute information, and subject information related to the subject based on the acquired video, and outputs the thermal comfort of the subject, and an output unit that outputs the estimated comfort information.

[0007] A thermal comfort estimation method according to one aspect of the present invention is a thermal comfort estimation method performed by a thermal comfort estimation system, which includes a thermal comfort estimation device, a first environmental sensor that senses the environment of a first space in which a subject is present, and a first imaging device.The thermal comfort estimation method includes an acquisition step of acquiring environmental information indicating the environment of the first space output from the first environmental sensor, attribute information indicating the attributes of the subject, and a video image of the subject captured by the first imaging device, an estimation step of estimating comfort information indicating the thermal comfort using a trained machine learning model that inputs the acquired environmental information, the acquired attribute information, and subject information related to the subject based on the acquired video, and outputs the thermal comfort of the subject, and an output step of outputting the estimated comfort information.

[0008] A thermal comfort estimation device according to one aspect of the present invention includes an acquisition unit that acquires environmental information indicating the environment of a first space in which a subject is present, output from a first environmental sensor that senses the environment of the first space, attribute information indicating the attributes of the subject, and a video image of the subject captured by a first imaging device; an estimation unit that estimates comfort information indicating the thermal comfort using a trained machine learning model that receives as input the acquired environmental information, the acquired attribute information, and subject information related to the subject based on the acquired video, and outputs the thermal comfort of the subject; and an output unit that outputs the estimated comfort information.

[0009] The thermal comfort estimation system of the present invention can estimate thermal comfort with higher accuracy.

[0010] FIG. 1 is a block diagram showing the functional configuration of a thermal comfort estimation system according to an embodiment. FIG. 2 is a diagram showing a space to which the thermal comfort estimation system according to an embodiment is applied. FIG. 3 is a diagram showing thermal comfort according to an embodiment. FIG. 4 is a diagram showing an example in which a subject is detected according to an embodiment. FIG. 5 is a diagram showing an example in which the skeleton of a subject is estimated according to an embodiment. FIG. 6 is a diagram showing a first correspondence table according to an embodiment. FIG. 7 is a diagram showing a second correspondence table according to an embodiment. FIG. 8 is a flowchart of an operation example 1 according to an embodiment. FIG. 9 is a block diagram showing the functional configuration of a thermal comfort estimation system according to a first modification of an embodiment. FIG. 10 is a diagram showing a space to which the thermal comfort estimation system according to the first modification of an embodiment is applied. FIG. 11 is a block diagram showing the functional configuration of a thermal comfort estimation system according to a second modification of an embodiment. FIG. 12 is a diagram showing a space to which the thermal comfort estimation system according to the second modification of an embodiment is applied.

[0011] Hereinafter, embodiments will be described with reference to the drawings. Note that the embodiments described below are all comprehensive or specific examples. The numerical values, shapes, materials, components, component placement and connection forms, steps, and step order shown in the following embodiments are merely examples and are not intended to limit the present invention. Furthermore, among the components in the following embodiments, components not recited in independent claims will be described as optional components.

[0012] It should be noted that the drawings are schematic diagrams and are not necessarily strict illustrations. In addition, in the drawings, substantially the same components are denoted by the same reference numerals, and overlapping descriptions may be omitted or simplified.

[0013] (Embodiment) [Configuration] First, the configuration of a thermal comfort estimation system 1 according to an embodiment will be described.

[0014] Fig. 1 is a block diagram showing the functional configuration of a thermal comfort estimation system 1 according to this embodiment. Fig. 2 is a diagram showing a space 90 to which the thermal comfort estimation system 1 according to this embodiment is applied.

[0015] The thermal comfort estimation system 1 is a system that estimates comfort information indicating the thermal comfort of a subject T in a space 90 and outputs the estimated comfort information. Thermal comfort is an index that indicates an individual's level of satisfaction with a thermal environment.

[0016] FIG. 3 is a diagram showing thermal comfort according to this embodiment. The comfort information is information indicating which of a plurality of levels of thermal comfort the subject T corresponds to. For example, the comfort information indicates which of seven levels of thermal comfort, −3, −2, −1, 0, +1, +2, and +3, the subject T's thermal comfort level corresponds to. Note that when the thermal comfort value is 0, the subject T feels comfortable; the smaller the thermal comfort value, the colder the subject T feels, and the lower the comfort level; and the larger the thermal comfort value, the hotter the subject T feels, and the lower the comfort level.

[0017] The space 90 is an example of a first space. The space 90 is, for example, an indoor space. The space 90 according to the present embodiment is an office (more specifically, a room in an office), but is not limited thereto. For example, the space 90 may be an educational facility such as an elementary school, a junior high school, a high school, or a university, a public facility such as a community center or a library, or may be used as a store or a commercial facility.

[0018] As shown in FIG. 1 , the thermal comfort estimation system 1 includes an environmental sensor 10 , an imaging device 20 , an information terminal 30 , a thermal comfort estimation device 40 , a display device 50 , and an air conditioning device 60 .

[0019] The environmental sensor 10 is an example of a first environmental sensor, and is a sensor that senses the environment of the space 90 in which the subject T is present. The environmental sensor 10 is a sensor that senses the temperature, humidity, radiant temperature, and wind speed of the space 90. The environmental sensor 10 outputs environmental information indicating the environment of the space 90 to the thermal comfort estimation device 40. The environmental information is information indicating the temperature, humidity, radiant temperature, and wind speed of the space 90. The environmental sensor 10 is placed in the space 90, and in this embodiment, is placed on the ceiling of the space 90, for example.

[0020] The imaging device 20 is an example of a first imaging device, and is a device that captures moving images of a subject T present in the space 90. The imaging device 20 outputs the moving images of the subject T to the thermal comfort estimation device 40. The imaging device 20 is placed in the space 90, and in this embodiment, is placed on the ceiling of the space 90, for example.

[0021] The imaging device 20 according to this embodiment is a visible light camera. That is, the imaging device 20 is a device that is sensitive to the wavelength region of visible light and captures visible light from a subject (e.g., subject T) at a predetermined frame rate to generate moving images. The moving images captured by the imaging device 20 are visible light moving images. Visible light cameras are general-purpose cameras that are inexpensive and can be introduced inexpensively into spaces 90 such as offices. Note that the imaging device 20 is not limited to a visible light camera and may be an infrared camera, for example.

[0022] The information terminal 30 is a terminal carried by the subject T. The information terminal 30 is, for example, a smartphone owned by the subject T, but is not limited to this and may be a mobile terminal such as a tablet terminal or a PDA (Personal Digital Assistant).

[0023] The information terminal 30 has an operation reception unit and a communication unit. The operation reception unit receives an operation from the subject T indicating the attributes of the subject T, more specifically, an operation from the subject T indicating the age, sex, race, height, and weight of the subject T. The communication unit outputs attribute information including the age, sex, race, height, and weight of the subject T indicated by the operation received by the operation reception unit to the thermal comfort estimation device 40. More specifically, the attribute information according to this embodiment is information indicating the age, sex, race, height, and weight of the subject T. Note that the attribute information is associated with an identifier that identifies the subject T carrying the information terminal 30.

[0024] The attribute information may be information indicating the attributes of the subject T. The attributes of the subject T include age, gender, race, hair color, eye color, facial features, nationality, height, and weight. In other words, the attribute information may include at least one of age, gender, race, hair color, eye color, facial features, nationality, height, and weight. Therefore, the attributes of the subject T accepted by the operation accepting unit are not limited to the age, gender, race, height, and weight of the subject T. For example, the operation accepting unit may accept an operation from the subject T indicating at least one of the age, gender, race, hair color, eye color, facial features, nationality, height, and weight of the subject T. In this case, the attribute information is information including at least one of these.

[0025] Note that an information processing device such as an edge computer or a cloud computer may be used instead of the information terminal 30. In this case, attribute information indicating the age, sex, race, height, and weight of the subject T is stored in advance in the information processing device, and the information processing device outputs this attribute information to the thermal comfort estimation device 40.

[0026] The thermal comfort estimation device 40 is a device that estimates and outputs comfort information based on environmental information indicating the environment of the space 90 output from the environmental sensor 10, attribute information output from the information terminal 30, and moving images output from the imaging device 20. The thermal comfort estimation device 40 is, for example, an edge computer installed in the space 90, but is not limited to this and may also be a cloud computer installed in a location remote from the space 90.

[0027] The thermal comfort estimation device 40 has a communication unit 41 , an estimation unit 42 , and a memory unit 43 .

[0028] The communication unit 41 is a communication module (communication circuit) that enables the thermal comfort estimation device 40 to communicate with each of the environmental sensor 10, the imaging device 20, the information terminal 30, the display device 50, and the air conditioning device 60. The communication unit 41 may be a wired communication circuit, but in this case it is a wireless communication circuit for performing wireless communication (more specifically, radio wave communication).

[0029] The communication unit 41 is a processing unit having an acquisition unit 411 and an output unit 412 .

[0030] The acquisition unit 411 acquires the environmental information output from the environmental sensor 10 , the attribute information output from the information terminal 30 , and the moving image output from the imaging device 20 .

[0031] The output unit 412 outputs the comfort information estimated by the estimation unit 42. The output unit 412 outputs the comfort information to each of the display device 50 and the air conditioner 60, for example.

[0032] The estimation unit 42 is a processing unit that estimates comfort information indicating the thermal comfort of the subject T using a trained machine learning model 421. This machine learning model 421 is a trained model that receives as input the environmental information acquired by the acquisition unit 411, the attribute information acquired by the acquisition unit 411, and subject information related to the subject T based on the moving images acquired by the acquisition unit 411, and outputs the thermal comfort of the subject T. The estimation unit 42 has the machine learning model 421.

[0033] Here, the subject information will be described.

[0034] The subject information is information estimated by the estimation unit 42 based on the moving image acquired by the acquisition unit 411, and is information related to the subject T. More specifically, the subject information is information indicating the amount of clothing worn by the subject T and the metabolic rate of the subject T. The procedure by which the estimation unit 42 estimates the subject information will be described below.

[0035] First, the estimation unit 42 detects the subject T depicted in the acquired video based on the acquired video. A known means can be used to detect the subject T. Examples of such known means include Haar-like features HOG (Histograms of Oriented Gradients), and classifiers such as SVM (Support Vector Machines), Decision Trees, Trees, and Deep Convolutional Networks. In this embodiment, the estimation unit 42 detects the subject T depicted in the acquired video using the acquired video and a known means.

[0036] 4 is a diagram showing an example of detection of a subject T according to the present embodiment. As shown in FIG. 4, when the subject T is detected by the estimation unit 42, the subject T is surrounded by a rectangle, for example.

[0037] Next, the estimation unit 42 estimates the skeleton of the detected subject T based on the above detection result. A known means can be used to estimate the skeleton of the subject T. As this known means, Deep Convolutional Networks or the like can be preferably used. In the present embodiment, the estimation unit 42 estimates the skeleton of the subject T using the above detection result and this known means.

[0038] 5 is a diagram showing an example of an estimated skeleton of a subject T according to the present embodiment. In FIG. 5, the skeleton of the subject T estimated by the estimation unit 42 is indicated by circles and lines.

[0039] Next, the estimation unit 42 estimates the clothing worn by the subject T based on the estimated skeleton of the subject T. A known means can be used as the means for estimating the clothing worn by the subject T. As this known means, Deep Convolutional Networks or the like can be preferably used. In this embodiment, this known means is used to estimate the clothing worn by the subject T. Furthermore, the known means used here is trained in advance using a dataset including multiple sets, where clothing and attributes of the clothing are treated as one set. Note that in this set, the attributes of the clothing are annotated to the clothing. The estimation unit 42 estimates the clothing worn by the subject T using another machine learning model that inputs the estimated skeleton of the subject T and outputs the clothing worn by the subject T.

[0040] For example, as shown in FIG. 5, the estimation unit 42 estimates that the clothes worn by the subject T are a long-sleeved top and pants.

[0041] Next, the estimation unit 42 estimates the amount of clothing worn by the subject T based on the estimated clothing and the first correspondence table. The first correspondence table includes data in which each set of clothing and the claw value of the clothing is included. FIG. 6 is a diagram showing the first correspondence table according to this embodiment. As shown in FIG. 6, the first correspondence table shows multiple sets.

[0042] For example, when the estimated clothing is "pants," the estimation unit 42 estimates that the claw value of "pants" is 0.240. When the subject T is wearing multiple pieces of clothing, the estimation unit 42 estimates the claw value for each of the multiple pieces of clothing worn by the subject T, and estimates that the total value of the claw values ​​for each of the multiple pieces of clothing is the amount of clothing worn by the subject T.

[0043] Furthermore, the estimation unit 42 estimates the movement of the detected subject T based on the above detection result (the detection result that the subject T captured in the acquired moving image is detected). A known means can be used to estimate the movement of the subject T. As this known means, Deep Convolutional Networks or the like can be preferably used. In the present embodiment, the estimation unit 42 estimates the movement of the subject T using the above detection result and this known means.

[0044] Next, the estimation unit 42 estimates the metabolic rate of the estimated movement based on the estimated movement and the second correspondence table. The second correspondence table includes data in which a movement and a metabolic rate of the movement are grouped into one set. FIG. 7 is a diagram showing the second correspondence table according to this embodiment. As shown in FIG. 7, the second correspondence table shows a plurality of sets.

[0045] For example, when the estimated motion is "standing up," the estimation unit 42 estimates that the metabolic rate (MET) for "standing up" is 1.2.

[0046] As described above, the estimation unit 42 estimates the amount of clothing worn by the subject T and the metabolic rate of the subject T, that is, estimates subject information.

[0047] In addition, the estimation unit 42 may use known facial recognition technology or the like to associate an identifier that identifies the subject T with subject information indicating the amount of clothing worn by the subject T and the metabolic rate of the subject T.

[0048] Next, the machine learning model 421 held by the estimation unit 42 will be described.

[0049] For example, the machine learning model 421 may be trained as follows.

[0050] For example, the thermal comfort estimation device 40 may have a learning unit, which learns and constructs the machine learning model 421. The learning unit provides the constructed machine learning model 421 to the estimation unit 42. Note that the learning unit is not an essential component and does not necessarily have to be included in the thermal comfort estimation device 40.

[0051] This machine learning model 421 is a model for generating comfort information.

[0052] In this embodiment, the machine learning model 421 is a model constructed by machine learning using one or more datasets as training data. Here, multiple items related to one dataset will be described. The multiple items include the temperature, humidity, radiant temperature, and wind speed of a specified space in which a specific person is present, the amount of clothing worn by the specific person, the metabolic rate of the specific person's movements, and at least one item indicated by the subject information for the specific person. For example, in this embodiment, the multiple items include the temperature, humidity, radiant temperature, and wind speed of a specified space in which a specific person is present, the amount of clothing worn by the specific person, the metabolic rate of the specific person's movements, and the age, sex, race, height, and weight of the specific person. One dataset is composed of a combination of multiple items and the specific person's thermal comfort.

[0053] In other words, the machine learning model 421 is a recognition model constructed by machine learning using one or more datasets as training data. More specifically, the machine learning model 421 is a recognition model constructed using multiple items belonging to each of the one or more datasets that are the training data as input data and the thermal comfort of a specific person related to the multiple items belonging to the datasets as output data.

[0054] In this embodiment, training data is used to construct the machine learning model 421. For example, as the training data, a dataset including multiple sets is used, where each set is a result of measuring the thermal environment in a room (space) by an instrument and a subjective thermal comfort assessment by a person in the room.

[0055] The learning unit learns the model using machine learning, for example, as described above. Therefore, in this embodiment, the model is the machine learning model 421.

[0056] The learning unit may also train the machine learning model 421 using, for example, decision tree or gradient boosting algorithms. The learning unit may also train the machine learning model 421 using a learning algorithm other than those described above.

[0057] The estimation unit 42 inputs the environmental information, the attribute information, and the subject information estimated by the estimation unit 42 based on the moving image to the machine learning model 421. More specifically, the estimation unit 42 inputs the temperature, humidity, radiant temperature, and wind speed of the space 90 indicated by the environmental information, the age, sex, race, height, and weight of the subject T indicated by the attribute information, and the amount of clothing worn by the subject T and the metabolic rate of the subject T indicated by the subject information to the machine learning model 421. As a result, thermal comfort is output from the machine learning model 421. The thermal comfort output from the machine learning model 421 corresponds to the thermal comfort of the subject T indicated by the comfort information estimated by the estimation unit 42.

[0058] As described above, the attribute information is associated with an identifier that identifies the subject T, and the subject information is also associated with an identifier that identifies the subject T. Therefore, the estimation unit 42 can estimate the comfort information of the subject T by regarding both the attribute information and the subject information as information related to the subject T.

[0059] Specifically, the estimation unit 42 is realized by a processor or a microcomputer. The function of the estimation unit 42 is realized by the processor or microcomputer constituting the estimation unit 42 executing a computer program stored in the storage unit 43.

[0060] The storage unit 43 is a storage device that stores information necessary for the information processing performed by the estimation unit 42. The information stored in the storage unit 43 includes a computer program executed by the estimation unit 42. The storage unit 43 also stores the first and second correspondence tables described above. Specifically, the storage unit 43 is realized by a semiconductor memory, a HDD, or the like.

[0061] The display device 50 is a display installed in the space 90. The display device 50 is realized by a display panel such as a liquid crystal panel or an organic EL panel. The display device 50 has a communication unit. The communication unit acquires comfort information output from the communication unit 41 (output unit 412) of the thermal comfort estimation device 40. The display device 50 displays the thermal comfort value of the subject T indicated by the comfort information acquired by the communication unit. The display device 50 may display an indication corresponding to the thermal comfort value of the subject T (for example, a smiley face mark shown in FIG. 2 ).

[0062] The air conditioning device 60 is a device that conditions the space 90, and more specifically, is a device that can control the temperature of the space 90. The air conditioning device 60 is disposed in the space 90. The air conditioning device 60 has a communication unit. The communication unit acquires comfort information output from the communication unit 41 (output unit 412) of the thermal comfort estimation device 40. The air conditioning device 60 conditions the space 90 based on the comfort information acquired by the communication unit. For example, if the thermal comfort level of the subject T indicated by the acquired comfort information is −3, −2, or −1, the air conditioning device 60 conditions the space 90 so as to increase the temperature of the space 90. Furthermore, for example, if the thermal comfort level of the subject T indicated by the acquired comfort information is +1, +2, or +3, the air conditioning device 60 conditions the space 90 so as to decrease the temperature of the space 90. Also, for example, if the thermal comfort of the subject T indicated by the acquired comfort information is 0, the air conditioning device 60 conditions the space 90 so as to maintain the temperature of the space 90.

[0063] Next, the operation of the thermal comfort estimation method executed by the thermal comfort estimation system 1 in this embodiment configured as above will be described.

[0064] [Operation Example 1] As described above, in the thermal comfort estimation system 1, the thermal comfort estimation device 40 estimates comfort information indicating thermal comfort, and the display device 50 and the air conditioning device 60 operate based on the estimated thermal comfort. Such an operation (Operation Example 1) will be described below.

[0065] FIG. 8 is a flowchart of an operation example 1 according to this embodiment.

[0066] First, the environmental sensor 10 senses the environment of the space 90 and outputs environmental information indicating the temperature, humidity, radiant temperature, and wind speed of the space 90 to the thermal comfort estimation device 40 (S10).

[0067] Furthermore, the imaging device 20 outputs the moving image of the subject T to the thermal comfort estimation device 40 (S20).

[0068] Next, the operation reception unit of the information terminal 30 receives an operation from the subject T indicating the age, sex, race, height, and weight of the subject T. The communication unit of the information terminal 30 outputs attribute information indicating the age, sex, race, height, and weight of the subject T indicated by the operation received by the operation reception unit to the thermal comfort estimation device 40 (S30).

[0069] The acquisition unit 411 acquires the environmental information output from the environmental sensor 10, the attribute information output from the information terminal 30, and the moving image output from the imaging device 20 (S40).

[0070] Then, the estimation unit 42 estimates subject information indicating the amount of clothing worn by the subject T and the metabolic rate of the subject T based on the moving image acquired by the acquisition unit 411 (S50).

[0071] The estimation unit 42 estimates comfort information indicating the thermal comfort of the subject T based on the environmental information acquired by the acquisition unit 411, the attribute information acquired by the acquisition unit 411, and the subject information estimated by the estimation unit 42 (S60). Here, the estimation unit 42 inputs the environmental information, the attribute information, and the subject information to the machine learning model 421. As a result, thermal comfort is output from the machine learning model 421, and this output thermal comfort corresponds to the thermal comfort of the subject T indicated by the comfort information estimated by the estimation unit 42.

[0072] The output unit 412 outputs the comfort information estimated by the estimation unit 42 to each of the display device 50 and the air conditioner 60 (S70).

[0073] The display device 50 displays the numerical value of the thermal comfort of the subject T indicated by the comfort information output by the output unit 412 (S80).

[0074] The air conditioner 60 conditions the space 90 based on the comfort information output by the output unit 412 (S90).

[0075] Here, the accuracy of the thermal comfort of the subject T indicated by the comfort information estimated by the thermal comfort estimation system 1 according to this embodiment will be described.

[0076] First, conventional thermal comfort will be explained.

[0077] For example, a comfort index called PMV (Predicted Mean Vote) has been used to estimate conventional thermal comfort. The PMV of the subject T can be calculated based on six items consisting of the temperature, humidity, radiant temperature, and wind speed of the space 90, as well as the metabolic rate and amount of clothing worn by the subject T. Note that the PMV is also an index expressed in seven stages from -3 to +3, as in the present embodiment.

[0078] In contrast, the thermal comfort estimation system 1 according to the present embodiment estimates comfort information indicating the thermal comfort of the subject T based on the above-mentioned multiple items. That is, in this embodiment, compared to PMV, comfort information is estimated based on at least one of the above-mentioned items indicated by the attribute information, more specifically, based on five items (age, sex, race, height, and weight). Therefore, the thermal comfort estimation system 1 according to the present embodiment can estimate thermal comfort with higher accuracy than PMV.

[0079] Furthermore, when the inventors compared the thermal comfort of the PMV with that of the present embodiment, it became clear that the thermal comfort estimation system 1 of the present embodiment can estimate thermal comfort with higher accuracy than the PMV.

[0080] [Modification 1 of the embodiment] Modification 1 of the embodiment will be described below. The following description will focus on the differences from the embodiment, and the description of the commonalities will be omitted or simplified.

[0081] Fig. 9 is a block diagram showing the functional configuration of a thermal comfort estimation system 1b according to a first modified embodiment of the present invention. Fig. 10 is a diagram showing a space 90 to which the thermal comfort estimation system 1b according to the first modified embodiment of the present invention is applied.

[0082] The thermal comfort estimation system 1b according to this modification has the same configuration as the thermal comfort estimation system 1 according to the embodiment, except that it further includes an information terminal 30b. This modification differs from the embodiment in that there are multiple subjects T (more specifically, two subjects T) in the space 90.

[0083] 10, there are two subjects T in a space 90. When identification is necessary, one subject T will be referred to as subject TA, and the other subject T will be referred to as subject TB.

[0084] The imaging device 20 according to this modification is a device that captures moving images of two subjects TA and TB present in a space 90. The imaging device 20 outputs the moving images to the thermal comfort estimation device 40.

[0085] The information terminal 30 is a terminal carried by the subject TA, and the information terminal 30b is a terminal carried by the subject TB. The information terminal 30b is, for example, a smartphone owned by the subject TB, but is not limited to this and may be a tablet terminal, a PDA, or other portable terminal.

[0086] An operation reception unit included in the information terminal 30 according to this modification receives an operation from the subject TA indicating the age, sex, race, height, and weight of the subject TA. A communication unit included in the information terminal 30 outputs attribute information (hereinafter, for identification purposes, sometimes referred to as attribute information A) indicating the age, sex, race, height, and weight of the subject TA indicated by the operation received by the operation reception unit to the thermal comfort estimation device 40. Note that the attribute information A is associated with an identifier (hereinafter, identifier A) that identifies the subject TA carrying the information terminal 30.

[0087] The information terminal 30b according to this modification has the same configuration as the information terminal 30. Accordingly, an operation reception unit included in the information terminal 30b receives an operation from the subject TB indicating the age, sex, race, height, and weight of the subject TB. A communication unit included in the information terminal 30b outputs attribute information (hereinafter, sometimes referred to as attribute information B for identification purposes) indicating the age, sex, race, height, and weight of the subject TB indicated by the operation received by the operation reception unit to the thermal comfort estimation device 40. Note that the attribute information B is associated with an identifier (hereinafter, identifier B) that identifies the subject TB carrying the information terminal 30b.

[0088] In this modified example, the communication unit 41 (more specifically, the acquisition unit 411) of the thermal comfort estimation device 40 acquires attribute information A output from the information terminal 30 and attribute information B output from the information terminal 30b. The acquisition unit 411 also acquires environmental information output from the environmental sensor 10 and moving images output from the imaging device 20.

[0089] The estimation unit 42 estimates comfort information indicating the thermal comfort of the subject TA and comfort information indicating the thermal comfort of the subject TB using the trained machine learning model 421. For ease of identification, the comfort information indicating the thermal comfort of the subject TA may be referred to as comfort information A, and the comfort information indicating the thermal comfort of the subject TB may be referred to as comfort information B.

[0090] The estimation unit 42 estimates the comfort information A and B as follows.

[0091] First, as in the embodiment, the estimation unit 42 estimates subject information indicating the amount of clothing worn by subject TA and the metabolic rate of subject TA, and subject information indicating the amount of clothing worn by subject TB and the metabolic rate of subject TB, based on the acquired moving images. Note that, for identification purposes, the subject information for subject TA may be referred to as subject information A, and the subject information for subject TB may be referred to as subject information B.

[0092] The procedure by which the estimation unit 42 estimates the subject information A is the same as the procedure by which the estimation unit 42 estimates the subject information according to the embodiment. That is, the estimation unit 42 detects the subject TA depicted in the acquired moving image, estimates the skeleton of the detected subject TA, estimates the clothing worn by the subject TA based on the estimated skeleton of the subject TA, and estimates the amount of clothing worn by the subject TA based on the estimated clothing. Furthermore, the estimation unit 42 estimates the movements of the subject TA based on the detected subject TA, and estimates the metabolic rate of the subject TA based on the estimated movements. In this way, the estimation unit 42 estimates the subject information A indicating the amount of clothing worn by the subject TA and the metabolic rate of the subject TA. The procedure by which the estimation unit 42 estimates the subject information B is also the same as described above.

[0093] The estimation unit 42 may use known face recognition technology or the like to associate identifier A that identifies subject TA with subject information A, and identifier B that identifies subject TB with subject information B.

[0094] The estimation unit 42 then inputs the environmental information, the attribute information A, and the subject information A estimated by the estimation unit 42 based on the moving image to the machine learning model 421. More specifically, the estimation unit 42 inputs the temperature, humidity, radiant temperature, and wind speed of the space 90 indicated by the environmental information, the age, sex, race, height, and weight of the subject TA indicated by the attribute information A, and the amount of clothing worn by the subject TA and the metabolic rate of the subject TA indicated by the subject information A to the machine learning model 421. As a result, thermal comfort for the subject TA is output from the machine learning model 421. The thermal comfort output from the machine learning model 421 corresponds to the thermal comfort of the subject TA indicated by the comfort information A estimated by the estimation unit 42.

[0095] As described above, the attribute information A is associated with an identifier A that identifies the subject TA, and the subject information A is also associated with an identifier A that identifies the subject TA. Therefore, the estimation unit 42 can estimate the comfort information A of the subject TA by regarding both the attribute information A and the subject information A as information related to the subject TA.

[0096] Furthermore, the estimation unit 42 inputs the environmental information, the attribute information B, and the subject information B estimated by the estimation unit 42 based on the moving image to the machine learning model 421. More specifically, the estimation unit 42 inputs the temperature, humidity, radiant temperature, and wind speed of the space 90 indicated by the environmental information, the age, sex, race, height, and weight of the subject TB indicated by the attribute information B, and the amount of clothing worn by the subject TB and the metabolic rate of the subject TB indicated by the subject information B to the machine learning model 421. As a result, thermal comfort for the subject TB is output from the machine learning model 421. The thermal comfort output from the machine learning model 421 corresponds to the thermal comfort of the subject TB indicated by the comfort information B estimated by the estimation unit 42.

[0097] As described above, the attribute information B is associated with an identifier B that identifies the subject T B, and the subject information B is also associated with an identifier B that identifies the subject T B. Therefore, the estimation unit 42 can estimate the comfort information B of the subject T B by regarding both the attribute information B and the subject information B as information related to the subject T B.

[0098] Furthermore, the communication unit 41 (more specifically, the output unit 412) according to this modification outputs the comfort information A and B estimated by the estimation unit 42. The output unit 412 outputs the comfort information A and B to, for example, the display device 50 and the air conditioning device 60, respectively.

[0099] The communication unit of the display device 50 acquires the comfort information A and B output from the output unit 412. The display device 50 displays the thermal comfort value of the subject TA indicated by the comfort information A acquired by the communication unit, and the thermal comfort value of the subject TB indicated by the acquired comfort information B. Fig. 10 shows that the thermal comfort value of the subject TA (Mr. TA) is 0, and the thermal comfort value of the subject TB (Mr. TB) is +1.

[0100] The communication unit of the air conditioner 60 acquires the comfort information A and B output from the output unit 412. The air conditioner 60 conditions the space 90 based on the comfort information A and B acquired by the communication unit.

[0101] Here, the air conditioner 60 conditions the space 90 based on the average value of the thermal comfort value of the subject TA indicated by the comfort information A and the thermal comfort value of the subject TB indicated by the comfort information B. For example, if the average value is greater than or equal to −3 and less than 0, the air conditioner 60 conditions the space 90 so as to increase the temperature of the space 90. Also, for example, if the average value is greater than 0 and less than or equal to +3, the air conditioner 60 conditions the space 90 so as to decrease the temperature of the space 90. Also, for example, if the average value is 0, the air conditioner 60 conditions the space 90 so as to maintain the temperature of the space 90.

[0102] [Modification 2 of the embodiment] Modification 2 of the embodiment will be described below. The following description will focus on the differences from Modification 1 of the embodiment, and description of commonalities will be omitted or simplified.

[0103] Fig. 11 is a block diagram showing the functional configuration of a thermal comfort estimation system 1c according to Modification 2 of the embodiment. Fig. 12 is a diagram showing a space 90 and a space 90c to which the thermal comfort estimation system 1c according to Modification 2 of the embodiment is applied.

[0104] The thermal comfort estimation system 1c has the same configuration as the thermal comfort estimation system 1b relating to the first variant of the embodiment, except that it further includes an environmental sensor 10c, an imaging device 20c, information terminals 30c and 30d, a display device 50c, and an air conditioning device 60c.

[0105] 12, two subjects T (more specifically, subjects TA and TB) are present in space 90, which is an example of a first space. Furthermore, two subjects T other than subjects TA and TB are present in space 90c, which is an example of a second space different from the first space. Note that when it is necessary to distinguish between the two subjects T present in space 90c, one subject T will be referred to as subject TC and the other subject T as subject TD.

[0106] The space 90 and the space 90c according to this modification are, for example, spaces arranged in a single office. The space 90 and the space 90c are adjacent spaces separated by, for example, a partition.

[0107] In the space 90c, an environmental sensor 10c, an imaging device 20c, information terminals 30c and 30d, a display device 50c, and an air conditioner 60c are arranged.

[0108] The environmental sensor 10c is an example of a second environmental sensor, and is a sensor that senses the environment of the space 90c where the two subjects TC and TD are located. The environmental sensor 10c is a sensor that senses the temperature, humidity, radiant temperature, and wind speed of the space 90c. The environmental sensor 10c outputs environmental information indicating the environment of the space 90c to the thermal comfort estimation device 40. The environmental information indicating the environment of the space 90c is information indicating the temperature, humidity, radiant temperature, and wind speed of the space 90c. The environmental sensor 10c has the same configuration as the environmental sensor 10. The environmental sensor 10c is placed in the space 90c, and in this modified example, is placed on the ceiling of the space 90c, etc.

[0109] In the following, environmental information indicating the environment of space 90 obtained by sensing using environmental sensor 10 may be referred to as environmental information A, and environmental information indicating the environment of space 90c obtained by sensing using environmental sensor 10c may be referred to as environmental information C.

[0110] The imaging device 20c is an example of a second imaging device, and is a device that captures moving images of two subjects TC and TD present in the space 90c. The imaging device 20c outputs the moving images of the two subjects TC and TD to the thermal comfort estimation device 40. The imaging device 20c has the same configuration as the imaging device 20. The imaging device 20c is placed in the space 90c, and in this modified example, is placed on the ceiling of the space 90c.

[0111] In the following, a moving image showing two subjects TA and TB in space 90 may be referred to as moving image A, and a moving image showing two subjects TC and TD in space 90c may be referred to as moving image C.

[0112] The information terminal 30c is a terminal carried by the subject TC, and the information terminal 30d is a terminal carried by the subject TD. The information terminals 30c and 30d are, for example, smartphones, but are not limited to these, and may be mobile terminals such as tablet terminals or PDAs. Each of the information terminals 30c and 30d has the same configuration as the information terminal 30.

[0113] An operation reception unit included in the information terminal 30c according to this modification receives an operation from the subject TC indicating the age, sex, race, height, and weight of the subject TC. A communication unit included in the information terminal 30c outputs attribute information (hereinafter, sometimes referred to as attribute information C for identification purposes) indicating the age, sex, race, height, and weight of the subject TC indicated by the operation received by the operation reception unit to the thermal comfort estimation device 40. Note that the attribute information C is associated with an identifier C that identifies the subject TC carrying the information terminal 30c.

[0114] An operation reception unit included in the information terminal 30d according to this modification receives an operation from the subject TD indicating the age, sex, race, height, and weight of the subject TD. A communication unit included in the information terminal 30d outputs attribute information (hereinafter, sometimes referred to as attribute information D for identification purposes) indicating the age, sex, race, height, and weight of the subject TD indicated by the operation received by the operation reception unit to the thermal comfort estimation device 40. Note that the attribute information D is associated with an identifier D that identifies the subject TD carrying the information terminal 30d.

[0115] In this modification, the communication unit 41 (acquisition unit 411) of the thermal comfort estimation device 40 acquires attribute information A, B, C, and D output from the information terminals 30, 30b, 30c, and 30d, respectively. The acquisition unit 411 also acquires environmental information A and C output from the environmental sensors 10 and 10c, and moving images A and C output from the imaging devices 20 and 20c.

[0116] The estimation unit 42 estimates comfort information A indicating the thermal comfort of the subject TA, comfort information B indicating the thermal comfort of the subject TB, comfort information indicating the thermal comfort of the subject TC, and comfort information indicating the thermal comfort of the subject TD using the trained machine learning model 421. For ease of identification, the comfort information indicating the thermal comfort of the subject TC may be referred to as comfort information C, and the comfort information indicating the thermal comfort of the subject TD may be referred to as comfort information D.

[0117] First, the estimation unit 42 according to this modification estimates the comfort information A and B in the same procedure as the estimation unit 42 according to the first modification of the embodiment.

[0118] That is, the estimation unit 42 inputs the environmental information A, the attribute information A, and the subject information A estimated by the estimation unit 42 based on the moving image A to the machine learning model 421. As a result, thermal comfort for the subject TA is output from the machine learning model 421. The thermal comfort output from the machine learning model 421 corresponds to the thermal comfort of the subject TA indicated by the comfort information A estimated by the estimation unit 42. The estimation unit 42 estimates comfort information B using a procedure similar to the procedure for estimating comfort information A.

[0119] Furthermore, the estimation unit 42 estimates subject information indicating the amount of clothing worn by the subject TC and the metabolic rate of the subject TC, and subject information indicating the amount of clothing worn by the subject TD and the metabolic rate of the subject TD, based on the acquired moving image C. The acquired moving image C is a moving image showing multiple subjects T (two subjects TC and TD) present in the space 90c. For identification purposes, the subject information about the subject TC may be referred to as subject information C, and the subject information about the subject TD may be referred to as subject information D.

[0120] The procedure by which the estimation unit 42 estimates the subject information C is the same as the procedure by which the estimation unit 42 estimates the subject information A. That is, the estimation unit 42 detects the subject TC depicted in the acquired moving image C, estimates the skeleton of the detected subject TC, estimates the clothing worn by the subject TC based on the estimated skeleton of the subject TC, and estimates the amount of clothing worn by the subject TC based on the estimated clothing. Furthermore, the estimation unit 42 estimates the movements of the subject TC based on the detected subject TC, and estimates the metabolic rate of the subject TC based on the estimated movements. In this way, the estimation unit 42 estimates subject information C indicating the amount of clothing worn by the subject TC and the metabolic rate of the subject TC. The procedure by which the estimation unit 42 estimates subject information D is also the same as the above.

[0121] The estimation unit 42 may use a known face recognition technique or the like to associate an identifier C that identifies the subject TC with the subject information C, and an identifier D that identifies the subject TD with the subject information D.

[0122] The estimation unit 42 also inputs the environmental information C, the attribute information C, and the subject information C estimated by the estimation unit 42 based on the moving image C to the machine learning model 421. More specifically, the estimation unit 42 inputs the temperature, humidity, radiant temperature, and wind speed of the space 90c indicated by the environmental information C, the age, sex, race, height, and weight of the subject TC indicated by the attribute information C, and the amount of clothing worn by the subject TC and the metabolic rate of the subject TC indicated by the subject information C to the machine learning model 421. As a result, thermal comfort for the subject TC is output from the machine learning model 421. The thermal comfort output from the machine learning model 421 corresponds to the thermal comfort of the subject TC indicated by the comfort information C estimated by the estimation unit 42. The estimation unit 42 estimates comfort information D using a procedure similar to the procedure for estimating the comfort information C.

[0123] As described above, the attribute information C is associated with an identifier C that identifies the subject TC, and the subject information C is also associated with an identifier C that identifies the subject TC. Therefore, the estimation unit 42 can estimate the comfort information C of the subject TC by assuming that the attribute information C and the subject information C are both information related to the subject TC. Similarly, the estimation unit 42 can estimate the comfort information D of the subject TD by assuming that the attribute information D and the subject information D associated with the identifier D are both information related to the subject TD.

[0124] The communication unit 41 (more specifically, the output unit 412) according to this modification outputs the comfort information A and B estimated by the estimation unit 42. The output unit 412 also outputs the comfort information (comfort information C and D) estimated by the estimation unit 42 for the plurality of subjects T (two subjects TC and TD) present in the space 90c. The output unit 412 outputs the comfort information A and B to the display device 50 and the air conditioning device 60, respectively, and outputs the comfort information C and D to the display device 50c and the air conditioning device 60c, respectively, for example. In other words, the output unit 412 outputs the comfort information for any subject T to a display device and an air conditioning device arranged in a space in which the subject T is present.

[0125] The communication unit of the display device 50 arranged in the space 90 acquires the comfort information A and B output from the output unit 412. The display device 50 displays the thermal comfort value of the subject TA indicated by the comfort information A acquired by the communication unit, and the thermal comfort value of the subject TB indicated by the acquired comfort information B.

[0126] The display device 50c is a display installed in the space 90c. The display device 50c has the same configuration as the display device 50. The communication unit of the display device 50c acquires the comfort information C and D output from the output unit 412 of the thermal comfort estimation device 40. The display device 50c displays the thermal comfort value of the subject TC indicated by the comfort information C acquired by the communication unit, and the thermal comfort value of the subject TD indicated by the acquired comfort information D.

[0127] The communication unit of the air conditioner 60 acquires the comfort information A and B output from the output unit 412. The air conditioner 60 conditions the space 90 based on the comfort information A and B acquired by the communication unit.

[0128] The air conditioning device 60c is a device that conditions the space 90c, and more specifically, is a device that can control the temperature of the space 90c. The air conditioning device 60c is arranged in the space 90c. The air conditioning device 60c has the same configuration as the air conditioning device 60. A communication unit included in the air conditioning device 60c acquires the comfort information C and D output from the output unit 412 of the thermal comfort estimation device 40. The air conditioning device 60c conditions the space 90c based on the comfort information C and D acquired by the communication unit. Similar to the air conditioning device 60, the air conditioning device 60c conditions the space 90c based on, for example, the average value of the thermal comfort value of the subject TC indicated by the comfort information C and the thermal comfort value of the subject TD indicated by the comfort information D.

[0129] [Effects, etc.] Invention 1 is a thermal comfort estimation system 1 that includes a thermal comfort estimation device 40, a first environmental sensor (environment sensor 10) that senses the environment of a first space (space 90) where a subject T is present, and a first imaging device (imaging device 20). The thermal comfort estimation device 40 has an acquisition unit 411 that acquires environmental information indicating the environment of the first space output from the first environmental sensor, attribute information indicating the attributes of the subject T, and a video image of the subject T captured by the first imaging device, an estimation unit 42 that estimates comfort information indicating thermal comfort using a trained machine learning model 421 that receives as input the acquired environmental information, the acquired attribute information, and subject information related to the subject T based on the acquired video, and outputs the thermal comfort of the subject T, and an output unit 412 that outputs the estimated comfort information. For example, the attribute information is information including at least one of the attributes of the subject T, namely, age, sex, race, hair color, eye color, facial features, nationality, height, and weight.

[0130] As a result, the thermal comfort estimation system 1 according to the embodiment can estimate comfort information indicating the thermal comfort of the subject T based on at least one of the above-mentioned characteristics of the subject T, for which, for example, PMV according to the conventional technology has not been used. Therefore, the thermal comfort estimation system 1 according to the embodiment can estimate thermal comfort with higher accuracy than, for example, PMV according to the conventional technology.

[0131] A second aspect of the present invention is the thermal comfort estimation system 1 according to the first aspect, wherein the first imaging device is a visible light camera.

[0132] That is, a general-purpose camera can be used as the first imaging device according to the embodiment. Compared to the case where a specialized camera such as a camera is used, the thermal comfort estimation system 1 according to the embodiment can be introduced into a space 90 such as an office at low cost.

[0133] Invention 3 is the thermal comfort estimation system 1 according to Invention 1, wherein the subject information indicates the amount of clothing worn by the subject T and the metabolic rate of the subject T.

[0134] This makes it possible to estimate comfort information based on the amount of clothing worn by the subject T and the subject T's metabolic rate.

[0135] A fourth aspect of the present invention is the thermal comfort estimation system 1 according to the first aspect, wherein the environmental information indicates the temperature, humidity, radiant temperature, and wind speed of the first space.

[0136] As a result, the thermal comfort estimation system 1 according to the embodiment estimates comfort information indicating the thermal comfort of the subject T based on the temperature, humidity, radiant temperature, and wind speed of the space 90 in which the subject T is present, the amount of clothing worn by the subject T, the metabolic rate of the subject T's movements, and at least one of the above information about the subject T. That is, in the present embodiment, compared to PMV, comfort information is estimated based on at least one of the above information. Therefore, the thermal comfort estimation system 1 according to the embodiment can estimate thermal comfort with higher accuracy than, for example, PMV according to conventional technology.

[0137] A fifth aspect of the present invention is the thermal comfort estimation system 1 according to the first aspect, further comprising an air conditioning device 60 that performs air conditioning of the first space based on the output comfort information.

[0138] This allows the space 90 to be air-conditioned based on comfort information that indicates thermal comfort estimated with greater accuracy, thereby realizing a thermal comfort estimation system 1 that can further increase the comfort of the subject T.

[0139] Invention 6 is a thermal comfort estimation system 1b according to any one of Inventions 1 to 5, in which the acquisition unit 411 acquires attribute information (two pieces of attribute information A and B) of each of multiple subjects T (two subjects TA and TB) present in the first space, and video images of the multiple subjects T captured by a first imaging device, the estimation unit 42 estimates comfort information (two pieces of comfort information A and B) indicating the thermal comfort of each of the multiple subjects T using a machine learning model 421 that receives as input the acquired environmental information, the acquired multiple pieces of attribute information (two pieces of attribute information A and B), and subject information (two pieces of subject information A and B) relating to each of the multiple subjects T based on the acquired video images, and outputs the thermal comfort of each of the multiple subjects T, and the output unit 412 outputs the estimated multiple pieces of comfort information (two pieces of comfort information A and B).

[0140] This realizes a thermal comfort estimation system 1b that can estimate comfort information (comfort information A and B) for each of two subjects TA and TB even when there are multiple subjects T (two subjects TA and TB) in space 90.

[0141] A seventh aspect of the present invention is the thermal comfort estimation system 1b according to the sixth aspect of the present invention, which includes a display device 50 that displays the outputted plurality of pieces of comfort information (two pieces of comfort information A and B).

[0142] This allows the thermal comfort estimation system 1b to notify a person (e.g., subject TA) present in the space 90 of the thermal comfort of another person (e.g., subject TB) present in the space 90.

[0143] Invention 8 includes a second environmental sensor (environmental sensor 10c) different from the first environmental sensor, and a second imaging device (imaging device 20c) different from the first imaging device, and the second environmental sensor senses the environment of a second space (space 90c) different from the first space, and an acquisition unit 411 acquires environmental information (environmental information C) indicating the environment of the second space output from the second environmental sensor, attribute information (two pieces of attribute information C and D) indicating the attributes of each of a plurality of subjects T (two subjects TC and TD) present in the second space, and a moving image (moving image C) of the plurality of subjects T present in the second space captured by the second imaging device, and an estimation unit 42 estimates the acquired environmental information (environmental information C) indicating the environment of the second space and the acquired attribute information (two pieces of attribute information C and D) of each of the plurality of subjects T present in the second space. The thermal comfort estimation system 1c described in Invention 6 estimates comfort information (two pieces of comfort information C and D) indicating the thermal comfort of each of the multiple subjects T in the second space using a trained machine learning model 421 that takes as input attribute information (two pieces of attribute information C and D) indicating the attributes of each of the multiple subjects T in the second space and subject information (two pieces of subject information C and D) relating to each of the multiple subjects T in the second space based on an acquired moving image (moving image C) in which the multiple subjects T in the second space are captured, and outputs the thermal comfort of each of the multiple subjects T in the second space, and the output unit 412 outputs the estimated multiple pieces of comfort information (two pieces of comfort information C and D) of the multiple subjects T in the second space.

[0144] This realizes a thermal comfort estimation system 1c that can estimate comfort information (comfort information A to D) for each of multiple subjects T (four subjects TA to TD) even when there are multiple subjects T (two subjects TA and TB) in space 90 and multiple subjects T (two subjects TC and TD) in space 90c.

[0145] Invention 9 is a thermal comfort estimation method performed by a thermal comfort estimation system 1, which includes a thermal comfort estimation device 40, a first environmental sensor (environmental sensor 10) that senses the environment of a first space (space 90) in which a subject T is present, and a first imaging device (imaging device 20).The thermal comfort estimation method includes an acquisition step of acquiring environmental information indicating the environment of the first space output from the first environmental sensor, attribute information indicating the attributes of the subject T, and a video image of the subject T captured by the first imaging device, an estimation step of estimating comfort information indicating thermal comfort using a trained machine learning model 421 that inputs the acquired environmental information, the acquired attribute information, and subject information related to the subject T based on the acquired video, and outputs the thermal comfort of the subject T, and an output step of outputting the estimated comfort information. For example, the attribute information is information including at least one of the attributes of the subject T, namely, age, sex, race, hair color, eye color, facial features, nationality, height, and weight.

[0146] As a result, the thermal comfort estimation method according to the embodiment can estimate comfort information indicating the thermal comfort of the subject T based on at least one of the above-mentioned characteristics of the subject T, which was not used in the PMV according to the conventional technology, for example. Therefore, the thermal comfort estimation method according to the embodiment can estimate thermal comfort with higher accuracy than, for example, the PMV according to the conventional technology.

[0147] The invention 10 relates to a thermal comfort estimation device 40 including: an acquisition unit 411 that acquires environmental information indicating the environment of a first space (space 90) where a subject T is present, the environmental information being output from a first environmental sensor (environmental sensor 10) that senses the environment of the first space, attribute information indicating the attributes of the subject T, and a video image of the subject T captured by a first imaging device (imaging device 20); an estimation unit 42 that estimates comfort information indicating thermal comfort using a trained machine learning model 421 that receives as input the acquired environmental information, the acquired attribute information, and subject information related to the subject T based on the acquired video image and outputs the thermal comfort of the subject T; and an output unit 412 that outputs the estimated comfort information. For example, the attribute information is information including at least one of the attributes of the subject T, namely, age, sex, race, hair color, eye color, facial features, nationality, height, and weight.

[0148] As a result, the thermal comfort estimation device 40 according to the embodiment can estimate comfort information indicating the thermal comfort of the subject T based on at least one of the above-mentioned characteristics of the subject T, for which, for example, PMV according to the conventional technology has not been used. Therefore, the thermal comfort estimation device 40 according to the embodiment can estimate thermal comfort with higher accuracy than, for example, PMV according to the conventional technology.

[0149] (Other Embodiments) Although the embodiments have been described above, the present invention is not limited to the above-described embodiments and modifications.

[0150] Furthermore, for example, in the above embodiment, the thermal comfort estimation system is realized by multiple devices. When the thermal comfort estimation system is realized by multiple devices in this way, the components (particularly functional components) of the thermal comfort estimation system may be distributed in any way among the multiple devices. For example, some or all of the functions of the thermal comfort estimation device 40 may be provided by a server device. In other words, the communication unit 41, the estimation unit 42, and the memory unit 43 may be provided by the server device.

[0151] For example, the communication method between the devices in the above-described embodiment is not particularly limited. Furthermore, a relay device (such as a broadband router) (not shown) may be involved in the communication between the devices.

[0152] In the above-described embodiment, the processing performed by a specific processing unit may be performed by another processing unit. The order of multiple processing operations may be changed, or multiple processing operations may be performed in parallel.

[0153] In the above-described embodiments, each component may be realized by executing a software program suitable for that component, or by a program execution unit such as a CPU or processor reading and executing a software program recorded on a recording medium such as a hard disk or semiconductor memory.

[0154] Furthermore, each component may be realized by hardware. For example, each component may be a circuit (or integrated circuit). These circuits may form a single circuit as a whole, or each may be a separate circuit. Furthermore, each of these circuits may be a general-purpose circuit or a dedicated circuit.

[0155] Furthermore, the general or specific aspects of the present invention may be realized as a system, an apparatus, a method, an integrated circuit, a computer program, or a computer-readable recording medium such as a CD-ROM, or as any combination of a system, an apparatus, a method, an integrated circuit, a computer program, and a recording medium.

[0156] For example, the present invention may be realized as the thermal comfort estimation system of the above-described embodiment, or as a thermal comfort estimation method executed by a computer such as a thermal comfort estimation system. The present invention may also be realized as a program for causing a computer to execute such a thermal comfort estimation method, or as a non-transitory recording medium on which such a program is recorded.

[0157] In addition, the present invention also includes forms obtained by applying various modifications to each embodiment that a person skilled in the art would think of, or forms realized by arbitrarily combining the components and functions of each embodiment within the scope that does not deviate from the spirit of the present invention.

[0158] 1, 1b, 1c Thermal comfort estimation system 10, 10c Environmental sensor 20, 20c Imaging device 40 Thermal comfort estimation device 42 Estimation unit 50, 50c Display device 60, 60c Air conditioner 90, 90c Space 411 Acquisition unit 412 Output unit 421 Machine learning model T, TA, TB, TC, TD Subject

Claims

1. A thermal comfort estimation system comprising: a thermal comfort estimation device; a first environmental sensor that senses the environment of a first space in which a subject is present; and a first imaging device, wherein the thermal comfort estimation device has an acquisition unit that acquires environmental information indicating the environment of the first space output from the first environmental sensor, attribute information indicating the attributes of the subject, and a video image of the subject captured by the first imaging device; an estimation unit that estimates comfort information indicating the thermal comfort using a trained machine learning model that inputs the acquired environmental information, the acquired attribute information, and subject information related to the subject based on the acquired video, and outputs the thermal comfort of the subject; and an output unit that outputs the estimated comfort information.

2. The thermal comfort estimation system according to claim 1, wherein the first imaging device is a visible light camera.

3. The thermal comfort estimation system according to claim 1, wherein the subject information indicates the amount of clothing worn by the subject and the metabolic rate of the subject.

4. The thermal comfort estimation system according to claim 3, wherein the environmental information indicates the temperature, humidity, radiant temperature, and wind speed of the first space.

5. The thermal comfort estimation system according to claim 1, further comprising an air conditioning device that conditions the first space based on the output comfort information.

6. The thermal comfort estimation system described in any one of claims 1 to 5, wherein the acquisition unit acquires the attribute information of each of the multiple subjects present in the first space and the moving images of the multiple subjects captured by the first imaging device, the estimation unit estimates the comfort information indicating the thermal comfort of each of the multiple subjects using the machine learning model that receives as input the acquired environmental information, the acquired multiple pieces of attribute information, and the subject information for each of the multiple subjects based on the acquired moving images, and outputs the thermal comfort of each of the multiple subjects, and the output unit outputs the estimated multiple pieces of comfort information.

7. The thermal comfort estimation system according to claim 6, further comprising a display device that displays the outputted plurality of pieces of comfort information.

8. A system comprising: a second environmental sensor different from the first environmental sensor; and a second imaging device different from the first imaging device; wherein the second environmental sensor senses an environment of a second space different from the first space; wherein the acquisition unit acquires environmental information indicating the environment of the second space output from the second environmental sensor, attribute information indicating attributes of each of a plurality of subjects present in the second space, and video images of the plurality of subjects present in the second space captured by the second imaging device; and wherein the estimation unit estimates comfort information indicating the thermal comfort of each of the plurality of subjects present in the second space using the trained machine learning model that receives as input the acquired environmental information indicating the environment of the second space, the attribute information indicating attributes of each of the acquired subjects present in the second space, and subject information related to each of the plurality of subjects present in the second space based on the acquired video images of the plurality of subjects present in the second space, and outputs the thermal comfort of each of the plurality of subjects present in the second space; The thermal comfort estimation system according to claim 6 , wherein the output unit outputs the estimated plurality of pieces of comfort information for the plurality of subjects present in the second space.

9. A thermal comfort estimation method performed by a thermal comfort estimation system, wherein the thermal comfort estimation system comprises a thermal comfort estimation device, a first environmental sensor that senses the environment of a first space in which a subject is present, and a first imaging device, and the thermal comfort estimation method includes an acquisition step of acquiring environmental information indicative of the environment of the first space output from the first environmental sensor, attribute information indicative of the attributes of the subject, and a video image of the subject captured by the first imaging device, an estimation step of estimating comfort information indicative of the thermal comfort using a trained machine learning model that receives as input the acquired environmental information, the acquired attribute information, and subject information related to the subject based on the acquired video, and outputs the thermal comfort of the subject, and an output step of outputting the estimated comfort information.

10. A thermal comfort estimation device comprising: an acquisition unit that acquires environmental information indicating the environment of a first space in which a subject is present, output from a first environmental sensor that senses the environment of the first space, attribute information indicating the attributes of the subject, and a video image of the subject captured by a first imaging device; an estimation unit that estimates comfort information indicating the thermal comfort using a trained machine learning model that receives as input the acquired environmental information, the acquired attribute information, and subject information related to the subject based on the acquired video, and outputs the thermal comfort of the subject; and an output unit that outputs the estimated comfort information.

Images