High-speed train passenger thermal comfort identification system and method based on peripheral physiological features

Abstract

The invention relates to the field of high-speed train health and safety guarantee, in particular to a high-speed train passenger thermal comfort recognition system and method based on peripheral physiological characteristics, and the system comprises a physiological data and group information collection module, a data preprocessing and characteristic fusion module, a passenger thermal comfort state recognition module and a group decision and internet-of-things transmission module. The high-speed train passenger thermal comfort identification system and method based on the peripheral physiological features have the beneficial effects that the system effectively extracts rich feature information from a thermal imaging camera and a smart watch by combining intelligent sensing acquisition, intelligent preprocessing and feature fusion technologies; and the recognition capability of the model on targets with different scales is enhanced, and efficient feature extraction is realized.

Description

Technical Field

[0001] The present invention relates to the field of high-speed rail health and safety protection, and in particular to a high-speed train occupant thermal comfort identification system and method based on peripheral physiological characteristics. Background Art

[0002] The development of rail transport has led to the widespread use of high-speed rail, with hundreds of millions of passengers choosing this mode of travel. During the hours-long journey, prolonged exposure to uncomfortable heat can have adverse effects on the physical and mental health of high-speed rail passengers. Uncomfortable indoor temperatures can have an impact on all aspects of human health. For example, a colder environment increases the risk of cardiovascular and respiratory diseases, while also providing a more favorable environment for the survival of viruses; higher temperatures increase the risk of acute nonspecific symptoms, such as dry eyes and respiratory problems; insufficient indoor ventilation reduces air quality and increases the chance of infection with infectious diseases, further increasing the risk of acute and chronic diseases.

[0003] At present, the human physiological regulation system has the ability to actively adjust these parameters within a certain range of environmental conditions, so the physiological level and heat dissipation of the human body are not only determined by environmental factors. In other words, high-speed train passengers can adapt and maintain thermal balance according to environmental conditions through various physiological mechanisms, such as sweating or vasoconstriction. Through the peripheral physiological signals of passengers, we can gain a deep understanding of the physiological responses and adaptation mechanisms of passengers in different thermal environments. These physiological indicators provide valuable information for understanding and evaluating thermal comfort, and help to formulate effective strategies to improve the comfort and well-being of high-speed train passengers. Comfort, intelligence and efficiency are the directions for high-quality development of high-speed railway transportation. At present, there is a problem that the thermal comfort of high-speed train passengers cannot be accurately evaluated, which also affects the flexibility and accuracy of the air supply control of the air conditioning system of the carriage thermal environment. The inability to provide a comfortable in-car thermal environment also affects energy efficiency.

[0004] In the prior art, the invention patent "A method for controlling a heating and ventilation system based on multi-user thermal comfort data" (patent number CN 107120782 B) does not describe the data acquisition equipment, and this patent uses a unified formula to calculate thermal comfort, which is not applicable to a variety of scenarios. The human thermal comfort recognition method of the invention patent "Thermal comfort evaluation system based on environmental parameters and human physiological parameters" (patent number CN 104633866 B) has poor applicability and does not take into account the particularity of the railway system.

[0005] Therefore, it is necessary to propose a high-speed train occupant thermal comfort identification system and method based on thermal imaging cameras to overcome the problems that the existing technology is not suitable for various scenarios and does not take into account the particularity of the railway system. Summary of the invention

[0006] The purpose of the present invention is to provide a high-speed train occupant thermal comfort identification system and method based on peripheral physiological characteristics to solve the problems raised in the above background technology.

[0007] To achieve the above-mentioned object, the present invention provides the following technical solutions: a high-speed train passenger thermal comfort identification system based on peripheral physiological characteristics, the system comprising a physiological data and group information acquisition module, a data preprocessing and feature fusion module, a passenger thermal comfort state identification module and a group decision-making and IoT transmission module;

[0008] Physiological data and group information collection module, responsible for collecting occupant peripheral physiological signals from thermal imaging cameras and smart watch devices;

[0009] The data preprocessing and feature fusion module is used to receive the collected signals and perform feature preprocessing and fusion on the collected signals;

[0010] The occupant thermal comfort state recognition module receives real-time changes in the occupant's thermal comfort state;

[0011] The group decision-making and IoT transmission module stores the collected thermal physiological signals and thermal comfort status categories, aggregates them into group thermal comfort levels, and transmits them to the air-conditioning control system in the train.

[0012] Preferably, the physiological data and group information acquisition module collects group category information corresponding to the passengers from the ticket purchasing system. The physiological data and group information acquisition module includes a digital-to-analog converter, a data acquisition card, and a data transmission interface component, which are used to convert real-time scenes into digital signals for processing.

[0013] Preferably, the data preprocessing and feature fusion module is connected to the physiological data and group information collection module and the occupant thermal comfort status identification module, receives the collected signals, performs feature preprocessing and fusion on the collected signals, and facilitates classification by the thermal comfort level identification module. The data preprocessing and feature fusion module consists of a data processing unit, a cache unit, and a data wireless transmission unit.

[0014] Preferably, the occupant thermal comfort state recognition module is connected to the data preprocessing and feature fusion module and the group decision-making and Internet of Things transmission module to receive real-time changes in the occupant thermal comfort state, and accelerates the response speed and accuracy of the air conditioning air supply strategy by transmitting the information to the air conditioning control system in real time; the occupant thermal comfort state recognition module is composed of a central data processing unit, a cache unit, a data wireless transmission unit and a data transmission interface.

[0015] Preferably, the group decision-making and Internet of Things transmission module is connected to the passenger thermal comfort status identification module, and the collected thermal physiological signals and thermal comfort status categories are stored, aggregated into a group thermal comfort level, and transmitted to the air-conditioning control system in the train. The group decision-making and Internet of Things transmission module consists of a control chip, a storage controller, a data wireless transmission unit and an interface circuit.

[0016] A method for identifying thermal comfort of high-speed train passengers based on peripheral physiological characteristics, the method comprising the following steps:

[0017] Step S1: constructing a train occupant thermal comfort dataset based on peripheral physiological characteristics;

[0018] Step S2: data preprocessing and feature fusion;

[0019] Step S3: Identification of thermal comfort state of occupants, using the extracted features to accurately identify the thermal comfort level through a machine learning algorithm;

[0020] Step S4: construct a real-time monitoring and transmission device for the thermal comfort status of passengers in the vehicle.

[0021] Preferably, the specific steps of step S1 are as follows:

[0022] Step S1.1: Under different seasons and different air-conditioning system working conditions, the peripheral physiological signals of the subjects are fully collected, including skin temperature of various parts of the face, galvanic skin response GSR and blood volume pulse BVP, and the subjective thermal comfort questionnaire of the subjects is collected. The heart rate variability HRV can be extracted from the blood volume pulse BVP. Heart rate variability HRV refers to the change in the number of heart beats per minute, which reflects the fluctuation of the heartbeat interval. The time domain index SDNN and nonlinear index can be calculated by the following formula. SDNN is the calculation result of the standard deviation of all normal heartbeat intervals NN intervals, which is used to reflect the overall level of overall heart rate variability. N represents the total number of NN intervals, and NNi represents the i-th N interval. is the average value of all NN intervals; SD1 and SD2 are geometric analysis indicators based on the Poincaré scatter plot, which are used to quantify instantaneous heart rate variability and long-term heart rate variability. The Poincaré scatter plot is a point plotting each heart beat interval with its previous interval on a two-dimensional plane, i.e., point (NNi, NNi+1); SD1 represents the standard deviation on the short axis of the scatter plot, reflecting the instantaneous component of heart rate variability, and is used to evaluate parasympathetic nerve activity; SD2 represents the standard deviation on the long axis of the scatter plot, reflecting the long-term component of heart rate variability, and is used to evaluate the joint action of sympathetic and parasympathetic nerves; define the standard deviation on the diagonal and vertical lines, and calculate SD1 and SD2 respectively; where Var(NNi-NNi+1) represents the variance of the difference between all consecutive NN intervals;

[0023]

[0024] For GSR data, it is uniformly converted into a change value from the baseline, i.e., ΔGSR, to reduce the measurement error under different air-conditioning conditions in different seasons. The baseline GSR value is defined as the lowest GSR value within the first 2-5 minutes of the measurement. The average facial skin temperature MSTF is obtained according to the area-weighted value of each part, where Tsk is the local skin temperature of the face, dAsk is the corresponding local surface area of ​​each part, and A is the total facial area;

[0025] MSTF=∫T sk dA sk / A = forehead local skin temperature × 0.2 + eye local skin temperature × 0.1 + cheek local skin temperature × 0.4 + nose local skin temperature × 0.1 + ear local skin temperature × 0.2;

[0026] S1.2: Group information collection. Adding differentiated categories can improve the performance and accuracy of the model. Therefore, the group category labels of the riding season, seat row and seat column, gender, BMI and age corresponding to each passenger are collected to enhance the model's ability to recognize the target.

[0027] Preferably, the specific steps of step S2 are as follows:

[0028] S2.1: Use LabelEncoder and StandardScaler tools to standardize each identification factor feature. The LabelEncoder tool encodes the group category feature label using a value between 0 and n_classes-1, and the StandardScaler tool standardizes the numerical peripheral physiological features.

[0029] S2.2: The StandardScaler tool is based on data standardization. It performs linear transformation on numerical features so that each feature, that is, each attribute or column of data, follows a standard normal distribution, with a mean of 0 and a standard deviation of 1. For a given eigenvalue X, StandardScaler standardizes each feature. The transformed eigenvalue X′ is calculated according to the following formula, where μ is the mean of feature X and σ is the standard deviation of feature X:

[0030]

[0031] Preferably, the specific steps of step S3 are as follows:

[0032] S3.1: Thermal comfort identification of high-speed train passengers includes two parts: overall thermal sensation and local thermal sensation. The established data set includes passengers' HRV characteristics, GSR characteristics, wrist skin temperature, skin temperature of various parts of the face and average skin temperature of the face, as well as population division, seat identification and season;

[0033] S3.2: For overall thermal sensation recognition, HRV features, GSR features, facial mean skin temperature MSTF and wrist skin temperature are integrated as peripheral physiological recognition features; for local thermal sensation recognition of the face, the skin temperature of the corresponding part is selected, and group category labels are used to enhance the recognition accuracy of overall and local thermal sensations;

[0034] S3.3: Input the pre-processed feature data set into the machine learning model for training. During the training process, the performance of the model is optimized and overfitting is prevented through data set partitioning, cross-validation and hyperparameter optimization strategies.

[0035] S3.4: K-nearest neighbor algorithm is an instance-based non-parametric classification algorithm. It selects the nearest K samples by calculating the distance between the sample to be classified and each sample in the training set, and votes and classifies them according to the labels of these samples.

[0036] Preferably, the real-time monitoring and transmission device for the thermal comfort status of the occupants in step S4 includes a physiological data and group information acquisition module, a data preprocessing and feature fusion module, an occupant thermal comfort status recognition module and a group decision-making and IoT transmission module.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The high-speed train occupant thermal comfort identification system and method proposed in the present invention, by combining intelligent sensing acquisition, intelligent preprocessing and feature fusion technology, the system effectively extracts rich feature information from thermal imaging cameras and smart watches, enhances the model's recognition ability for targets of different scales, and realizes efficient feature extraction;

[0039] By integrating peripheral physiological and group category features with machine learning algorithms, the system can identify thermal comfort status in real time and quickly adjust to different air conditioning conditions in different seasons, thereby improving the accuracy and flexibility of identification and achieving real-time and accurate identification of occupant thermal comfort status;

[0040] The system is designed with the complexity of group physiological data acquisition in mind. The number and placement of thermal imaging cameras can be flexibly deployed to achieve easy acquisition and deployment.

[0041] With high-performance hardware configuration and optimized algorithm design, this system can realize real-time data processing and thermal comfort state recognition. This real-time performance ensures that the air conditioning control system can respond and adjust the air supply parameters in time according to the thermal comfort state of the occupants, providing immediate support for the improvement of thermal comfort and the reasonable allocation of energy consumption, with strong real-time performance;

[0042] The system supports personalized flexible management. In the group decision module, it allows the output of group thermal comfort decisions with individual emphasis based on the age distribution, number of passengers and density distribution of passengers. This personalized management and flexible decision-making function enables the system to respond to various situations more accurately, improving the pertinence and effectiveness of thermal comfort control. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a flow chart of the method of the present invention;

[0044] Figure 2 This is a system architecture diagram of the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose and technical solution of the present invention clearly and completely described, and the advantages more clearly understood, the embodiments of the present invention are further described in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are part of the embodiments of the present invention, rather than all of the embodiments, and are only used to explain the embodiments of the present invention, and are not used to limit the embodiments of the present invention. All other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] For example, see Figure 1-Figure 2 The present invention provides a technical solution: a high-speed train occupant thermal comfort identification system based on peripheral physiological characteristics, the system comprising a physiological data and group information acquisition module, a data preprocessing and feature fusion module, an occupant thermal comfort state identification module and a group decision-making and IoT transmission module.

[0047] The physiological data and group information acquisition module is responsible for collecting the peripheral physiological signals of the passengers from the thermal imaging camera and the smart watch device; the physiological data and group information acquisition module collects the group category information corresponding to the passengers from the ticket purchasing system. The physiological data and group information acquisition module includes a digital-to-analog converter, a data acquisition card, and a data transmission interface component, which is used to convert the real-time scene into a digital signal for processing; among them, the PPG sensor and EDA sensor of the smart watch accurately collect the blood volume pulse (BVP) and skin electrical signal changes (GSR) of the passengers. In addition, it can also read the skin temperature around the watch. At the same time, the thermal imaging camera arranged in the car monitors the skin temperature of various parts of the passengers' faces (forehead, eyes, nose, cheeks, ears) under different environmental parameters in real time; the physiological data and group information acquisition module is responsible for collecting the season, personal characteristics of the passengers (such as gender, age and body mass index BMI) and seats of the passengers (rows and columns).

[0048] The data preprocessing and feature fusion module is used to receive the collected signals, perform feature preprocessing and fusion on the collected signals, convert the group information features into the group category labels required by the recognition model, and extract the heart rate variability characteristics and the ΔGSR of the relative baseline change from the peripheral physiological signals, and convert the local facial skin temperature into the average facial skin temperature according to the surface area weights of each part of the face; the data preprocessing and feature fusion module is connected to the physiological data and group information collection module and the occupant thermal comfort state recognition module, receives the collected signals, performs feature preprocessing and fusion on the collected signals, and facilitates the thermal comfort level recognition module to perform classification. The data preprocessing and feature fusion module consists of a data processing unit, a cache unit, and a data wireless transmission unit.

[0049] The occupant thermal comfort state recognition module receives real-time changes in the occupant thermal comfort state; the occupant thermal comfort state recognition module is connected to the data preprocessing and feature fusion module and the group decision-making and Internet of Things transmission module, receives real-time changes in the occupant thermal comfort state, and accelerates the response speed and accuracy of the air conditioning air supply strategy by transmitting the information to the air conditioning control system in real time; the occupant thermal comfort state recognition module consists of a central data processing unit, a cache unit, a data wireless transmission unit and a data transmission interface.

[0050] The group decision-making and Internet of Things transmission module stores the collected thermal physiological signals and thermal comfort status categories, aggregates them into group thermal comfort levels, and transmits them to the air-conditioning control system in the train; the group decision-making and Internet of Things transmission module is connected to the passenger thermal comfort status identification module, stores the collected thermal physiological signals and thermal comfort status categories, aggregates them into group thermal comfort levels, and transmits them to the air-conditioning control system in the train. The group decision-making and Internet of Things transmission module consists of a control chip, a storage controller, a data wireless transmission unit and an interface circuit.

[0051] The high-speed train passenger thermal comfort identification system based on peripheral physiological characteristics effectively improves the recognition accuracy and efficiency of the thermal comfort levels of high-speed train passengers under different thermal environment conditions through the fusion of machine learning technology and peripheral physiological characteristics, and realizes accurate recognition of the overall and local thermal comfort of passengers in different seasons, with different personal characteristics and in different seats, so as to achieve more personalized and efficient in-vehicle thermal environment control.

[0052] Embodiment 2, based on embodiment 1, provides a method for identifying thermal comfort of high-speed train passengers based on peripheral physiological characteristics, the method comprising the following steps:

[0053] Step S1: constructing a train occupant thermal comfort dataset based on peripheral physiological characteristics;

[0054] Step S1.1: Under different seasons and different air-conditioning system working conditions, the peripheral physiological signals of the subjects are fully collected, including skin temperature of various parts of the face, galvanic skin response GSR and blood volume pulse BVP, and the subjective thermal comfort questionnaire of the subjects is collected. The heart rate variability HRV can be extracted from the blood volume pulse BVP. Heart rate variability HRV refers to the change in the number of heart beats per minute, which reflects the fluctuation of the heartbeat interval. The time domain index SDNN and nonlinear index can be calculated by the following formula. SDNN is the calculation result of the standard deviation of all normal heartbeat intervals NN intervals, which is used to reflect the overall level of overall heart rate variability. N represents the total number of NN intervals, and NNi represents the i-th N interval. is the average value of all NN intervals; SD1 and SD2 are geometric analysis indicators based on the Poincaré scatter plot, which are used to quantify instantaneous heart rate variability and long-term heart rate variability. The Poincaré scatter plot is a point plotting each heart beat interval with its previous interval on a two-dimensional plane, i.e., point (NNi, NNi+1); SD1 represents the standard deviation on the short axis of the scatter plot, reflecting the instantaneous component of heart rate variability, and is used to evaluate parasympathetic nerve activity; SD2 represents the standard deviation on the long axis of the scatter plot, reflecting the long-term component of heart rate variability, and is used to evaluate the joint action of sympathetic and parasympathetic nerves; define the standard deviation on the diagonal and vertical lines, and calculate SD1 and SD2 respectively; where Var(NNi-NNi+1) represents the variance of the difference between all consecutive NN intervals;

[0055]

[0056] For GSR data, it is uniformly converted into a change value from the baseline, i.e., ΔGSR, to reduce the measurement error under different air-conditioning conditions in different seasons. The baseline GSR value is defined as the lowest GSR value within the first 2-5 minutes of the measurement. The average facial skin temperature MSTF is obtained according to the area-weighted value of each part, where Tsk is the local skin temperature of the face, dAsk is the corresponding local surface area of ​​each part, and A is the total facial area;

[0057] MSTF=∫T sk dA sk / A = forehead local skin temperature × 0.2 + eye local skin temperature × 0.1 + cheek local skin temperature × 0.4 + nose local skin temperature × 0.1 + ear local skin temperature × 0.2;

[0058] S1.2: Group information collection. Adding differentiated categories can improve the performance and accuracy of the model. Therefore, the group category labels of the riding season, seat row and seat column, gender, BMI and age corresponding to each passenger are collected to enhance the model's ability to recognize the target.

[0059] Step S2: data preprocessing and feature fusion;

[0060] S2.1: Use LabelEncoder and StandardScaler tools to standardize each identification factor feature. The LabelEncoder tool encodes the group category feature label using a value between 0 and n_classes-1, and the StandardScaler tool standardizes the numerical peripheral physiological features.

[0061] S2.2: The StandardScaler tool is based on data standardization. It performs linear transformation on numerical features so that each feature, that is, each attribute or column of data, follows a standard normal distribution, with a mean of 0 and a standard deviation of 1. For a given eigenvalue X, StandardScaler standardizes each feature. The transformed eigenvalue X′ is calculated according to the following formula, where μ is the mean of feature X and σ is the standard deviation of feature X:

[0062]

[0063] Step S3: Identification of thermal comfort state of occupants, using the extracted features to accurately identify the thermal comfort level through a machine learning algorithm;

[0064] S3.1: Thermal comfort identification of high-speed train passengers includes two parts: overall thermal sensation and local thermal sensation. The established data set includes passengers' HRV characteristics, GSR characteristics, wrist skin temperature, skin temperature of various parts of the face and average skin temperature of the face, as well as population division, seat identification and season;

[0065] S3.2: For overall thermal sensation recognition, HRV features, GSR features, facial mean skin temperature MSTF and wrist skin temperature are integrated as peripheral physiological recognition features; for local thermal sensation recognition of the face, the skin temperature of the corresponding part is selected, and group category labels are used to enhance the recognition accuracy of overall and local thermal sensations;

[0066] S3.3: Input the pre-processed feature data set into the machine learning model for training. During the training process, the performance of the model is optimized and overfitting is prevented through data set partitioning, cross-validation and hyperparameter optimization strategies.

[0067] S3.4: K-nearest neighbor algorithm is an instance-based non-parametric classification algorithm. It selects the nearest K samples by calculating the distance between the sample to be classified and each sample in the training set, and votes and classifies them according to the labels of these samples.

[0068] Step S4: construct a real-time monitoring and transmission device for the thermal comfort status of passengers in the vehicle, including a physiological data and group information collection module, a data preprocessing and feature fusion module, an occupant thermal comfort status recognition module, and a group decision-making and IoT transmission module.

[0069] Example 3, based on Example 2, collects the following physiological characteristic input information table:

[0070] The following environmental diagnostic inputs were collected:

[0071]

[0072] The following group information characteristics were collected:

[0073] The following local thermal sensory outputs were collected:

[0074] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A high-speed train passenger thermal comfort identification system based on peripheral physiological characteristics, characterized by: The system includes a physiological data and group information acquisition module, a data preprocessing and feature fusion module, an occupant thermal comfort state recognition module, and a group decision-making and IoT transmission module; the physiological data and group information acquisition module is responsible for collecting occupant peripheral physiological signals from thermal imaging cameras and smart watch devices; The data preprocessing and feature fusion module is used to receive the collected signals and perform feature preprocessing and fusion on the collected signals; The occupant thermal comfort state recognition module receives real-time changes in the occupant's thermal comfort state; The group decision-making and IoT transmission module stores the collected thermal physiological signals and thermal comfort status categories, aggregates them into group thermal comfort levels, and transmits them to the air-conditioning control system in the train.

2. A high-speed train occupant thermal comfort identification system based on peripheral physiological characteristics according to claim 1, characterized in that: The physiological data and group information acquisition module collects group category information corresponding to passengers from the ticket purchasing system. The physiological data and group information acquisition module includes a digital-to-analog converter, a data acquisition card, and a data transmission interface component, which is used to convert real-time scenes into digital signals for processing.

3. The high-speed train occupant thermal comfort identification system based on peripheral physiological characteristics according to claim 1 is characterized by: The data preprocessing and feature fusion module is connected to the physiological data and group information collection module and the occupant thermal comfort state recognition module, receives the collected signals, and performs feature preprocessing and fusion on the collected signals to facilitate classification by the thermal comfort level recognition module. The data preprocessing and feature fusion module is composed of a data processing unit, a cache unit, and a data wireless transmission unit.

4. The high-speed train occupant thermal comfort identification system based on peripheral physiological characteristics according to claim 1 is characterized by: The occupant thermal comfort state recognition module is connected to the data preprocessing and feature fusion module and the group decision-making and Internet of Things transmission module, receives real-time changes in the occupant thermal comfort state, and accelerates the response speed and accuracy of the air conditioning air supply strategy by transmitting the information to the air conditioning control system in real time; the occupant thermal comfort state recognition module is composed of a central data processing unit, a cache unit, a data wireless transmission unit and a data transmission interface.

5. The high-speed train occupant thermal comfort identification system based on peripheral physiological characteristics according to claim 1 is characterized by: The group decision-making and Internet of Things transmission module is connected to the passenger thermal comfort status identification module, stores the collected thermal physiological signals and thermal comfort status categories, aggregates them into group thermal comfort levels, and transmits them to the air-conditioning control system in the train. The group decision-making and Internet of Things transmission module consists of a control chip, a storage controller, a data wireless transmission unit and an interface circuit.

6. A method for identifying thermal comfort of high-speed train passengers based on peripheral physiological characteristics according to the system for identifying thermal comfort of high-speed train passengers based on peripheral physiological characteristics as claimed in any one of claims 1 to 5, characterized in that: The method comprises the following steps: Step S1: constructing a train occupant thermal comfort dataset based on peripheral physiological characteristics; Step S2: data preprocessing and feature fusion; Step S3: Identification of thermal comfort state of occupants, using the extracted features to accurately identify the thermal comfort level through a machine learning algorithm; Step S4: construct a real-time monitoring and transmission device for the thermal comfort status of passengers in the vehicle.

7. The method for identifying thermal comfort of high-speed train passengers based on peripheral physiological characteristics according to claim 6, characterized in that: The specific steps of step S1 are as follows: Step S1.1: Under different seasons and different air-conditioning system working conditions, the peripheral physiological signals of the subjects are fully collected, including skin temperature of various parts of the face, galvanic skin response GSR and blood volume pulse BVP, and the subjective thermal comfort questionnaire of the subjects is collected. The heart rate variability HRV can be extracted from the blood volume pulse BVP. Heart rate variability HRV refers to the change in the number of heart beats per minute, which reflects the fluctuation of the heartbeat interval. The time domain index SDNN and nonlinear index can be calculated by the following formula. SDNN is the calculation result of the standard deviation of all normal heartbeat intervals NN intervals, which is used to reflect the overall level of overall heart rate variability. N represents the total number of NN intervals, and NNi represents the i-th N interval. is the average value of all NN intervals; SD1 and SD2 are geometric analysis indicators based on the Poincaré scatter plot, which are used to quantify instantaneous heart rate variability and long-term heart rate variability. The Poincaré scatter plot is a point plotting each heart beat interval with its previous interval on a two-dimensional plane, i.e., point (NNi, NNi+1); SD1 represents the standard deviation on the short axis of the scatter plot, reflecting the instantaneous component of heart rate variability, and is used to evaluate parasympathetic nerve activity; SD2 represents the standard deviation on the long axis of the scatter plot, reflecting the long-term component of heart rate variability, and is used to evaluate the joint action of sympathetic and parasympathetic nerves; define the standard deviation on the diagonal and vertical lines, and calculate SD1 and SD2 respectively; where Var(NNi-NNi+1) represents the variance of the difference between all consecutive NN intervals; For GSR data, it is uniformly converted into a change value from the baseline, i.e., ΔGSR, to reduce the measurement error under different air-conditioning conditions in different seasons. The baseline GSR value is defined as the lowest GSR value within the first 2-5 minutes of the measurement. The average facial skin temperature MSTF is obtained according to the area-weighted value of each part, where Tsk is the local skin temperature of the face, dAsk is the corresponding local surface area of ​​each part, and A is the total facial area; MSTF=∫T sl dA sk / A = forehead local skin temperature × 0.2 + eye local skin temperature × 0.1 + cheek local skin temperature × 0.4 + nose local skin temperature × 0.1 + ear local skin temperature × 0.2; S1.2: Group information collection. Adding differentiated categories can improve the performance and accuracy of the model. Therefore, the group category labels of the riding season, seat row and seat column, gender, BMI and age corresponding to each passenger are collected to enhance the model's ability to recognize the target.

8. The method for identifying thermal comfort of high-speed train passengers based on peripheral physiological characteristics according to claim 6, characterized in that: The specific steps of step S2 are as follows: S2.1: Use LabelEncoder and StandardScaler tools to standardize each identification factor feature. The LabelEncoder tool encodes the group category feature label using a value between 0 and n_classes-1, and the StandardScaler tool standardizes the numerical peripheral physiological features. S2.2: The StandardScaler tool is based on data standardization. It performs linear transformation on numerical features so that each feature, that is, each attribute or column of data, follows a standard normal distribution, with a mean of 0 and a standard deviation of 1. For a given eigenvalue X, StandardScaler standardizes each feature. The transformed eigenvalue X′ is calculated according to the following formula, where μ is the mean of feature X and σ is the standard deviation of feature X:

9. The method for identifying thermal comfort of high-speed train passengers based on peripheral physiological characteristics according to claim 6, characterized in that: The specific steps of step S3 are as follows: S3.1: Thermal comfort identification of high-speed train passengers includes two parts: overall thermal sensation and local thermal sensation. The established data set includes passengers' HRV characteristics, GSR characteristics, wrist skin temperature, skin temperature of various parts of the face and average skin temperature of the face, as well as population division, seat identification and season; S3.2: For overall thermal sensation recognition, HRV features, GSR features, facial mean skin temperature MSTF and wrist skin temperature are integrated as peripheral physiological recognition features; for local thermal sensation recognition of the face, the skin temperature of the corresponding part is selected, and group category labels are used to enhance the recognition accuracy of overall and local thermal sensations; S3.3: Input the pre-processed feature data set into the machine learning model for training. During the training process, the performance of the model is optimized and overfitting is prevented through data set partitioning, cross-validation and hyperparameter optimization strategies. S3.4: K-nearest neighbor algorithm is an instance-based non-parametric classification algorithm. It selects the nearest K samples by calculating the distance between the sample to be classified and each sample in the training set, and votes and classifies them according to the labels of these samples.

10. The method for identifying thermal comfort of high-speed train passengers based on peripheral physiological characteristics according to claim 6, characterized in that: The real-time monitoring and transmission device for the thermal comfort status of the occupants in step S4 includes a physiological data and group information acquisition module, a data preprocessing and feature fusion module, an occupant thermal comfort status recognition module, and a group decision-making and IoT transmission module.

Images