An intelligent wearable device and method for rapid identification of personal thermal comfort

Abstract

The application discloses a kind of intelligent wearable device and method for personal thermal comfort quick identification, comprising: (1) intelligent wearable device is by environmental temperature and humidity detection module, human skin temperature detection module, human heart rate detection module, thermal comfort voting module, wireless communication module, power supply module, display module etc. (2) small batch of environment / human physiological parameters are collected by intelligent wearable device, including environmental temperature and humidity, human skin temperature, human heart rate and thermal comfort voting value to construct target domain data set. (3) control computer selects similar data to construct source domain data set in the public pseudo-source domain data pool according to the target domain data uploaded by intelligent wearable device. (4) control computer constructs thermal comfort classifier. The application has the advantages of high sensor integration, small amount of data acquisition, high prediction accuracy, etc., which helps to improve indoor thermal comfort level while reducing indoor building energy consumption.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of indoor thermal environment regulation, and specifically belongs to an intelligent wearable device and method for rapid identification of thermal comfort of a crowd. BACKGROUND

[0002] Creating a good indoor space thermal comfort can not only save building energy consumption, but also improve people's happiness, productivity and work efficiency. Traditional indoor environment prediction methods include but are not limited to PMV (Predicted Mean Vote) method, adaptive comfort model prediction method, etc. The PMV model uses mathematical expressions of environmental factors (air temperature, radiant temperature, wind speed and humidity) and personal factors (metabolic rate and clothing thermal resistance) to express the index of the thermal feeling of the occupants. The personal factors are mainly estimated roughly, and cannot reflect the thermal comfort differences of the crowd, region, etc.; the adaptive model uses linear regression between thermal feeling and working temperature to define the acceptable thermal environment, and also has the problem of poor generalization performance. Patent CN114322230A discloses a thermal environment regulation system and method based on a smart bracelet. The smart bracelet in the patent collects two parameters of human body temperature and environmental temperature, and does not involve the collection of human heart rate and environmental humidity as described in the patent. Studies have shown that human temperature, heart rate and environmental temperature and humidity are closely related to personal thermal comfort. Secondly, the patent mentions using traditional data-driven models such as neural networks or support vector machines for thermal comfort prediction. Unlike this, the prediction model described in the patent only needs to collect a small amount of field data, and identifies the thermal comfort level through a transfer learning strategy, with small data collection amount and good rapidity. SUMMARY

[0003] The purpose of the application is to provide an intelligent wearable device and method for rapid identification of thermal comfort of a crowd, which can predict the overall thermal comfort level of the indoor crowd, improve the automation level of indoor thermal environment regulation, and thus reduce building energy consumption.

[0004] To achieve the above object, the present application provides the following technical solutions: a kind of intelligent wearable device for personal thermal comfort degree fast identification, the main components of the intelligent wearable device are arranged on the two surfaces of the double-layer PCB board of wrist width, and the main components include: environmental temperature and humidity detection module, for detecting environmental temperature and humidity;Skin temperature detection module, for obtaining skin temperature;Human heart rate detection module, for detecting human heart rate signal;Thermal comfort voting module, for user subjective thermal comfort degree voting, wireless communication module, for reading the data of each module and wireless transmission to control computer;Power supply module, for the direct current power supply of device;OLED display module, for displaying the collected data, the environmental temperature and humidity detection module, skin temperature detection module, human heart rate detection module, thermal comfort voting module, display module, wireless communication module and power supply module are connected with the intelligent wearable device.

[0005] The present application provides a kind of intelligent wearable method for personal thermal comfort degree fast identification, comprising the following steps: (1) intelligent wearable device is composed of environmental temperature and humidity detection module, human skin temperature detection module, human heart rate detection module, thermal comfort voting module, wireless communication module, power supply module, OLED display module and so on.(2) small batch of environmental / human physiological parameters are collected by intelligent wearable device, including environmental temperature and humidity, human skin temperature, human heart rate and thermal comfort voting value to construct target domain data set.(3) control computer constructs source domain data set according to the target domain data uploaded by intelligent wearable device in the similar data selected in the public pseudo-source domain data pool.(4) control computer constructs thermal comfort classifier.First, the parameters of the classifier are pre-trained using source domain data, and then the parameters of the classifier are fine-tuned using target domain data, and the obtained classifier can be used for the rapid identification of personal thermal comfort degree.

[0006] Further, the method used by the sensor in step (1) to collect heart rate is the optical volume scanning method (PPG) to measure heart rate. The basic principle of PPG is to use the absorption and reflection characteristics of light, use photoelectric sensor to detect the light intensity passing through skin, tissue and blood vessels, and calculate the heart rate. These changes are reflected in the PPG signal, and the fluctuation of the signal is synchronized with the heart cycle. Let I(t) be the light intensity measured at time t, then the PPG signal can be expressed as:

[0007] I(t)=I0+ΔI(t)

[0008] Where I0 is the basic light intensity (the constant part reflected by skin and tissue), and ΔI(t) is the light intensity change part caused by blood volume change. In the process of sensor detection, the peak value can be obviously detected, and the heart rate can be calculated by detecting these peak values. Heart rate (HR) can be expressed as:

[0009]

[0010] where N is the number of peaks, T is the time interval (in seconds), and 60 is used to convert seconds to minutes. The final HR is the heart rate per minute

[0011] Further, the smart wearable device is worn on the wrist of the user, the heart rate sensor and the skin temperature sensor are close to the skin of the user, and the control computer and the wireless communication module of the smart wearable device are in a communication state. In the data acquisition process, the smart wearable device display module is used to display the information collected by the sensor in real time, including: the skin temperature of the human body, the heart rate of the human body, the environmental temperature and humidity, and the thermal sensation voting value; the thermal sensation comfort voting module is used to vote for the thermal sensation of the user under the current environmental state.

[0012] Further, in step (2), the subject is required to vote for the thermal sensation value in a stable indoor thermal environment, and the smart wearable device is used to vote for the current thermal sensation value. After the voting is completed, the button is pressed to indicate confirmation. At this time, the OLED display screen on the smart wearable device displays a message sending state, and after the sending is completed, the display screen displays "OK", indicating that the information collection is completed. The smart wearable device is used to collect the physiological state information of the subject, the surrounding environment information, and the seven-level thermal sensation voting value. A small batch of 50-60 groups of data are collected to construct a target domain data set, and preparation is made for the migration learning of the model.

[0013] Further, in step (3), due to the influence of related factors such as region, climate, and subjects, it is found that the data distribution difference between the target data set and the public data set is too large, so a nearest neighbor search algorithm is used to find similar data in the public data set for migration learning. The main process of this method is as follows: assuming that all data are unknown categories, based on the Euclidean distance, the data with the smallest distance from the given query item in a search space is found.

[0014] The definition of Euclidean distance is:

[0015]

[0016] where x i =(x1,x2,...x n ) and y i =(y1,y2,...y n ) are two given n-dimensional data points, which respectively represent the data points in the target domain and the source domain data set in the present application. The indoor temperature, humidity, skin temperature, and human heart rate are selected as features, and the source domain (data in the public data set) is selected as the search space for nearest neighbor search. The calculation formula is as follows:

[0017]

[0018] Y = unique (y1, y2, …, y m )

[0019] where dist (q t , x) is the Euclidean distance between q t and x, q t is the data query point in the target domain dataset, t represents the number of features, x is the data point in the source domain dataset, and the argmin operation is used to find the point that minimizes the Euclidean distance as y t . Then, the unique function is used to exclude duplicate objects, and the searched data is used as the source domain dataset.

[0020] Further, the classifier model in step (4) is constructed based on a one-dimensional convolutional neural network (1D-CNN) model; wherein the 1D-CNN model includes two convolutional layers and one fully connected layer, and a ReLU activation function is applied after each convolutional layer to enhance the non-linear expression ability of the model; after the convolutional layer, a fully connected layer containing 50 neurons is used, which functions to combine and transform the features extracted by the convolutional layer, thereby further extracting high-level features; during the process of training the personal thermal comfort classifier using the target domain, the parameters of the fully connected layer 2 of the neural network model trained by the source domain are migrated to the fully connected layer 2 of the target domain model; finally, the output from the fully connected layer is transmitted to the output layer, and a softmax activation function is used to improve the classification efficiency and accuracy of the 1D-CNN model, thereby improving the precision of the model.

[0021] Further, the model is deployed on a control computer, at this time the subject wears the smart wearable device, which continuously collects physiological parameters, environmental information parameters and other data without affecting normal office work, and sends the data to the control computer through the wireless communication module; the control computer runs the personal thermal comfort classifier for prediction, and then obtains the thermal comfort condition of the subject under the current state.

[0022] The beneficial effects of the present application are:

[0023] The present application is aimed at the characteristics of being able to collect only small batches of data or insufficient data in an online environment, and through the introduction of model transfer learning method, the smart wearable device can quickly predict the thermal comfort of the crowd under the condition of small sample in the online situation, and the finally measured thermal comfort can be applied to the heating ventilation and air conditioning system to improve the automatic adjustment level of the temperature in the system and reduce the building energy consumption. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 is a method flowchart

[0025] Figure 2 is a method smart wearable device composition diagram

[0026] Figure 3 is the circuit PCB structure diagram of the intelligent wearable device in the application

[0027] Figure 4 is the thermal comfort voting value range;

[0028] Fig. 1, intelligent wearable device; 2, thermal comfort voting module; 3, human heart rate detection module; 4, environment temperature and humidity detection module; 5, human skin temperature detection module; 6, display module; 7, wireless communication module; 8, power supply module. DETAILED DESCRIPTION

[0029] The application discloses an intelligent wearable device and method for quickly identifying thermal comfort of a crowd, and aims to quickly predict the thermal comfort of the current state of the crowd by using the intelligent wearable device, so as to help improve the automatic adjustment performance of a heating ventilation air conditioning system. Figure 1 As shown in the overall flow chart, the method comprises the following steps:

[0030] Step 1: encapsulate a thermal comfort voting module 2, a human heart rate detection module 3, an environment temperature and humidity detection module 4, a skin temperature detection module 5, a display module 6, a wireless communication module 7 and a power supply module 8 into an intelligent wearable device as shown in Figure 2 .

[0031] The application adopts a double-layer PCB structure, including an upper layer and a lower layer, and the outer dimensions are 65*33*1.6MM as shown in Figure 3 , and various sensors and communication components are arranged respectively.

[0032] The thermal comfort voting module device is connected with the communication device through a GPIO port, and the device is a 4.5*4.5*5 vertical patch 4-pin micro-motion button switch, two of which are used to confirm that data is sent to a control computer, and when the button is pressed, the data is packaged and sent to the control through the communication module.

[0033] The human heart rate detection module adopts a MAX30102 human heart rate acquisition module, and the communication module adopts IIC to exchange data, wherein the heart rate detection module detects the human heart rate by using a light volume scanning method.

[0034] The environment temperature and humidity detection module adopts a DHT11 module to detect environment parameters in real time by using temperature-sensitive and humidity-sensitive resistors, and corresponding temperature and humidity data are obtained through an amplification circuit and a filter circuit in the circuit, the module can detect the temperature and humidity of the environment in real time, and transmit the data to a main controller, and exchange data with the communication module through a GPIO port.

[0035] The human skin temperature and humidity detection module adopts AHT10, uses SMD packaging, and is used for measuring the skin surface temperature through an improved MEMS semiconductor capacitive humidity sensor and a standard on-chip temperature sensor element. The module exchanges data with the communication module through a GPIO port.

[0036] The OLED display module adopts a 0.96-inch OLED display liquid crystal serial screen module, which communicates through IIC.

[0037] The wireless communication module adopts ESP8266 NODEMCU, which is composed of an ESP8266 WiFi module, a USB-to-TTL chip, a 5V-to-3.3V LDO, and an automatic download circuit.

[0038] The power supply module adopts a 3.7V-300mAh polymer lithium battery, and the lithium battery charging and discharging circuit adopts a chip model tp4056 lithium battery management chip, which adopts a SOP-8 packaging technology, with a size of 5.2*4.9*1.2mm, a working temperature of -40℃-130℃, and a working voltage of 1-8V.

[0039] Step two: The intelligent wearable device collects a small batch of environmental / human physiological parameters, including environmental temperature and humidity, human skin temperature, human heart rate, and thermal comfort voting value to construct a target domain data set.

[0040] The basic process of collecting small batch of thermal comfort data on site is as follows: first, keep the target indoor thermal environment stable, the subject wears the intelligent wearable device into the room, and the control computer keeps stable communication with the intelligent wearable device. After the subject's thermal sensation state is stable, input his / her 7-level thermal comfort state in the intelligent wearable device, and the thermal comfort value range is as shown in Figure 4 When the user presses the confirmation button, the display screen displays the word "OK", indicating that the intelligent wearable device has packaged and sent the data to the control computer, and a set of data collection is completed. Then, the subject adjusts the air conditioning temperature according to his / her thermal preference, and when the current thermal environment is stable, the subject performs thermal sensation voting again on the intelligent wearable device. The number of collections and the collection time are not limited, and 10-50 sets of experimental samples can be collected on site. The control computer pre-processes these data sets, including abnormal value detection and normalization processing, and the processed data sets are used as the target domain data set.

[0041] Step three: The control computer selects similar data from the public pseudo-source domain data pool to construct the source domain data set according to the target domain data uploaded by the intelligent wearable device.

[0042] The ASHRAE RP-884 database is one of the public databases widely used in human thermal comfort research. The RP-884 dataset involves more than 25000 observations collected from 52 studies in different climate zones and 26 cities around the world. The target domain dataset generated by step two, using the nearest neighbor search algorithm (NNS) to filter similar data in the RP-884 dataset, the data features include: environmental temperature, environmental humidity, human heart rate, skin temperature, thermal sensation vote, and the filtered data constitute the source domain dataset in this method.

[0043] Step four: control computer builds thermal comfort classifier.

[0044] First, the source domain data is used to pre-train the classifier parameters, and then the target domain data is used to fine-tune the classifier parameters, and the obtained classifier can be used for rapid identification of human thermal comfort. The parameters for building the classifier are shown in Table 1, where the Dropout parameter can handle the overfitting problem by randomly reducing the number of neurons in the hidden layer during training. Batch size is related to the generalization of model performance, by training the entire dataset, dividing them into small parts can reduce memory costs, speed up the classifier to make predictions, and improve real-time performance.

[0045] Table 1

[0046]

[0047] Step five: online collection of physiological and environmental information using intelligent wearable devices, based on the built thermal comfort model for rapid identification and prediction of individual thermal comfort.

[0048] Through the sensors on the wearable device, the human physiological signals and surrounding environmental parameters are collected online and in real time, and the data is uploaded to the control computer through the communication component. The control computer inputs the collected data into the built thermal comfort classifier, and the obtained prediction result is the individual thermal comfort level under the current thermal environment.

Claims

1. A smart wearable method for rapid identification of personal thermal comfort, characterized in that, The steps include: (1) The smart wearable device includes an environmental temperature and humidity detection module, a human skin temperature detection module, a human heart rate detection module, a thermal comfort voting module, a wireless communication module, a power supply module, and a display module; (2) The smart wearable device collects a small batch of environmental / human physiological parameters, including environmental temperature and humidity, human skin temperature, human heart rate, and thermal comfort voting values ​​to construct a target domain dataset; (3) The control computer selects similar data from a publicly available pseudo-source domain data pool based on the target domain data uploaded by the smart wearable device to construct a source domain dataset; (4) The control computer constructs a thermal comfort classifier, first pre-trains the classifier parameters using source domain data, and then fine-tunes the classifier parameters using target domain data. The resulting classifier can be used for rapid identification of personal thermal comfort. In step (3), due to the influence of factors such as region, climate, and subject-related factors, it was found that the data distribution difference between the target dataset and the public dataset was too large. It is necessary to use the nearest neighbor search algorithm to find data in the public dataset that is similar to the target domain for transfer learning. The main process of this method is as follows: Assuming that all data are of unknown category, based on Euclidean distance, find the data with the smallest distance to the given query item in a search space. The Euclidean distance is defined as: ; in and Given two n-dimensional data points, which in this invention represent data points in the target domain and source domain datasets respectively, indoor temperature, humidity, skin temperature, and human heart rate are selected as features. Nearest neighbor search is performed using the source domain as the search space, and the calculation formula is as follows: ， ； in It is a calculation and The Euclidean distance between them For each data query point in the target domain dataset, t represents the number of features. For data points in the source domain dataset, The operation is used to find the point that minimizes the Euclidean distance. Then use The function excludes duplicate objects and uses the searched data as the source dataset. In step (4), the classifier model is constructed based on a one-dimensional convolutional neural network (1D-CNN) model. The 1D-CNN model contains two convolutional layers and one fully connected layer. The ReLU activation function is applied after each convolutional layer to enhance the non-linear expressive ability of the model. After the convolutional layers, the model uses a fully connected layer with 50 neurons, which combines and transforms the features extracted by the convolutional layers to further extract high-level features. In the process of training the personal thermal comfort classifier using the target domain, the parameters in the fully connected layer 2 of the source domain training neural network model are transferred to the fully connected layer 2 of the target domain model. Finally, the output from the fully connected layer is passed to the output layer, and the softmax activation function is used to improve the classification efficiency and accuracy of the 1D-CNN model, thereby improving the accuracy of the model.

2. The smart wearable method for rapid identification of personal thermal comfort according to claim 1, characterized in that, In step (1), the human heart rate detection module uses photoplethysmography (PPG) to measure heart rate. The basic principle of PPG is to utilize the absorption and reflection properties of light, using a photoelectric sensor to detect the light intensity passing through the skin, tissues, and blood vessels, thereby calculating the heart rate. These changes are reflected in the PPG signal, and the signal fluctuations are synchronized with the heart cycle. If the light intensity is measured at time t, then the PPG signal can be expressed as: ; in, It is the basic light intensity, a constant component reflected by the skin and tissues. The change in light intensity is due to changes in blood volume. During sensor detection, peak values ​​can be clearly detected. By detecting these peak values, heart rate can be calculated. Heart rate (HR) can be expressed as: ; In the formula, N is the number of peak values, T is the time interval, 60 is used to convert seconds to minutes, and the final HR is the heart rate per minute.

3. The smart wearable method for rapid identification of personal thermal comfort according to claim 1, characterized in that, The smart wearable device is worn on the user's wrist, with the human heart rate detection module and skin temperature detection module in close contact with the user's skin. The control computer maintains communication with the wireless communication module of the smart wearable device. During data acquisition, the display module is used to display the information collected by the sensors in real time, including: human skin temperature, human heart rate, ambient temperature and humidity, and thermal comfort voting value. The thermal comfort voting module is used to allow the user to vote on the thermal comfort under the current environmental conditions.

4. The smart wearable method for rapid identification of personal thermal comfort according to claim 1, characterized in that, In step (2), the subject is required to vote on the thermal sensation value in a stable indoor thermal environment. The subject uses a smart wearable device to vote on the current thermal sensation value. After voting, the subject presses a button to confirm. At this time, the display module on the smart wearable device will show the message sending status. After the message is sent, the display module will show "OK", which means that a set of information collection is completed. The subject's physiological state information, surrounding environment information, and seven-level thermal sensation voting values ​​are collected using the smart wearable device. A small batch of 50 to 60 sets of data are collected to construct the target domain dataset to prepare for the transfer learning of the model.

5. A smart wearable method for rapid identification of personal thermal comfort as described in claim 1, characterized in that, To address the instability in accuracy of the human heart rate detection module due to external interference and noise, a mean filtering algorithm is used to calculate the arithmetic mean of the sampled heart rates within a fixed time window as the heart rate value. The filtered heart rate value at time point t... The calculation formula is: ; in, It is the window width. These are the original sampled values ​​from the heart rate sensor.

Images