AI indoor environment adjusting system and method
Through the AI indoor environment regulation system, combining emotional and physical condition analysis to adjust the indoor environment, the problem that the existing system cannot be intelligently adjusted is solved, and a personalized health and comfort experience is achieved.
Patent Information
- Application Number
- CN202411363508.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-27
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing smart home system fails to fully consider the physical and mental health of users and cannot intelligently adjust the indoor environment to provide a high-quality experience.
An AI indoor environment regulation system was designed, including an emotion analysis module, a physical condition analysis module and an indoor environment regulation module. User text information, voice information and facial expression data were collected through the mobile phone, and the user's emotions were analyzed in combination with a convolutional neural network. Physical condition and adjust the indoor temperature, humidity, ventilation and fragrance according to the analysis results.
It realizes real-time adjustment of the indoor environment according to the user's emotions and physical condition, provides a personalized health and comfort experience, and enhances the user's physical and mental health experience.
Smart Images

Figure CN118935701A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of software information technology, and in particular to an AI indoor environment adjustment system and method. Background Art
[0002] With the continuous iteration of smart home ecological management system technology, the adjustment of indoor environmental elements such as indoor temperature, humidity, ventilation, etc. has become more and more convenient, but it still fails to fully consider the physical and mental health of users. A more intelligent indoor environment adjustment system is needed to adjust various parameters of the indoor environment and bring users a better indoor environment experience. Summary of the invention
[0003] In order to realize intelligent adjustment of the indoor environment according to the physical and mental health of the user, the present invention designs an AI indoor environment adjustment system and method.
[0004] The technical solution adopted by the present invention is an AI indoor environment adjustment system: It includes emotion analysis module, physical condition analysis module and indoor environment adjustment module.
[0005] The emotion analysis module is used to analyze the user's emotions in real time through the data collected by the mobile phone. The specific method is as follows: A1, collects text information input by users on mobile phones, including social interaction information, comment information, and status display information; A2, collects the voice information input by the user on the mobile phone, including speech speed, fundamental frequency, and volume; A3, collects the user’s facial expression information through the front camera of the mobile phone; A4, classifies emotions into multiple types, including peace, happiness, excitement, anger, sadness, and anxiety; A5, Analyze user emotions based on text information: ① Create an emotional text sample library, collect existing texts expressing emotions, and classify them according to the emotional types in A4; ② Based on the text input by the user on the mobile phone within 30 minutes, extract the text that is the same as the text in the emotional text sample library, classify it according to the emotional type in A4, and count the number of texts belonging to each emotional type; ③ Assume that the number of words entered by the user on the mobile phone within 30 minutes is m, and the number of words classified into a certain emotion type is n, then the score of this emotion is S 1i =100×n÷m; i=1, 2, 3, 4, 5, 6, S 11 , S 12 , S 13 , S 14 , S 15 , S 16They represent the six emotion scores obtained through text information analysis: peace, happiness, excitement, anger, sadness, and anxiety.
[0006] A6, analyzing user emotions based on voice information: Calculate the speech rate, fundamental frequency, and volume based on the voice information input by the user on the mobile phone within 30 minutes; the unit of speech rate is SPM (Syllables Per Minute), the unit of fundamental frequency is Hz (Hertz), and the unit of volume is dB (decibel); Suppose the fluctuation threshold of speech rate, fundamental frequency and volume is Q=k×sqrt{[(x1-u) 2 +(x2-u) 2 +…+(x n -u) 2 ]÷n}, u is the average value of the data within 30 minutes, k is the coefficient, x j For each data point, j = 1, 2, ... n, sqrt{} represents the square root, and the coefficient k in the fluctuation threshold formula of speech rate, fundamental frequency, and volume is determined according to the respective test results; u=(x1+x2+…+x n )÷n, where n is the total number of data points; The fluctuation value of speech speed, fundamental frequency and volume P=x j -u; Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through voice information analysis are S 21 , S 22 , S 23 , S 24 , S 25 , S 26 ; ① When the fluctuation values of speech speed, fundamental frequency and volume are all less than their respective fluctuation thresholds, the peaceful emotion score S 21 =50, sad emotion score S 25 =50, other types of emotions scored as 0; ② When the fluctuation values of speech speed and volume are both less than their respective fluctuation thresholds, and the fluctuation value of fundamental frequency exceeds its fluctuation threshold, the pleasure score S 22 =50, anxiety score S 26 =50, other types of emotions scored as 0; ③ When the fluctuation values of speech speed, fundamental frequency and volume are all greater than their respective fluctuation thresholds, the excitement score S 23 =50, anger score S 24 =50, and other types of emotions scored as 0.
[0007] A7, analyzing user emotions based on user facial expression information: ① Collect facial expression image datasets and annotate the emotion types. The facial expression image dataset contains samples of peace, happiness, excitement, anger, sadness, and anxiety. Adjust the images in the samples to grayscale images of 48x48 pixels. ②Build a convolutional neural network model, Input layer: The input is processed into a 48x48 pixel grayscale image of facial expression; Convolutional layer: Multiple convolutional layers are used to extract features, the convolution kernel size is 3x3, and the ReLU activation function is used; Pooling layer: reduces feature dimensions, reduces computational effort, and increases model robustness; Fully connected layer: flattens the extracted features and performs sentiment classification; Output layer: Use the softmax activation function to output the probability distribution of each emotion.
[0008] ③Train the model, divide the data set into a training set and a validation set, and select the optimal parameters for the model through cross-validation.
[0009] ④ Emotion recognition: Use the trained model to recognize the user's facial expression image, output the corresponding emotion type and probability, multiply the probability by 100, and obtain a score of 100 for each emotion. Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through facial expression information analysis are S, 31 , S 32 , S 33 , S 34 , S 35 , S 36 .
[0010] A8: Comprehensively judge the user's emotions. The total score of each emotion is calculated as follows: ① The total score of peaceful mood is S 11 +S 21 +S 31 ; ②The total score of pleasant emotion is S 12 +S 22 +S 32 ; ③The total score of excitement is S 13 +S 23 +S 33 ; ④The total score of anger is S 14 +S 24 +S 34 ; ⑤The total score of sadness is S 15 +S 25 +S 35 ; ⑥The total score of anxiety is S16 +S 26 +S 36 .
[0011] The emotion type with the highest comprehensive score is determined to be the user's emotion.
[0012] The physical condition analysis module is used to analyze the user's physical condition through the data collected by the mobile phone, specifically in the following ways: B1, the mobile phone collects the user's physical condition data from the health monitoring portable device worn by the user, including heart rate, blood oxygen saturation, body temperature, and sleep data.
[0013] B2, analyze heart rate data; The user's real-time resting heart rate is compared with the historical resting heart rate average value, and the excellent heart rate difference, good heart rate difference, medium heart rate difference, and poor heart rate difference are set. 0<excellent heart rate difference<good heart rate difference<medium heart rate difference<poor heart rate difference; When (historical resting heart rate average value - heart rate excellent difference) ≤ real-time resting heart rate ≤ (historical resting heart rate average value + heart rate excellent difference), the current heart rate is judged to be excellent; When (historical resting heart rate average value - good heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - excellent heart rate difference), or (historical resting heart rate average value + excellent heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + good heart rate difference), the current heart rate is determined to be good; When (historical resting heart rate average value - heart rate medium difference) ≤ real-time resting heart rate < (historical resting heart rate average value - heart rate good difference value), or (historical resting heart rate average value + heart rate good difference value) < real-time resting heart rate ≤ (historical resting heart rate average value + heart rate medium difference value), the current heart rate is determined to be medium; When (historical resting heart rate average value - poor heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - medium heart rate difference), or (historical resting heart rate average value + medium heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + poor heart rate difference), the current heart rate is determined to be poor.
[0014] B3, analyze blood oxygen saturation data: When 95%<blood oxygen saturation≤100%, it means that the blood oxygen content is excellent; When 90%<blood oxygen saturation≤95%, it means that the blood oxygen content is good; When 85%<blood oxygen saturation≤90%, it means that the blood oxygen content is moderate; When the blood oxygen saturation is ≤85%, it means that the blood oxygen content is poor.
[0015] B4, analyze body temperature data: The user's real-time body temperature is compared with the historical average body temperature, and the excellent temperature difference, good temperature difference, medium temperature difference, and poor temperature difference are set. 0<excellent temperature difference<good temperature difference<medium temperature difference<poor temperature difference; When (historical temperature average value - temperature difference) ≤ real-time temperature ≤ (historical temperature average value + temperature difference), the current temperature is considered excellent; When (historical temperature average value - good temperature difference) ≤ real-time temperature < (historical temperature average value - excellent temperature difference), or (historical temperature average value + excellent temperature difference) < real-time temperature ≤ (historical temperature average value + good temperature difference), the current temperature is considered good; When (historical temperature average value - medium temperature difference) ≤ real-time temperature < (historical temperature average value - good temperature difference), or (historical temperature average value + good temperature difference) < real-time temperature ≤ (historical temperature average value + medium temperature difference), the current temperature is determined to be medium; When (historical body temperature average value - abnormal body temperature difference) ≤ real-time body temperature < (historical body temperature average value - average body temperature difference), or (historical body temperature average value + average body temperature difference) < real-time body temperature ≤ (historical body temperature average value + abnormal body temperature difference), the current body temperature is determined to be abnormal.
[0016] B5, Analyze sleep data: Sleep efficiency V = actual sleep time ÷ time in bed × 100%; When V ≥ 90%, it is excellent sleep; When 85%≤V<90%, it is a good sleep; When 75%≤V<85%, it is moderate sleep; When V is less than 75%, it is poor sleep.
[0017] B6, classifies physical condition into four categories: excellent, good, moderate, and poor; When heart rate, blood oxygen saturation, body temperature and sleep all reach excellent levels, the physical condition is judged to be excellent; When the heart rate, blood oxygen saturation, body temperature, and sleep are at good levels and there are no medium or bad levels, it is judged that the physical condition is good; When the heart rate, blood oxygen saturation, body temperature, and sleep are at medium levels and there are no bad levels, the physical condition is judged to be medium; When heart rate, blood oxygen saturation, body temperature and sleep are at bad levels, it is judged that the physical condition is poor.
[0018] The indoor environment adjustment module adjusts the indoor environment according to the user's mood and physical condition, specifically in the following ways: C1, all objects that need to be located indoors are positioned using WiFi to create an indoor coordinate system. The indoor environment adjustment module connects all temperature, humidity, ventilation, and fragrance control systems to number and locate all temperature, humidity, ventilation, and fragrance equipment and their vents; locate all user mobile phones; C2, the initial indoor temperature was adjusted to 22-23°C, the initial indoor humidity was adjusted to 45-50%RH, and the ventilation intensity was 1-2ACH; C3, when there is only one user in the room; ① When the user's physical condition is at an excellent level, the initial indoor temperature, humidity, and ventilation intensity remain unchanged; When users want to feel peaceful and happy, they don’t use fragrance; When the user is in an excited mood, use lavender fragrance; When the user is in an angry mood, use sandalwood fragrance; When the user is in a sad mood, use orange blossom fragrance; When users are anxious, they use clary sage fragrance.
[0019] ② When the user's physical condition is at a good level, based on the initial indoor temperature, humidity, and ventilation data, the temperature rises by 1°C, the humidity remains unchanged, and the ventilation intensity increases by 1ACH; no fragrance is used; ③ When the user's physical condition is at a moderate level, based on the initial indoor temperature, humidity, and ventilation data, the temperature is increased by 2°C, the humidity is increased by 5%RH, and the ventilation intensity is increased by 2ACH; no fragrance is used; ④ When the user's physical condition is at a poor level, based on the initial indoor temperature, humidity, and ventilation data, the temperature increases by 3°C, the humidity increases by 10%RH, and the ventilation intensity increases by 3ACH; no fragrance is used.
[0020] C4, when there are multiple users in the room; ① When all users are in excellent physical condition, the initial indoor temperature, humidity and ventilation intensity remain unchanged; Count the number of users with each emotion, and use the fragrance based on the emotion with the largest number of users, in the same way as ① in C3; when there are multiple emotions with the largest number of users, switch the corresponding fragrance type in turn every 2 hours, in the same way as ① in C3; ② When there are users in good, medium or poor physical condition in the room, do not use fragrance; When the worst physical condition among all users in the room is at a good level, adjust the humidity and ventilation intensity according to ② in C3; When the worst physical condition among all users in the room is at a medium level, adjust the humidity and ventilation intensity according to ③ in C3; When the worst physical condition among all users in the room is at the bad level, adjust the humidity and ventilation intensity according to ④ in C3; The circle with a diameter of 3m on the ground directly below each air outlet is used as the statistical area, and the physical condition of the users directly below each air outlet is counted: When the worst physical condition of the user under a certain air outlet is in good condition, the air outlet temperature of this air outlet increases by 1°C; When the worst physical condition of the users under a certain air outlet is at the intermediate level, the air outlet temperature of this air outlet increases by 2°C; When the worst physical condition among the users under a certain air outlet is at the bad level, the air outlet temperature of this air outlet increases by 3°C.
[0021] An AI indoor environment adjustment method: Step 1: Analyze the user's emotions in real time through the data collected by the mobile phone. The specific method is as follows: A1, collects text information input by users on mobile phones, including social interaction information, comment information, and status display information; A2, collects the voice information input by the user on the mobile phone, including speech speed, fundamental frequency, and volume; A3, collects the user’s facial expression information through the front camera of the mobile phone; A4, classifies emotions into multiple types, including peace, happiness, excitement, anger, sadness, and anxiety; A5, Analyze user emotions based on text information: ① Create an emotional text sample library, collect existing texts expressing emotions, and classify them according to the emotional types in A4; ② Based on the text input by the user on the mobile phone within 30 minutes, extract the text that is the same as the text in the emotional text sample library, classify it according to the emotional type in A4, and count the number of texts belonging to each emotional type; ③ Assume that the number of words entered by the user on the mobile phone within 30 minutes is m, and the number of words classified into a certain emotion type is n, then the score of this emotion is S 1i =100×n÷m; i=1, 2, 3, 4, 5, 6, S 11 , S 12 , S 13 , S 14 , S 15 , S 16 They represent the six emotion scores obtained through text information analysis: peace, happiness, excitement, anger, sadness, and anxiety.
[0022] A6, analyzing user emotions based on voice information: Calculate the speech rate, fundamental frequency, and volume based on the voice information input by the user on the mobile phone within 30 minutes; the unit of speech rate is SPM (Syllables Per Minute), the unit of fundamental frequency is Hz (Hertz), and the unit of volume is dB (decibel); Suppose the fluctuation threshold of speech rate, fundamental frequency and volume is Q=k×sqrt{[(x1-u) 2 +(x2-u) 2 +…+(x n -u) 2 ]÷n}, u is the average value of the data within 30 minutes, k is the coefficient, x j For each data point, j = 1, 2, ... n, sqrt{} represents the square root, and the coefficient k in the fluctuation threshold formula of speech rate, fundamental frequency, and volume is determined according to the respective test results; u=(x1+x2+…+x n )÷n, where n is the total number of data points; The fluctuation value of speech speed, fundamental frequency and volume P=x j -u; Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through voice information analysis are S 21 , S 22 , S 23 , S 24 , S 25 , S 26 ; ① When the fluctuation values of speech speed, fundamental frequency and volume are all less than their respective fluctuation thresholds, the peaceful emotion score S 21 =50, sad emotion score S 25 =50, other types of emotions scored as 0; ② When the fluctuation values of speech speed and volume are both less than their respective fluctuation thresholds, and the fluctuation value of fundamental frequency exceeds its fluctuation threshold, the pleasure score S 22 =50, anxiety score S 26 =50, other types of emotions scored as 0; ③ When the fluctuation values of speech speed, fundamental frequency and volume are all greater than their respective fluctuation thresholds, the excitement score S 23 =50, anger score S 24 =50, and other types of emotions scored as 0.
[0023] A7, analyzing user emotions based on user facial expression information: ① Collect facial expression image datasets and annotate the emotion types. The facial expression image dataset contains samples of peace, happiness, excitement, anger, sadness, and anxiety. Adjust the images in the samples to grayscale images of 48x48 pixels. ②Build a convolutional neural network model, Input layer: The input is processed into a 48x48 pixel grayscale image of facial expression; Convolutional layer: Multiple convolutional layers are used to extract features, the convolution kernel size is 3x3, and the ReLU activation function is used; Pooling layer: reduces feature dimensions, reduces computational effort, and increases model robustness; Fully connected layer: flattens the extracted features and performs sentiment classification; Output layer: Use the softmax activation function to output the probability distribution of each emotion.
[0024] ③Train the model, divide the data set into a training set and a validation set, and select the optimal parameters for the model through cross-validation.
[0025] ④ Emotion recognition: Use the trained model to recognize the user's facial expression image, output the corresponding emotion type and probability, multiply the probability by 100, and obtain a score of 100 for each emotion. Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through facial expression information analysis are S, 31 , S 32 , S 33 , S 34 , S 35 , S 36 .
[0026] A8: Comprehensively judge the user's emotions. The total score of each emotion is calculated as follows: ① The total score of peaceful mood is S 11 +S 21 +S 31 ; ②The total score of pleasant emotion is S 12 +S 22 +S 32 ; ③The total score of excitement is S 13 +S 23 +S 33 ; ④The total score of anger is S 14 +S 24 +S 34 ; ⑤The total score of sadness is S 15 +S 25 +S 35 ; ⑥The total score of anxiety is S 16 +S 26 +S 36 .
[0027] The emotion type with the highest comprehensive score is determined to be the user's emotion.
[0028] Step 2: Analyze the user's physical condition through the data collected by the mobile phone. The specific method is as follows: B1, the mobile phone collects the user's physical condition data from the health monitoring portable device worn by the user, including heart rate, blood oxygen saturation, body temperature, and sleep data.
[0029] B2, analyze heart rate data; The user's real-time resting heart rate is compared with the historical resting heart rate average value, and the excellent heart rate difference, good heart rate difference, medium heart rate difference, and poor heart rate difference are set. 0<excellent heart rate difference<good heart rate difference<medium heart rate difference<poor heart rate difference; When (historical resting heart rate average value - heart rate excellent difference) ≤ real-time resting heart rate ≤ (historical resting heart rate average value + heart rate excellent difference), the current heart rate is judged to be excellent; When (historical resting heart rate average value - good heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - excellent heart rate difference), or (historical resting heart rate average value + excellent heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + good heart rate difference), the current heart rate is determined to be good; When (historical resting heart rate average value - heart rate medium difference) ≤ real-time resting heart rate < (historical resting heart rate average value - heart rate good difference value), or (historical resting heart rate average value + heart rate good difference value) < real-time resting heart rate ≤ (historical resting heart rate average value + heart rate medium difference value), the current heart rate is determined to be medium; When (historical resting heart rate average value - poor heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - medium heart rate difference), or (historical resting heart rate average value + medium heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + poor heart rate difference), the current heart rate is determined to be poor.
[0030] B3, analyze blood oxygen saturation data: When 95%<blood oxygen saturation≤100%, it means that the blood oxygen content is excellent; When 90%<blood oxygen saturation≤95%, it means that the blood oxygen content is good; When 85%<blood oxygen saturation≤90%, it means that the blood oxygen content is moderate; When the blood oxygen saturation is ≤85%, it means that the blood oxygen content is poor.
[0031] B4, analyze body temperature data: The user's real-time body temperature is compared with the historical average body temperature, and the excellent temperature difference, good temperature difference, medium temperature difference, and poor temperature difference are set. 0<excellent temperature difference<good temperature difference<medium temperature difference<poor temperature difference; When (historical temperature average value - temperature difference) ≤ real-time temperature ≤ (historical temperature average value + temperature difference), the current temperature is considered excellent; When (historical temperature average value - good temperature difference) ≤ real-time temperature < (historical temperature average value - excellent temperature difference), or (historical temperature average value + excellent temperature difference) < real-time temperature ≤ (historical temperature average value + good temperature difference), the current temperature is considered good; When (historical temperature average value - medium temperature difference) ≤ real-time temperature < (historical temperature average value - good temperature difference), or (historical temperature average value + good temperature difference) < real-time temperature ≤ (historical temperature average value + medium temperature difference), the current temperature is determined to be medium; When (historical body temperature average value - abnormal body temperature difference) ≤ real-time body temperature < (historical body temperature average value - average body temperature difference), or (historical body temperature average value + average body temperature difference) < real-time body temperature ≤ (historical body temperature average value + abnormal body temperature difference), the current body temperature is determined to be abnormal.
[0032] B5, Analyze sleep data: Sleep efficiency V = actual sleep time ÷ time in bed × 100%; When V ≥ 90%, it is excellent sleep; When 85%≤V<90%, it is a good sleep; When 75%≤V<85%, it is moderate sleep; When V is less than 75%, it is poor sleep.
[0033] B6, classifies physical condition into four categories: excellent, good, moderate, and poor; When heart rate, blood oxygen saturation, body temperature and sleep all reach excellent levels, the physical condition is judged to be excellent; When the heart rate, blood oxygen saturation, body temperature, and sleep are at good levels and there are no medium or bad levels, it is judged that the physical condition is good; When the heart rate, blood oxygen saturation, body temperature, and sleep are at medium levels and there are no bad levels, the physical condition is judged to be medium; When heart rate, blood oxygen saturation, body temperature and sleep are at bad levels, it is judged that the physical condition is poor.
[0034] Step 3: Adjust the indoor environment according to the user's mood and physical condition. The specific method is as follows: C1, all objects that need to be located indoors are positioned using WiFi, creating an indoor coordinate system, centrally controlling all temperature, humidity, ventilation, and fragrance equipment, numbering and locating all temperature, humidity, ventilation, and fragrance equipment and their vents; locating all user mobile phones; C2, the initial indoor temperature was adjusted to 22-23°C, the initial indoor humidity was adjusted to 45-50%RH, and the ventilation intensity was 1-2ACH; C3, when there is only one user in the room; ① When the user's physical condition is at an excellent level, the initial indoor temperature, humidity, and ventilation intensity remain unchanged; When users want to feel peaceful and happy, they don’t use fragrance; When the user is in an excited mood, use lavender fragrance; When the user is in an angry mood, use sandalwood fragrance; When the user is in a sad mood, use orange blossom fragrance; When users are anxious, they use clary sage fragrance.
[0035] ② When the user's physical condition is at a good level, based on the initial indoor temperature, humidity, and ventilation data, the temperature rises by 1°C, the humidity remains unchanged, and the ventilation intensity increases by 1ACH; no fragrance is used; ③ When the user's physical condition is at a moderate level, based on the initial indoor temperature, humidity, and ventilation data, the temperature is increased by 2°C, the humidity is increased by 5%RH, and the ventilation intensity is increased by 2ACH; no fragrance is used; ④ When the user's physical condition is at a poor level, based on the initial indoor temperature, humidity, and ventilation data, the temperature increases by 3°C, the humidity increases by 10%RH, and the ventilation intensity increases by 3ACH; no fragrance is used.
[0036] C4, when there are multiple users in the room; ① When all users are in excellent physical condition, the initial indoor temperature, humidity and ventilation intensity remain unchanged; Count the number of users with each emotion, and use the fragrance based on the emotion with the largest number of users, in the same way as ① in C3; when there are multiple emotions with the largest number of users, switch the corresponding fragrance type in turn every 2 hours, in the same way as ① in C3; ② When there are users in good, medium or poor physical condition in the room, do not use fragrance; When the worst physical condition among all users in the room is at a good level, adjust the humidity and ventilation intensity according to ② in C3; When the worst physical condition among all users in the room is at a medium level, adjust the humidity and ventilation intensity according to ③ in C3; When the worst physical condition among all users in the room is at the bad level, adjust the humidity and ventilation intensity according to ④ in C3; The circle with a diameter of 3m on the ground directly below each air outlet is used as the statistical area, and the physical condition of the users directly below each air outlet is counted: When the worst physical condition of the user under a certain air outlet is in good condition, the air outlet temperature of this air outlet increases by 1°C; When the worst physical condition of the users under a certain air outlet is at the intermediate level, the air outlet temperature of this air outlet increases by 2°C; When the worst physical condition among the users under a certain air outlet is at the bad level, the air outlet temperature of this air outlet increases by 3°C.
[0037] The principle of the present invention is: Convolutional neural network is a deep learning model that can effectively extract the spatial features of images.
[0038] The ReLU activation function f(z)=max(0,z). If z is greater than 0, the output is z itself. If z is less than or equal to 0, the output is 0. In deep learning, the ReLU activation function can accelerate the convergence of the model.
[0039] The Softmax activation function converts an H-dimensional vector into an H-dimensional probability distribution, which converts the output into the probability of each category.
[0040] The symbol °C refers to degrees Celsius; the symbol %RH refers to relative humidity, which expresses the amount of water vapor in the air as a percentage of the maximum amount of water vapor at the same temperature; the symbol ACH refers to air change rate, which expresses the number of times the indoor air is replaced by fresh air within one hour.
[0041] Aromatherapy is a therapy that uses the scent of plants to improve emotional states. Different aroma ingredients can evoke different emotional responses. When users feel excited, lavender can help soothe emotions and make it easier to relax. Sandalwood can help users calm anger and achieve inner peace. The scent of orange blossom can fight sadness and bring a sense of warmth. The scent of clary sage can help relieve anxiety.
[0042] The AI indoor environment adjustment system and method of the present invention have the following advantages: (1) The indoor environment is adjusted based on the user's mood and physical condition, which is a novel concept; (2) Create a fluctuation threshold formula and determination method when determining user emotions based on user voice information; use AI algorithms to solve image analysis problems when analyzing user emotions based on user facial expression images, and the method is reliable; (3) Indoor environment control equipment and air vents are numbered and located, and differentiated adjustments are made based on the user's mood, physical condition, and location, which is highly creative.
[0043] Other features and advantages of the present invention will be set forth in the following description or may be learned by practicing the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.
[0045] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION
[0046] The following will further describe in detail an AI indoor environment adjustment system and method of the present invention in conjunction with the accompanying drawings and embodiments.
[0047] The technical solution adopted by the present invention is an AI indoor environment adjustment system: It includes emotion analysis module, physical condition analysis module and indoor environment adjustment module.
[0048] The emotion analysis module is used to analyze the user's emotions in real time through the data collected by the mobile phone. The specific method is as follows: A1, collects text information input by users on mobile phones, including social interaction information, comment information, and status display information; A2, collects the voice information input by the user on the mobile phone, including speech speed, fundamental frequency, and volume; A3, collects the user’s facial expression information through the front camera of the mobile phone; A4, classifies emotions into multiple types, including peace, happiness, excitement, anger, sadness, and anxiety; A5, Analyze user emotions based on text information: ① Create an emotional text sample library, collect existing texts expressing emotions, and classify them according to the emotional types in A4; ② Based on the text input by the user on the mobile phone within 30 minutes, extract the text that is the same as the text in the emotional text sample library, classify it according to the emotional type in A4, and count the number of texts belonging to each emotional type; ③ Assume that the number of words entered by the user on the mobile phone within 30 minutes is m, and the number of words classified into a certain emotion type is n, then the score of this emotion is S 1i =100×n÷m; i=1, 2, 3, 4, 5, 6, S 11 , S 12 , S 13 , S 14 , S 15 , S 16 They represent the six emotion scores obtained through text information analysis: peace, happiness, excitement, anger, sadness, and anxiety.
[0049] A6, analyzing user emotions based on voice information: Calculate the speech rate, fundamental frequency, and volume based on the voice information input by the user on the mobile phone within 30 minutes; the unit of speech rate is SPM (Syllables Per Minute), the unit of fundamental frequency is Hz (Hertz), and the unit of volume is dB (decibel); Suppose the fluctuation threshold of speech rate, fundamental frequency and volume is Q=k×sqrt{[(x1-u) 2 +(x2-u) 2 +…+(x n -u) 2 ]÷n}, u is the average value of the data within 30 minutes, k is the coefficient, x j For each data point, j = 1, 2, ... n, sqrt{} represents the square root, and the coefficient k in the fluctuation threshold formula of speech rate, fundamental frequency, and volume is determined according to the respective test results; u=(x1+x2+…+x n )÷n, where n is the total number of data points; The fluctuation value of speech speed, fundamental frequency and volume P=x j -u; Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through voice information analysis are S 21 , S 22 , S 23 , S 24 , S 25 , S 26 ; ① When the fluctuation values of speech speed, fundamental frequency and volume are all less than their respective fluctuation thresholds, the peaceful emotion score S 21 =50, sad emotion score S 25 =50, other types of emotions scored as 0; ② When the fluctuation values of speech speed and volume are both less than their respective fluctuation thresholds, and the fluctuation value of fundamental frequency exceeds its fluctuation threshold, the pleasure score S 22 =50, anxiety score S 26 =50, other types of emotions scored as 0; ③ When the fluctuation values of speech speed, fundamental frequency and volume are all greater than their respective fluctuation thresholds, the excitement score S 23 =50, anger score S 24 =50, and other types of emotions scored as 0.
[0050] A7, analyzing user emotions based on user facial expression information: ① Collect facial expression image datasets and annotate the emotion types. The facial expression image dataset contains samples of peace, happiness, excitement, anger, sadness, and anxiety. Adjust the images in the samples to grayscale images of 48x48 pixels. ②Build a convolutional neural network model, Input layer: The input is processed into a 48x48 pixel grayscale image of facial expression; Convolutional layer: Multiple convolutional layers are used to extract features, the convolution kernel size is 3x3, and the ReLU activation function is used; Pooling layer: reduces feature dimensions, reduces computational effort, and increases model robustness; Fully connected layer: flattens the extracted features and performs sentiment classification; Output layer: Use the softmax activation function to output the probability distribution of each emotion.
[0051] ③Train the model, divide the data set into a training set and a validation set, and select the optimal parameters for the model through cross-validation.
[0052] ④ Emotion recognition: Use the trained model to recognize the user's facial expression image, output the corresponding emotion type and probability, multiply the probability by 100, and obtain a score of 100 for each emotion. Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through facial expression information analysis are S, 31 , S 32 , S 33 , S 34 , S 35 , S 36 .
[0053] A8: Comprehensively judge the user's emotions. The total score of each emotion is calculated as follows: ① The total score of peaceful mood is S 11 +S 21 +S 31 ; ②The total score of pleasant emotion is S 12 +S 22 +S 32 ; ③The total score of excitement is S 13 +S 23 +S 33 ; ④The total score of anger is S 14 +S 24 +S 34 ; ⑤The total score of sadness is S 15 +S 25 +S 35 ; ⑥The total score of anxiety is S 16 +S 26 +S 36 .
[0054] The emotion type with the highest comprehensive score is determined to be the user's emotion.
[0055] The physical condition analysis module is used to analyze the user's physical condition through the data collected by the mobile phone, specifically in the following ways: B1, the mobile phone collects the user's physical condition data from the health monitoring portable device worn by the user, including heart rate, blood oxygen saturation, body temperature, and sleep data.
[0056] B2, analyze heart rate data; The user's real-time resting heart rate is compared with the historical resting heart rate average value, and the excellent heart rate difference, good heart rate difference, medium heart rate difference, and poor heart rate difference are set. 0<excellent heart rate difference<good heart rate difference<medium heart rate difference<poor heart rate difference; When (historical resting heart rate average value - heart rate excellent difference) ≤ real-time resting heart rate ≤ (historical resting heart rate average value + heart rate excellent difference), the current heart rate is judged to be excellent; When (historical resting heart rate average value - good heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - excellent heart rate difference), or (historical resting heart rate average value + excellent heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + good heart rate difference), the current heart rate is determined to be good; When (historical resting heart rate average value - heart rate medium difference) ≤ real-time resting heart rate < (historical resting heart rate average value - heart rate good difference value), or (historical resting heart rate average value + heart rate good difference value) < real-time resting heart rate ≤ (historical resting heart rate average value + heart rate medium difference value), the current heart rate is determined to be medium; When (historical resting heart rate average value - poor heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - medium heart rate difference), or (historical resting heart rate average value + medium heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + poor heart rate difference), the current heart rate is determined to be poor.
[0057] B3, analyze blood oxygen saturation data: When 95%<blood oxygen saturation≤100%, it means that the blood oxygen content is excellent; When 90%<blood oxygen saturation≤95%, it means that the blood oxygen content is good; When 85%<blood oxygen saturation≤90%, it means that the blood oxygen content is moderate; When the blood oxygen saturation is ≤85%, it means that the blood oxygen content is poor.
[0058] B4, analyze body temperature data: The user's real-time body temperature is compared with the historical average body temperature, and the excellent temperature difference, good temperature difference, medium temperature difference, and poor temperature difference are set. 0<excellent temperature difference<good temperature difference<medium temperature difference<poor temperature difference; When (historical temperature average value - temperature difference) ≤ real-time temperature ≤ (historical temperature average value + temperature difference), the current temperature is considered excellent; When (historical temperature average value - good temperature difference) ≤ real-time temperature < (historical temperature average value - excellent temperature difference), or (historical temperature average value + excellent temperature difference) < real-time temperature ≤ (historical temperature average value + good temperature difference), the current temperature is considered good; When (historical temperature average value - medium temperature difference) ≤ real-time temperature < (historical temperature average value - good temperature difference), or (historical temperature average value + good temperature difference) < real-time temperature ≤ (historical temperature average value + medium temperature difference), the current temperature is determined to be medium; When (historical body temperature average value - abnormal body temperature difference) ≤ real-time body temperature < (historical body temperature average value - average body temperature difference), or (historical body temperature average value + average body temperature difference) < real-time body temperature ≤ (historical body temperature average value + abnormal body temperature difference), the current body temperature is determined to be abnormal.
[0059] B5, Analyze sleep data: Sleep efficiency V = actual sleep time ÷ time in bed × 100%; When V ≥ 90%, it is excellent sleep; When 85%≤V<90%, it is a good sleep; When 75%≤V<85%, it is moderate sleep; When V is less than 75%, it is poor sleep.
[0060] B6, classifies physical condition into four categories: excellent, good, moderate, and poor; When heart rate, blood oxygen saturation, body temperature and sleep all reach excellent levels, the physical condition is judged to be excellent; When the heart rate, blood oxygen saturation, body temperature, and sleep are at good levels and there are no medium or bad levels, the body is judged to be in good condition; When the heart rate, blood oxygen saturation, body temperature, and sleep are at medium levels and there are no bad levels, the physical condition is judged to be medium; When heart rate, blood oxygen saturation, body temperature and sleep are at bad levels, it is judged that the physical condition is poor.
[0061] The indoor environment adjustment module adjusts the indoor environment according to the user's mood and physical condition, specifically in the following ways: C1, all objects that need to be located indoors are positioned using WiFi to create an indoor coordinate system. The indoor environment adjustment module connects all temperature, humidity, ventilation, and fragrance control systems to number and locate all temperature, humidity, ventilation, and fragrance equipment and their vents; locate all user mobile phones; C2, the initial indoor temperature was adjusted to 22-23°C, the initial indoor humidity was adjusted to 45-50%RH, and the ventilation intensity was 1-2ACH; C3, when there is only one user in the room; ① When the user's physical condition is at an excellent level, the initial indoor temperature, humidity, and ventilation intensity remain unchanged; When users want to feel peaceful and happy, they don’t use fragrance; When the user is in an excited mood, use lavender fragrance; When the user is in an angry mood, use sandalwood fragrance; When the user is in a sad mood, use orange blossom fragrance; When users are anxious, they use clary sage fragrance.
[0062] ② When the user's physical condition is at a good level, based on the initial indoor temperature, humidity, and ventilation data, the temperature rises by 1°C, the humidity remains unchanged, and the ventilation intensity increases by 1ACH; no fragrance is used; ③ When the user's physical condition is at a moderate level, based on the initial indoor temperature, humidity, and ventilation data, the temperature is increased by 2°C, the humidity is increased by 5%RH, and the ventilation intensity is increased by 2ACH; no fragrance is used; ④ When the user's physical condition is at a poor level, based on the initial indoor temperature, humidity, and ventilation data, the temperature increases by 3°C, the humidity increases by 10%RH, and the ventilation intensity increases by 3ACH; no fragrance is used.
[0063] C4, when there are multiple users in the room; ① When all users are in excellent physical condition, the initial indoor temperature, humidity and ventilation intensity remain unchanged; Count the number of users with each emotion, and use the fragrance based on the emotion with the largest number of users, in the same way as ① in C3; when there are multiple emotions with the largest number of users, switch the corresponding fragrance type in turn every 2 hours, in the same way as ① in C3; ② When there are users in good, medium or poor physical condition in the room, do not use fragrance; When the worst physical condition among all users in the room is at a good level, adjust the humidity and ventilation intensity according to ② in C3; When the worst physical condition among all users in the room is at a medium level, adjust the humidity and ventilation intensity according to ③ in C3; When the worst physical condition among all users in the room is at the bad level, adjust the humidity and ventilation intensity according to ④ in C3; The circle with a diameter of 3m on the ground directly below each air outlet is used as the statistical area, and the physical condition of the users directly below each air outlet is counted: When the worst physical condition of the user under a certain air outlet is in good condition, the air outlet temperature of this air outlet increases by 1°C; When the worst physical condition of the users under a certain air outlet is at the intermediate level, the air outlet temperature of this air outlet increases by 2°C; When the worst physical condition among the users under a certain air outlet is at the bad level, the air outlet temperature of this air outlet increases by 3°C.
[0064] like Figure 1 As shown, an AI indoor environment adjustment method: Step 1: Analyze the user's emotions in real time through the data collected by the mobile phone. The specific method is as follows: A1, collects text information input by users on mobile phones, including social interaction information, comment information, and status display information; A2, collects the voice information input by the user on the mobile phone, including speech speed, fundamental frequency, and volume; A3, collects the user’s facial expression information through the front camera of the mobile phone; A4, classifies emotions into multiple types, including peace, happiness, excitement, anger, sadness, and anxiety; A5, Analyze user emotions based on text information: ① Create an emotional text sample library, collect existing texts expressing emotions, and classify them according to the emotional types in A4; ② Based on the text input by the user on the mobile phone within 30 minutes, extract the text that is the same as the text in the emotional text sample library, classify it according to the emotional type in A4, and count the number of texts belonging to each emotional type; ③ Assume that the number of words entered by the user on the mobile phone within 30 minutes is m, and the number of words classified into a certain emotion type is n, then the score of this emotion is S 1i =100×n÷m; i=1, 2, 3, 4, 5, 6, S 11 , S 12 , S 13 , S 14 , S 15 , S 16 They represent the six emotion scores obtained through text information analysis: peace, happiness, excitement, anger, sadness, and anxiety.
[0065] A6, analyzing user emotions based on voice information: Calculate the speech rate, fundamental frequency, and volume based on the voice information input by the user on the mobile phone within 30 minutes; the unit of speech rate is SPM (Syllables Per Minute), the unit of fundamental frequency is Hz (Hertz), and the unit of volume is dB (decibel); Suppose the fluctuation threshold of speech rate, fundamental frequency and volume is Q=k×sqrt{[(x1-u)2 +(x2-u) 2 +…+(x n -u) 2 ]÷n}, u is the average value of the data within 30 minutes, k is the coefficient, x j For each data point, j = 1, 2, ... n, sqrt{} represents the square root, and the coefficient k in the fluctuation threshold formula of speech rate, fundamental frequency, and volume is determined according to the respective test results; u=(x1+x2+…+x n )÷n, where n is the total number of data points; The fluctuation value of speech speed, fundamental frequency and volume P=x j -u; Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through voice information analysis are S 21 , S 22 , S 23 , S 24 , S 25 , S 26 ; ① When the fluctuation values of speech speed, fundamental frequency and volume are all less than their respective fluctuation thresholds, the peaceful emotion score S 21 =50, sad emotion score S 25 =50, other types of emotions scored as 0; ② When the fluctuation values of speech speed and volume are both less than their respective fluctuation thresholds, and the fluctuation value of fundamental frequency exceeds its fluctuation threshold, the pleasure score S 22 =50, anxiety score S 26 =50, other types of emotions scored as 0; ③ When the fluctuation values of speech speed, fundamental frequency and volume are all greater than their respective fluctuation thresholds, the excitement score S 23 =50, anger score S 24 =50, and other types of emotions scored as 0.
[0066] A7, analyzing user emotions based on user facial expression information: ① Collect facial expression image datasets and annotate the emotion types. The facial expression image dataset contains samples of peace, happiness, excitement, anger, sadness, and anxiety. Adjust the images in the samples to grayscale images of 48x48 pixels. ②Build a convolutional neural network model, Input layer: The input is processed into a 48x48 pixel grayscale image of facial expression; Convolutional layer: Multiple convolutional layers are used to extract features, the convolution kernel size is 3x3, and the ReLU activation function is used; Pooling layer: reduces feature dimensions, reduces computational effort, and increases model robustness; Fully connected layer: flattens the extracted features and performs sentiment classification; Output layer: Use the softmax activation function to output the probability distribution of each emotion.
[0067] ③Train the model, divide the data set into a training set and a validation set, and select the optimal parameters for the model through cross-validation.
[0068] ④ Emotion recognition: Use the trained model to recognize the user's facial expression image, output the corresponding emotion type and probability, multiply the probability by 100, and obtain a score of 100 for each emotion. Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through facial expression information analysis are S, 31 , S 32 , S 33 , S 34 , S 35 , S 36 .
[0069] A8: Comprehensively judge the user's emotions. The total score of each emotion is calculated as follows: ① The total score of peaceful mood is S 11 +S 21 +S 31 ; ②The total score of pleasant emotion is S 12 +S 22 +S 32 ; ③The total score of excitement is S 13 +S 23 +S 33 ; ④The total score of anger is S 14 +S 24 +S 34 ; ⑤The total score of sadness is S 15 +S 25 +S 35 ; ⑥The total score of anxiety is S 16 +S 26 +S 36 .
[0070] The emotion type with the highest comprehensive score is determined to be the user's emotion.
[0071] Step 2: Analyze the user's physical condition through the data collected by the mobile phone. The specific method is as follows: B1, the mobile phone collects the user's physical condition data from the health monitoring portable device worn by the user, including heart rate, blood oxygen saturation, body temperature, and sleep data.
[0072] B2, analyze heart rate data; The user's real-time resting heart rate is compared with the historical resting heart rate average value, and the excellent heart rate difference, good heart rate difference, medium heart rate difference, and poor heart rate difference are set. 0<excellent heart rate difference<good heart rate difference<medium heart rate difference<poor heart rate difference; When (historical resting heart rate average value - heart rate excellent difference) ≤ real-time resting heart rate ≤ (historical resting heart rate average value + heart rate excellent difference), the current heart rate is judged to be excellent; When (historical resting heart rate average value - good heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - excellent heart rate difference), or (historical resting heart rate average value + excellent heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + good heart rate difference), the current heart rate is determined to be good; When (historical resting heart rate average value - heart rate medium difference) ≤ real-time resting heart rate < (historical resting heart rate average value - heart rate good difference value), or (historical resting heart rate average value + heart rate good difference value) < real-time resting heart rate ≤ (historical resting heart rate average value + heart rate medium difference value), the current heart rate is determined to be medium; When (historical resting heart rate average value - poor heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - medium heart rate difference), or (historical resting heart rate average value + medium heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + poor heart rate difference), the current heart rate is determined to be poor.
[0073] B3, analyze blood oxygen saturation data: When 95%<blood oxygen saturation≤100%, it means that the blood oxygen content is excellent; When 90%<blood oxygen saturation≤95%, it means that the blood oxygen content is good; When 85%<blood oxygen saturation≤90%, it means that the blood oxygen content is moderate; When the blood oxygen saturation is ≤85%, it means that the blood oxygen content is poor.
[0074] B4, analyze body temperature data: The user's real-time body temperature is compared with the historical average body temperature, and the excellent body temperature difference, good body temperature difference, medium body temperature difference, and poor body temperature difference are set, 0 < excellent body temperature difference < good body temperature difference < medium body temperature difference < poor body temperature difference;.
[0075] When (historical temperature average value - temperature difference) ≤ real-time temperature ≤ (historical temperature average value + temperature difference), the current temperature is considered excellent; When (historical temperature average value - good temperature difference) ≤ real-time temperature < (historical temperature average value - excellent temperature difference), or (historical temperature average value + excellent temperature difference) < real-time temperature ≤ (historical temperature average value + good temperature difference), the current temperature is considered good; When (historical temperature average value - medium temperature difference) ≤ real-time temperature < (historical temperature average value - good temperature difference), or (historical temperature average value + good temperature difference) < real-time temperature ≤ (historical temperature average value + medium temperature difference), the current temperature is determined to be medium; When (historical body temperature average value - abnormal body temperature difference) ≤ real-time body temperature < (historical body temperature average value - average body temperature difference), or (historical body temperature average value + average body temperature difference) < real-time body temperature ≤ (historical body temperature average value + abnormal body temperature difference), the current body temperature is determined to be abnormal.
[0076] B5, Analyze sleep data: Sleep efficiency V = actual sleep time ÷ time in bed × 100%; When V ≥ 90%, it is excellent sleep; When 85%≤V<90%, it is a good sleep; When 75%≤V<85%, it is moderate sleep; When V is less than 75%, it is poor sleep.
[0077] B6, classifies physical condition into four categories: excellent, good, moderate, and poor; When heart rate, blood oxygen saturation, body temperature and sleep all reach excellent levels, the physical condition is judged to be excellent; When the heart rate, blood oxygen saturation, body temperature, and sleep are at good levels and there are no medium or bad levels, it is judged that the physical condition is good; When the heart rate, blood oxygen saturation, body temperature, and sleep are at medium levels and there are no bad levels, the physical condition is judged to be medium; When heart rate, blood oxygen saturation, body temperature and sleep are at bad levels, it is judged that the physical condition is poor.
[0078] Step 3: Adjust the indoor environment according to the user's mood and physical condition. The specific method is as follows: C1, all objects that need to be located indoors are positioned using WiFi, creating an indoor coordinate system, centrally controlling all temperature, humidity, ventilation, and fragrance equipment, numbering and locating all temperature, humidity, ventilation, and fragrance equipment and their vents; locating all user mobile phones; C2, the initial indoor temperature was adjusted to 22-23°C, the initial indoor humidity was adjusted to 45-50%RH, and the ventilation intensity was 1-2ACH; C3, when there is only one user in the room; ① When the user's physical condition is at an excellent level, the initial indoor temperature, humidity, and ventilation intensity remain unchanged; When users want to feel peaceful and happy, they don’t use fragrance; When the user is in an excited mood, use lavender fragrance; When the user is in an angry mood, use sandalwood fragrance; When the user is in a sad mood, use orange blossom fragrance; When users are anxious, they use clary sage fragrance.
[0079] ② When the user's physical condition is at a good level, based on the initial indoor temperature, humidity, and ventilation data, the temperature rises by 1°C, the humidity remains unchanged, and the ventilation intensity increases by 1ACH; no fragrance is used; ③ When the user's physical condition is at a moderate level, based on the initial indoor temperature, humidity, and ventilation data, the temperature is increased by 2°C, the humidity is increased by 5%RH, and the ventilation intensity is increased by 2ACH; no fragrance is used; ④ When the user's physical condition is at a poor level, based on the initial indoor temperature, humidity, and ventilation data, the temperature increases by 3°C, the humidity increases by 10%RH, and the ventilation intensity increases by 3ACH; no fragrance is used.
[0080] C4, when there are multiple users in the room; ① When all users are in excellent physical condition, the initial indoor temperature, humidity and ventilation intensity remain unchanged; Count the number of users with each emotion, and use the fragrance based on the emotion with the largest number of users, in the same way as ① in C3; when there are multiple emotions with the largest number of users, switch the corresponding fragrance type in turn every 2 hours, in the same way as ① in C3; ② When there are users in good, medium or poor physical condition in the room, do not use fragrance; When the worst physical condition among all users in the room is at a good level, adjust the humidity and ventilation intensity according to ② in C3; When the worst physical condition among all users in the room is at a medium level, adjust the humidity and ventilation intensity according to ③ in C3; When the worst physical condition among all users in the room is at the bad level, adjust the humidity and ventilation intensity according to ④ in C3; The circle with a diameter of 3m on the ground directly below each air outlet is used as the statistical area, and the physical condition of the users directly below each air outlet is counted: When the worst physical condition of the user under a certain air outlet is in good condition, the air outlet temperature of this air outlet increases by 1°C; When the worst physical condition of the users under a certain air outlet is at the intermediate level, the air outlet temperature of this air outlet increases by 2°C; When the worst physical condition among the users under a certain air outlet is at the bad level, the air outlet temperature of this air outlet increases by 3°C.
[0081] Convolutional neural network is a deep learning model that can effectively extract the spatial features of images. It belongs to the field of AI algorithms. The 3x3 convolution kernel represents a small matrix with a size of 3 rows and 3 columns.
[0082] The ReLU activation function f(z)=max(0,z). If z is greater than 0, the output is z itself. If z is less than or equal to 0, the output is 0. In deep learning, the ReLU activation function can accelerate the convergence of the model.
[0083] The Softmax activation function converts an H-dimensional vector into an H-dimensional probability distribution, which converts the output into the probability of each category.
[0084] The symbol °C refers to degrees Celsius; the symbol %RH refers to relative humidity, which expresses the amount of water vapor in the air as a percentage of the maximum amount of water vapor at the same temperature; the symbol ACH refers to air change rate, which expresses the number of times the indoor air is replaced by fresh air within one hour.
[0085] Aromatherapy is a therapy that uses the scent of plants to improve emotional states. Different aroma ingredients can evoke different emotional responses. When users feel excited, lavender can help soothe emotions and make it easier to relax. Sandalwood can help users calm anger and achieve inner peace. The scent of orange blossom can fight sadness and bring a sense of warmth. The scent of clary sage can help relieve anxiety.
[0086] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.
Claims
1. An AI indoor environment adjustment system, characterized in that: Including emotion analysis module, physical condition analysis module, and indoor environment adjustment module; The emotion analysis module is used to analyze the user's emotions in real time through the data collected by the mobile phone. The specific method is as follows: A1, collects text information input by users on mobile phones, including social interaction information, comment information, and status display information; A2, collects the voice information input by the user on the mobile phone, including speech speed, fundamental frequency, and volume; A3, collects the user’s facial expression information through the front camera of the mobile phone; A4, classifies emotions into multiple types, including peace, happiness, excitement, anger, sadness, and anxiety; A5, Analyze user emotions based on text information: ① Create an emotional text sample library, collect existing texts expressing emotions, and classify them according to the emotional types in A4; ② Based on the text input by the user on the mobile phone within 30 minutes, extract the text that is the same as the text in the emotional text sample library, classify it according to the emotional type in A4, and count the number of texts belonging to each emotional type; ③ Assume that the number of words entered by the user on the mobile phone within 30 minutes is m, and the number of words classified into a certain emotion type is n, then the score of this emotion is S 1i =100×n÷m; i=1, 2, 3, 4, 5, 6, S 11 , S 12 , S 13 , S 14 , S 15 , S 16 They represent the six emotional scores obtained through text information analysis: peace, happiness, excitement, anger, sadness, and anxiety; A6, analyzing user emotions based on voice information: Calculate the speech rate, fundamental frequency, and volume based on the voice information input by the user on the mobile phone within 30 minutes; the unit of speech rate is syllable / second, the unit of fundamental frequency is Hz, and the unit of volume is decibel; Suppose the fluctuation threshold of speech rate, fundamental frequency and volume is Q=k×sqrt{[(x1-u) 2 +(x2-u) 2 +…+(x n -u) 2 ]÷n}, u is the average value of the data within 30 minutes, k is the coefficient, x j For each data point, j = 1, 2, ... n, sqrt{} represents the square root, and the coefficient k in the fluctuation threshold formula of speech rate, fundamental frequency, and volume is determined according to the respective test results; u=(x1+x2+…+x n )÷n, where n is the total number of data points; The fluctuation value of speech speed, fundamental frequency and volume P=x j -u; Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through voice information analysis are S 21 , S 22 , S 23 , S 24 , S 25 , S 26 ; ① When the fluctuation values of speech speed, fundamental frequency and volume are all less than their respective fluctuation thresholds, the peaceful emotion score S 21 =50, sad emotion score S 25 =50, other types of emotions scored as 0; ② When the fluctuation values of speech speed and volume are both less than their respective fluctuation thresholds, and the fluctuation value of fundamental frequency exceeds its fluctuation threshold, the pleasure score S 22 =50, anxiety score S 26 =50, other types of emotions scored as 0; ③ When the fluctuation values of speech speed, fundamental frequency and volume are all greater than their respective fluctuation thresholds, the excitement score S 23 =50, anger score S 24 =50, other types of emotions scored as 0; A7, analyzing user emotions based on user facial expression information: ① Collect facial expression image datasets and annotate the emotion types. The facial expression image dataset contains samples of peace, happiness, excitement, anger, sadness, and anxiety. Adjust the images in the samples to grayscale images of 48x48 pixels. ②Build a convolutional neural network model, Input layer: The input is processed into a 48x48 pixel grayscale image of facial expression; Convolutional layer: Multiple convolutional layers are used to extract features, the convolution kernel size is 3x3, and the ReLU activation function is used; Pooling layer: reduces feature dimensions, reduces computational effort, and increases model robustness; Fully connected layer: flattens the extracted features and performs sentiment classification; Output layer: Use softmax activation function to output the probability distribution of each emotion; ③Train the model, divide the data set into a training set and a validation set, and select the optimal parameters for the model through cross-validation; ④ Emotion recognition: Use the trained model to recognize the user's facial expression image, output the corresponding emotion type and probability, multiply the probability by 100, and obtain a score of 100 for each emotion. Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through facial expression information analysis are S, 31 , S 32 , S 33 , S 34 , S 35 , S 36 ; A8: Comprehensively judge the user's emotions. The total score of each emotion is calculated as follows: ①The total score of peaceful emotion is S 11 +S 21 +S 31 ; ②The total score of pleasant emotion is S 12 +S 22 +S 32 ; ③The total score of excitement is S 13 +S 23 +S 33 ; ④The total score of anger emotion is S 14 +S 24 +S 34 ; ⑤The total score of sadness is S 15 +S 25 +S 35 ; ⑥The total score of anxiety is S 16 +S 26 +S 36 ; The emotion type with the highest comprehensive score is determined to be the user's emotion; The physical condition analysis module is used to analyze the user's physical condition through the data collected by the mobile phone, specifically in the following ways: B1, the mobile phone collects the user's physical condition data from the health monitoring portable device worn by the user, including heart rate, blood oxygen saturation, body temperature, and sleep data; B2, analyze heart rate data; The user's real-time resting heart rate is compared with the historical resting heart rate average value, and the excellent heart rate difference, good heart rate difference, medium heart rate difference, and poor heart rate difference are set. 0<excellent heart rate difference<good heart rate difference<medium heart rate difference<poor heart rate difference; When (historical resting heart rate average value - heart rate excellent difference) ≤ real-time resting heart rate ≤ (historical resting heart rate average value + heart rate excellent difference), the current heart rate is judged to be excellent; When (historical resting heart rate average value - good heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - excellent heart rate difference), or (historical resting heart rate average value + excellent heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + good heart rate difference), the current heart rate is determined to be good; When (historical resting heart rate average value - heart rate medium difference) ≤ real-time resting heart rate < (historical resting heart rate average value - heart rate good difference value), or (historical resting heart rate average value + heart rate good difference value) < real-time resting heart rate ≤ (historical resting heart rate average value + heart rate medium difference value), the current heart rate is determined to be medium; When (historical resting heart rate average value - poor heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - moderate heart rate difference), or (historical resting heart rate average value + moderate heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + poor heart rate difference), the current heart rate is determined to be poor; B3, analyze blood oxygen saturation data: When 95%<blood oxygen saturation≤100%, it means that the blood oxygen content is excellent; When 90%<blood oxygen saturation≤95%, it means that the blood oxygen content is good; When 85%<blood oxygen saturation≤90%, it means that the blood oxygen content is moderate; When the blood oxygen saturation is ≤85%, it means that the blood oxygen content is poor; B4, analyze body temperature data: The user's real-time body temperature is compared with the historical average body temperature, and the excellent temperature difference, good temperature difference, medium temperature difference, and poor temperature difference are set. 0<excellent temperature difference<good temperature difference<medium temperature difference<poor temperature difference; When (historical temperature average value - temperature difference) ≤ real-time temperature ≤ (historical temperature average value + temperature difference), the current temperature is considered excellent; When (historical temperature average value - good temperature difference) ≤ real-time temperature < (historical temperature average value - excellent temperature difference), or (historical temperature average value + excellent temperature difference) < real-time temperature ≤ (historical temperature average value + good temperature difference), the current temperature is considered good; When (historical temperature average value - medium temperature difference) ≤ real-time temperature < (historical temperature average value - good temperature difference), or (historical temperature average value + good temperature difference) < real-time temperature ≤ (historical temperature average value + medium temperature difference), the current temperature is determined to be medium; When (historical temperature average value - temperature difference) ≤ real-time temperature < (historical temperature average value - temperature difference), or (historical temperature average value + temperature difference) < real-time temperature ≤ (historical temperature average value + temperature difference), the current temperature is considered to be abnormal. B5, Analyze sleep data: Sleep efficiency V = actual sleep time ÷ time in bed × 100%; When V ≥ 90%, it is excellent sleep; When 85%≤V<90%, it is a good sleep; When 75%≤V<85%, it is moderate sleep; When V < 75%, it is poor sleep; B6, classifies physical condition into four categories: excellent, good, moderate, and poor; When heart rate, blood oxygen saturation, body temperature and sleep all reach excellent levels, the physical condition is judged to be excellent; When the heart rate, blood oxygen saturation, body temperature, and sleep are at good levels and there are no medium or bad levels, it is judged that the physical condition is good; When the heart rate, blood oxygen saturation, body temperature, and sleep are at medium levels and there are no bad levels, the physical condition is judged to be medium; When heart rate, blood oxygen saturation, body temperature, and sleep are at bad levels, it is judged that the physical condition is bad; The indoor environment adjustment module adjusts the indoor environment according to the user's mood and physical condition, specifically in the following ways: C1, all objects that need to be located indoors are positioned using WiFi to create an indoor coordinate system. The indoor environment adjustment module connects all temperature, humidity, ventilation, and fragrance control systems to number and locate all temperature, humidity, ventilation, and fragrance equipment and their vents; locate all user mobile phones; C2, the initial indoor temperature was adjusted to 22-23°C, the initial indoor humidity was adjusted to 45-50%RH, and the ventilation intensity was 1-2ACH; C3, when there is only one user in the room; ① When the user's physical condition is at an excellent level, the initial indoor temperature, humidity, and ventilation intensity remain unchanged; When users want to feel peaceful and happy, they don’t use fragrance; When the user is in an excited mood, use lavender fragrance; When the user is in an angry mood, use sandalwood fragrance; When the user is in a sad mood, use orange blossom fragrance; When users are anxious, they use clary sage fragrance; ② When the user's physical condition is at a good level, based on the initial indoor temperature, humidity, and ventilation data, the temperature rises by 1°C, the humidity remains unchanged, and the ventilation intensity increases by 1ACH; no fragrance is used; ③ When the user's physical condition is at a moderate level, based on the initial indoor temperature, humidity, and ventilation data, the temperature is increased by 2°C, the humidity is increased by 5%RH, and the ventilation intensity is increased by 2ACH; no fragrance is used; ④ When the user's physical condition is at a poor level, based on the initial indoor temperature, humidity, and ventilation data, the temperature is increased by 3°C, the humidity is increased by 10%RH, and the ventilation intensity is increased by 3ACH; no fragrance is used; C4, when there are multiple users in the room; ① When all users are in excellent physical condition, the initial indoor temperature, humidity and ventilation intensity remain unchanged; Count the number of users with each emotion, and use the fragrance based on the emotion with the largest number of users, in the same way as ① in C3; when there are multiple emotions with the largest number of users, switch the corresponding fragrance type in turn every 2 hours, in the same way as ① in C3; ② When there are users in good, medium or poor physical condition in the room, do not use fragrance; When the worst physical condition among all users in the room is at a good level, adjust the humidity and ventilation intensity according to ② in C3; When the worst physical condition among all users in the room is at a medium level, adjust the humidity and ventilation intensity according to ③ in C3; When the worst physical condition among all users in the room is at the bad level, adjust the humidity and ventilation intensity according to ④ in C3; The circle with a diameter of 3m on the ground directly below each air outlet is used as the statistical area, and the physical condition of the users directly below each air outlet is counted: When the worst physical condition of the user under a certain air outlet is in good condition, the air outlet temperature of this air outlet increases by 1°C; When the worst physical condition of the users under a certain air outlet is at the intermediate level, the air outlet temperature of this air outlet increases by 2°C; When the worst physical condition among the users under a certain air outlet is at the bad level, the air outlet temperature of this air outlet increases by 3°C.
2. An AI indoor environment adjustment method, characterized in that: Step 1: Analyze the user's emotions in real time through the data collected by the mobile phone. The specific method is as follows: A1, collects text information input by users on mobile phones, including social interaction information, comment information, and status display information; A2, collects the voice information input by the user on the mobile phone, including speech speed, fundamental frequency, and volume; A3, collects the user’s facial expression information through the front camera of the mobile phone; A4, classifies emotions into multiple types, including peace, happiness, excitement, anger, sadness, and anxiety; A5, Analyze user emotions based on text information: ① Create an emotional text sample library, collect existing texts expressing emotions, and classify them according to the emotional types in A4; ② Based on the text input by the user on the mobile phone within 30 minutes, extract the text that is the same as the text in the emotional text sample library, classify it according to the emotional type in A4, and count the number of texts belonging to each emotional type; ③ Assume that the number of words entered by the user on the mobile phone within 30 minutes is m, and the number of words classified into a certain emotion type is n, then the score of this emotion is S 1i =100×n÷m; i=1, 2, 3, 4, 5, 6, S 11 , S 12 , S 13 , S 14 , S 15 , S 16 They represent the six emotional scores obtained through text information analysis: peace, happiness, excitement, anger, sadness, and anxiety; A6, analyzing user emotions based on voice information: Calculate the speech rate, fundamental frequency, and volume based on the voice information input by the user on the mobile phone within 30 minutes; the unit of speech rate is syllable / second, the unit of fundamental frequency is Hz, and the unit of volume is decibel; Suppose the fluctuation threshold of speech rate, fundamental frequency and volume is Q=k×sqrt{[(x1-u) 2 +(x2-u) 2 +…+(x n -u) 2 ]÷n}, u is the average value of the data within 30 minutes, k is the coefficient, x j For each data point, j = 1, 2, ... n, sqrt{} represents the square root, and the coefficient k in the fluctuation threshold formula of speech rate, fundamental frequency, and volume is determined according to the respective test results; u=(x1+x2+…+x n )÷n, where n is the total number of data points; The fluctuation value of speech speed, fundamental frequency and volume P=x j -u; Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through voice information analysis are S 21 , S 22 , S 23 , S 24 , S 25 , S 26 ; ① When the fluctuation values of speech speed, fundamental frequency and volume are all less than their respective fluctuation thresholds, the peaceful emotion score S 21 =50, sad emotion score S 25 =50, other types of emotions scored as 0; ② When the fluctuation values of speech speed and volume are both less than their respective fluctuation thresholds, and the fluctuation value of fundamental frequency exceeds its fluctuation threshold, the pleasure score S 22 =50, anxiety score S 26 =50, other types of emotions scored as 0; ③ When the fluctuation values of speech speed, fundamental frequency and volume are all greater than their respective fluctuation thresholds, the excitement score S 23 =50, anger score S 24 =50, other types of emotions scored as 0; A7, analyzing user emotions based on user facial expression information: ① Collect facial expression image datasets and annotate the emotion types. The facial expression image dataset contains samples of peace, happiness, excitement, anger, sadness, and anxiety. Adjust the images in the samples to grayscale images of 48x48 pixels. ②Build a convolutional neural network model, Input layer: The input is processed into a 48x48 pixel grayscale image of facial expression; Convolutional layer: Multiple convolutional layers are used to extract features, the convolution kernel size is 3x3, and the ReLU activation function is used; Pooling layer: reduces feature dimensions, reduces computational effort, and increases model robustness; Fully connected layer: flattens the extracted features and performs sentiment classification; Output layer: Use softmax activation function to output the probability distribution of each emotion; ③Train the model, divide the data set into a training set and a validation set, and select the optimal parameters for the model through cross-validation; ④ Emotion recognition: Use the trained model to recognize the user's facial expression image, output the corresponding emotion type and probability, multiply the probability by 100, and obtain a score of 100 for each emotion. Suppose the scores of peace, happiness, excitement, anger, sadness, and anxiety obtained through facial expression information analysis are S, 31 , S 32 , S 33 , S 34 , S 35 , S 36 ; A8: Comprehensively judge the user's emotions. The total score of each emotion is calculated as follows: ①The total score of peaceful emotion is S 11 +S 21 +S 31 ; ②The total score of pleasant emotion is S 12 +S 22 +S 32 ; ③The total score of excitement is S 13 +S 23 +S 33 ; ④The total score of anger emotion is S 14 +S 24 +S 34 ; ⑤The total score of sadness is S 15 +S 25 +S 35 ; ⑥The total score of anxiety is S 16 +S 26 +S 36 ; The emotion type with the highest comprehensive score is determined to be the user's emotion; Step 2: Analyze the user's physical condition through the data collected by the mobile phone. The specific method is as follows: B1, the mobile phone collects the user's physical condition data from the health monitoring portable device worn by the user, including heart rate, blood oxygen saturation, body temperature, and sleep data; B2, analyze heart rate data; The user's real-time resting heart rate is compared with the historical resting heart rate average value, and the excellent heart rate difference, good heart rate difference, medium heart rate difference, and poor heart rate difference are set. 0<excellent heart rate difference<good heart rate difference<medium heart rate difference<poor heart rate difference; When (historical resting heart rate average value - heart rate excellent difference) ≤ real-time resting heart rate ≤ (historical resting heart rate average value + heart rate excellent difference), the current heart rate is judged to be excellent; When (historical resting heart rate average value - good heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - excellent heart rate difference), or (historical resting heart rate average value + excellent heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + good heart rate difference), the current heart rate is determined to be good; When (historical resting heart rate average value - heart rate medium difference) ≤ real-time resting heart rate < (historical resting heart rate average value - heart rate good difference value), or (historical resting heart rate average value + heart rate good difference value) < real-time resting heart rate ≤ (historical resting heart rate average value + heart rate medium difference value), the current heart rate is determined to be medium; When (historical resting heart rate average value - poor heart rate difference) ≤ real-time resting heart rate < (historical resting heart rate average value - moderate heart rate difference), or (historical resting heart rate average value + moderate heart rate difference) < real-time resting heart rate ≤ (historical resting heart rate average value + poor heart rate difference), the current heart rate is determined to be poor; B3, analyze blood oxygen saturation data: When 95%<blood oxygen saturation≤100%, it means that the blood oxygen content is excellent; When 90%<blood oxygen saturation≤95%, it means that the blood oxygen content is good; When 85%<blood oxygen saturation≤90%, it means that the blood oxygen content is moderate; When the blood oxygen saturation is ≤85%, it means that the blood oxygen content is poor; B4, analyze body temperature data: The user's real-time body temperature is compared with the historical average body temperature, and the excellent temperature difference, good temperature difference, medium temperature difference, and poor temperature difference are set. 0<excellent temperature difference<good temperature difference<medium temperature difference<poor temperature difference; When (historical temperature average value - temperature difference) ≤ real-time temperature ≤ (historical temperature average value + temperature difference), the current temperature is considered excellent; When (historical temperature average value - good temperature difference) ≤ real-time temperature < (historical temperature average value - excellent temperature difference), or (historical temperature average value + excellent temperature difference) < real-time temperature ≤ (historical temperature average value + good temperature difference), the current temperature is considered good; When (historical temperature average value - medium temperature difference) ≤ real-time temperature < (historical temperature average value - good temperature difference), or (historical temperature average value + good temperature difference) < real-time temperature ≤ (historical temperature average value + medium temperature difference), the current temperature is determined to be medium; When (historical temperature average value - temperature difference) ≤ real-time temperature < (historical temperature average value - temperature difference), or (historical temperature average value + temperature difference) < real-time temperature ≤ (historical temperature average value + temperature difference), the current temperature is considered to be abnormal. B5, Analyze sleep data: Sleep efficiency V = actual sleep time ÷ time in bed × 100%; When V ≥ 90%, it is excellent sleep; When 85%≤V<90%, it is a good sleep; When 75%≤V<85%, it is moderate sleep; When V < 75%, it is poor sleep; B6, classifies physical condition into four categories: excellent, good, moderate, and poor; When heart rate, blood oxygen saturation, body temperature and sleep all reach excellent levels, the physical condition is judged to be excellent; When the heart rate, blood oxygen saturation, body temperature, and sleep are at good levels and there are no medium or bad levels, it is judged that the physical condition is good; When the heart rate, blood oxygen saturation, body temperature, and sleep are at medium levels and there are no bad levels, the physical condition is judged to be medium; When heart rate, blood oxygen saturation, body temperature, and sleep are at bad levels, it is judged that the physical condition is bad; Step 3: Adjust the indoor environment according to the user's mood and physical condition. The specific method is as follows: C1, all objects that need to be located indoors are positioned using WiFi, creating an indoor coordinate system, centrally controlling all temperature, humidity, ventilation, and fragrance equipment, numbering and locating all temperature, humidity, ventilation, and fragrance equipment and their vents; locating all user mobile phones; C2, the initial indoor temperature was adjusted to 22-23°C, the initial indoor humidity was adjusted to 45-50%RH, and the ventilation intensity was 1-2ACH; C3, when there is only one user in the room; ① When the user's physical condition is at an excellent level, the initial indoor temperature, humidity, and ventilation intensity remain unchanged; When users want to feel peaceful and happy, they don’t use fragrance; When the user is in an excited mood, use lavender fragrance; When the user is in an angry mood, use sandalwood fragrance; When the user is in a sad mood, use orange blossom fragrance; When users are anxious, they use clary sage fragrance; ② When the user's physical condition is at a good level, based on the initial indoor temperature, humidity, and ventilation data, the temperature rises by 1°C, the humidity remains unchanged, and the ventilation intensity increases by 1ACH; no fragrance is used; ③ When the user's physical condition is at a moderate level, based on the initial indoor temperature, humidity, and ventilation data, the temperature is increased by 2°C, the humidity is increased by 5%RH, and the ventilation intensity is increased by 2ACH; no fragrance is used; ④ When the user's physical condition is at a poor level, based on the initial indoor temperature, humidity, and ventilation data, the temperature is increased by 3°C, the humidity is increased by 10%RH, and the ventilation intensity is increased by 3ACH; no fragrance is used; C4, when there are multiple users in the room; ① When all users are in excellent physical condition, the initial indoor temperature, humidity and ventilation intensity remain unchanged; Count the number of users with each emotion, and use the fragrance based on the emotion with the largest number of users, in the same way as ① in C3; when there are multiple emotions with the largest number of users, switch the corresponding fragrance type in turn every 2 hours, in the same way as ① in C3; ② When there are users in good, medium or poor physical condition in the room, do not use fragrance; When the worst physical condition among all users in the room is at a good level, adjust the humidity and ventilation intensity according to ② in C3; When the worst physical condition among all users in the room is at a medium level, adjust the humidity and ventilation intensity according to ③ in C3; When the worst physical condition among all users in the room is at the bad level, adjust the humidity and ventilation intensity according to ④ in C3; The circle with a diameter of 3m on the ground directly below each air outlet is used as the statistical area, and the physical condition of the users directly below each air outlet is counted: When the worst physical condition of the user under a certain air outlet is in good condition, the air outlet temperature of this air outlet increases by 1°C; When the worst physical condition of the users under a certain air outlet is at the intermediate level, the air outlet temperature of this air outlet increases by 2°C; When the worst physical condition among the users under a certain air outlet is at the bad level, the air outlet temperature of this air outlet increases by 3°C.
Citation Information
Cited By
Cooperative control system for indoor environment adjusting system
CN120140879A
