Capacitive sensing-based seat comfort detection system and comfort evaluation method
Through a capacitive sensing-based seat comfort detection system, combined with machine learning models, the existing seat comfort evaluation method is solved by subjects and seat morphological factors, and the precise evaluation of seat comfort and ergonomic design support is achieved.
Patent Information
- Application Number
- CN202510468687.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-15
- Publication Date
- 2025-08-01
AI Technical Summary
The existing seat comfort evaluation methods are affected by the subjects' own characteristics and seat morphology factors, and the evaluation is insufficient and ergonomic design is not fully considered.
The seat comfort detection system based on capacitive sensing is adopted, including a sensor module, a capacitance measurement module, a data processing module and a comfort evaluation module. The capacitive sensor is used to accurately sense the pressure distribution of the seat and the human body contact surface, and the comfort evaluation is carried out in combination with a machine learning model.
Accurate evaluation of seat comfort is achieved, and can analyze different human characteristics and seat designs, providing overall and local comfort scores and thermal maps, improving the objectivity and accuracy of the assessment.
Smart Images
Figure CN120404175A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of seat comfort detection systems, and specifically to a seat comfort detection system and a comfort evaluation method based on capacitive sensing. Background Art
[0002] With the improvement of people's living standards and the enhancement of the requirements for the quality of life, the attention to seat comfort has been increasing. The existing seat comfort evaluation methods include subjective evaluation method, physical measurement method, and biomechanical measurement method. These three methods mainly rely on the subjective perception and evaluation of the subjects on the seat comfort, measure the physical characteristics of the seat, such as hardness, elasticity, temperature, etc., and measure the physiological indexes of the human body when sitting on the seat, such as muscle activity, pressure distribution, etc. to simply evaluate the seat comfort. However, these evaluation methods are affected by factors such as the characteristics of the subjects themselves and the shape of the seat, and do not fully consider whether the ergonomic design is reasonable, and the rationality of their evaluation is insufficient. The present invention proposes a seat comfort detection facility based on capacitive sensing, aiming to overcome the deficiencies of the existing technology, and has the advantages of small volume, high sensitivity, low cost, good durability, and easy integration, making it an ideal choice for seat comfort detection. Summary of the Invention
[0003] To solve the above technical problems, the present invention proposes a seat comfort detection system and a comfort evaluation method based on capacitive sensing, which can be used to detect whether the shape of a seat conforms to ergonomics and can bring a better sitting experience.
[0004] To achieve the above object, the technical solution adopted by the present invention is:
[0005] A seat comfort detection system based on capacitive sensing, the system includes:
[0006] A sensor module, including a five-layer capacitive structure of two shielding layers, two conductive layers, and one dielectric layer, where the dielectric layer is in the middlemost position, the two conductive layers are separated on both sides of the dielectric layer, and the outermost layer is wrapped by two shielding layers; at the same time, the five layers are pasted after being placed in sequence to achieve fixation;
[0007] A capacitance measurement module, including a mutual capacitance measurement circuit, a microcontroller Arduino, and a control switch circuit. The mutual capacitance measurement circuit has the function of sending an excitation signal at the Tx end and receiving and processing the signal at the Rx end, and supports data interaction with the microcontroller Arduino through the I2C communication protocol. The microcontroller Arduino is responsible for providing power and controlling the mutual capacitance measurement circuit to complete the measurement of the capacitance value;
[0008] A data processing module for processing the measured capacitance values and mapping the obtained sensor readings into a two-dimensional matrix, which is the capacitance value corresponding to the position.
[0009] A comfort evaluation module for comprehensively evaluating the comfort of the seat. This module receives the recognized capacitance values and the user's body characteristic information and inputs them into a pre-deployed machine learning seat comfort evaluation model for calculation.
[0010] A model deployment module for pre-training and deploying a machine learning seat comfort evaluation model, including a user perception unit, a signal receiving unit, a normalization unit, a key point annotation unit, and a training unit, a total of five units.
[0011] Furthermore, the shielding layer and the dielectric layer of the sensor module are made of insulating fabric, and the conductive layer is a conductive material in the shape of a rhombus with a side length of 5 mm and a spacing of 2 mm.
[0012] Furthermore, the evaluation results of the comfort evaluation module include:
[0013] 1) Seat overall comfort score: Calculate the overall comfort index based on the whole-body capacitance distribution characteristics and the user's individual characteristics.
[0014] 2) Seat local comfort score: Calculate the local comfort index based on the whole-body capacitance distribution characteristics and the user's individual characteristics. The local areas include the head, upper back, middle back, lower back, buttocks, thighs, and calves.
[0015] 3) Comfort heat map: Based on the seat local comfort, convert the capacitance distribution situation into a comfort situation and generate a heat map to visually present the comfort distribution of each area of the seat.
[0016] Furthermore, in the model deployment module;
[0017] The user perception unit is used to subjectively evaluate the comfort of the seat. For a specific seat, multiple subjects with different body types are invited to take a test sit, and their individual characteristic information such as height, weight, BMI, limb length, and torso length is collected. The subjects use a standardized comfort assessment scale to score the comfort of the seat. The scale covers multiple evaluation dimensions such as overall comfort and local comfort, and the local comfort is divided into the head, upper back, middle back, lower back, buttocks, thighs, and calves.
[0018] The signal receiving unit is used to receive the capacitance values of the target seat during use obtained by the sensor module and the comfort scores obtained by the corresponding user perception unit.
[0019] The normalization unit is used to normalize the capacitance values processed by the calibration matrix.
[0020] The key point annotation unit collects capacitance sensor data, combines individual characteristic information and external reference devices to obtain human key point information, and establishes the corresponding relationship between capacitance distribution and human key points.
[0021] The training unit includes at least two sub-models, which are respectively used for human key point recognition, overall comfort evaluation, and local comfort evaluation. Each model uses the deep convolutional neural network ResNet for feature extraction and learning.
[0022] The comfort heat map drawing unit constructs a scalar field for local comfort evaluation based on capacitance values, and visually converts it into a heat map to achieve refined analysis of the comfort of different parts of the human body.
[0023] Furthermore, the external reference device of the key point annotation unit is an optical camera, a pressure sensor or an IMU.
[0024] For the seat comfort detection system based on capacitive sensing of the present invention, the specific functions of each sub-model of the training unit are as follows:
[0025] 1) The human body structure recognition model is used to analyze the mapping relationship between the two-dimensional capacitance distribution map and the human anatomical structure, and takes the normalized capacitance value, the individual characteristic information collected by the user perception unit, and the labeled human key point data as inputs to train the human key point recognition model.
[0026] 2) The overall comfort evaluation model is used to learn the relationship between capacitance data and the overall comfort score, and takes the normalized capacitance value, the individual characteristic information collected by the user perception unit, and the overall comfort obtained based on the standardized comfort evaluation scale as inputs to train the overall comfort score model.
[0027] 3) The local comfort evaluation model combines human body structure information and capacitance values to evaluate the comfort of local areas, and takes the normalized capacitance value, the individual characteristic information collected by the user perception unit, the local comfort obtained based on the standardized comfort evaluation scale, and the human body contour information provided by the human body structure recognition model as inputs to train the local comfort score model.
[0028] The present invention provides a comfort evaluation method for a seat comfort detection system based on capacitive sensing, which is characterized in that the steps include:
[0029] S1. Obtain the capacitance numerical information corresponding to the seat to be measured in the manned state through the sensor module and the capacitance measurement module.
[0030] S2. Obtain individual characteristic information of the subject, such as height, weight, BMI, limb length, trunk length, etc.
[0031] S3. Calibrate the data measured in S1 and use it as the identified capacitance value;
[0032] S4. Input the identified capacitance value obtained in step S3 and the individual characteristic information such as the subject's height, weight, BMI, limb length, and trunk length obtained in S2 into a pre-deployed machine learning identification model for identification, and output a comfort evaluation result.
[0033] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0034] By using capacitive sensing technology, it can accurately sense the change in the pressure distribution on the contact surface between the seat and the human body, with high sensitivity; the capacitive sensor has a relatively low cost, a small volume, and is easy to integrate into the seat structure, facilitating large-scale production and application; this device can realize multi-point measurement of the pressure on the contact surface between the seat and the human body, and can process and analyze the collected data through data analysis software to more comprehensively evaluate the comfort of the seat, and can conduct targeted analysis on different human characteristics and seat designs. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] Figure 1 is a schematic diagram of the sensor module of the present invention;
[0036] Figure 2 is a schematic diagram of the capacitance measurement module of the present invention;
[0037] Figure 3 is an overall assembly drawing of the present invention.
[0038] REFERENCE SIGNS:
[0039] 1. Seat; 2. Sensor module; 3. First capacitance measurement module; 4. Second capacitance measurement module; 5. Host computer. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0040] The present invention will be further described in detail below in conjunction with the drawings and the specific embodiments:
[0041] Figure 1 is the sensor module, including a five-layer capacitance structure of two shielding layers, two conductive layers and one dielectric layer. The dielectric layer is in the middle position, the two conductive layers are separated on both sides of the dielectric layer, and the outermost is wrapped by two shielding layers; the shielding layer and the dielectric layer are made of insulating cloth, and the conductive layer is a conductive material in the shape of a rhombus with a side length of 5 mm and a spacing of 2 mm. At the same time, the five layers are placed in sequence and then pasted to achieve fixation.
[0042] Figure 2It is a capacitance measurement module, including a mutual capacitance measurement circuit, a microcontroller Arduino, and a control switch circuit. Structure 2 is a sensor module, which is a cross electrode where the horizontal electrode and the vertical electrode do not contact each other. To ensure the insulation performance of the electrodes, insulating cloth is used to wrap both the front and back sides of the electrodes. Structure 3 is a mutual capacitance measurement circuit, which has the function of sending an excitation signal at the Tx end and receiving and processing signals at the Rx end, and can accurately complete the measurement task of the mutual capacitance in the circuit. In addition, it also supports data interaction with the microcontroller Arduino through the I2C communication protocol. Structure 4 is the microcontroller Arduino, which is responsible for providing power and controlling the mutual capacitance measurement circuit to complete the measurement of capacitance values. Structure 2 is connected to Structure 3, and Structure 3 is connected to Structure 4 respectively using wires. Structure 4 is connected to the power supply and the host computer respectively.
[0043] The principle of the mutual capacitance measurement circuit is as follows:
[0044] 1) Charge integration. Through switch control, the chip makes the Tx electrode apply a high-frequency pulse signal, whose electric field is coupled to the Rx electrode, and the change in capacitance is measured through a charge integration circuit Q = C mutual ·V tx , where Q is the electric charge quantity and V tx is the excitation voltage applied by Tx.
[0045] 2) Voltage conversion and ADC sampling. The analog-to-digital converter (ADC) converts the measured electric charge quantity into a voltage signal: where C feedback is the feedback capacitance built into the chip. Thus, we can calculate the capacitance value C mutual at a certain intersection point through the voltage input at the Tx end and the voltage output at the Rx end.
[0046] Figure 3 For the overall assembly, Structure 1 is the seat, Structure 2 is the sensor module, Structures 3 and 4 are the capacitance measurement modules, and Structure 5 is the host computer. When working normally, when a person contacts the seat, it will cause corresponding changes in the voltage values on the electrodes. The mutual capacitance measurement circuit can scan the touch coordinates at a rate of up to 60 Hz at the fastest, then measure the voltage value on the vertical electrode, and communicate with the microcontroller Arduino through the I2C protocol, so as to obtain the capacitance readings C mutual of each positive intersection point of the tactile sensor.
[0047] The data processing module is used to process the measured capacitance values. Map the obtained sensor readings into a two-dimensional matrix, that is, the capacitance values corresponding to the positions. Since in the case of no touch on the sensor module, the background capacitance C baselineThey may also be different, so it is necessary to establish a calibration matrix to unify the initial reference and enable it to respond to calibration requests. The steps of data processing are as follows:
[0048] 1) In the initial non-touch situation, measure each sensor unit (i.e., the capacitance value at the intersection of TX and RX) 10 times and take the average value: where: C k (i,j) is the capacitance value measured for the kth time, and C baseline (i,j) is the initial calibration capacitance at this position. Then the calibration matrix can be obtained.
[0049] 2) In the case of contact, obtain the sensor module reading C mutual . Then, by subtracting the directly measured capacitance value from the calibration capacitance value, the calibrated capacitance value can be obtained. That is, C new = C mutual - C baseline .
[0050] The comfort evaluation module is used to comprehensively evaluate the comfort of the seat. This module receives the recognized capacitance value C new and the user's body characteristic information such as height, weight, etc., and inputs them into the pre-deployed machine learning seat comfort evaluation model for calculation. The evaluation results include:
[0051] 1) Seat overall comfort score: Calculate the overall comfort index based on the whole-body capacitance distribution characteristics and the user's individual characteristics.
[0052] 2) Seat local comfort score: Calculate the local comfort index based on the whole-body capacitance distribution characteristics and the user's individual characteristics. The local areas include the head, upper back, middle back, lower back, buttocks, thighs, calves, etc.
[0053] 3) Comfort heat map: Based on the seat local comfort, convert the capacitance distribution situation into a comfort situation and generate a heat map to visually present the comfort distribution of each area of the seat.
[0054] The model deployment module is used to pre-train and deploy the machine learning seat comfort evaluation model. The model deployment section includes:
[0055] The user perception unit is used to subjectively evaluate the comfort of the seat. For a specific seat, invite multiple subjects with different body types to try sitting and collect their individual characteristic information such as height, weight, BMI, limb length, torso length, etc. The subjects use a standardized comfort assessment scale to score the comfort of the seat. The scale covers multiple evaluation dimensions such as overall comfort and local comfort, and the local comfort is divided into the head, upper back, middle back, lower back, buttocks, thighs, calves, etc.;
[0056] A signal receiving unit, configured to receive the capacitance value when the target seat is in use acquired by the sensor module and the comfort evaluation obtained by the corresponding user perception unit.
[0057] A normalization unit, configured to perform normalization processing on the capacitance value processed by using a calibration matrix, and the steps are as follows:
[0058] 1) Calculate the mean μ and standard deviation σ of the full data set:
[0059] 2) Normalize each sample:
[0060] A key point annotation unit, which will collect capacitance sensor data, combine individual characteristic information and external reference devices (optical camera, pressure sensor or IMU) to obtain human key point information, and establish a correspondence between the capacitance distribution and human key points.
[0061] A training unit, including multiple sub-models, which are respectively used for human key point recognition, overall comfort evaluation, and local comfort evaluation. Each model uses a deep convolutional neural network (ResNet) for feature extraction and learning. The specific functions are as follows:
[0062] 1) A human body structure recognition model (Model-1), which is used to analyze the mapping relationship between the two-dimensional capacitance distribution map and the human anatomical structure. Taking the normalized capacitance value, the individual characteristic information (such as height and weight) collected by the user perception unit, and the labeled human key point data as inputs, a human key point recognition model is trained.
[0063] 2) An overall comfort evaluation model (Model-2), which is used to learn the relationship between capacitance data and the overall comfort score. Taking the normalized capacitance value, the individual characteristic information (such as height and weight) collected by the user perception unit, and the overall comfort obtained based on a standardized comfort evaluation scale as inputs, an overall comfort score model is trained.
[0064] 3) A local comfort evaluation model (Model-3), which combines human body structure information (output of Model-1) and capacitance value to evaluate the comfort of a local area. Taking the normalized capacitance value, the individual characteristic information (such as height and weight) collected by the user perception unit, the local comfort obtained based on a standardized comfort evaluation scale, and the human body contour information provided by Model-1 as inputs, a local comfort score model is trained.
[0065] In the specific training of the three models, a deep convolutional neural network ResNet is used for training, and the specific principle is as follows:
[0066] 1) Feature extraction and residual learning. ResNet mainly relies on residual blocks to extract the spatial features of capacitance changes. The calculation form of each residual block is as follows: where x is the input feature, represents the transformed feature after multiple convolutional operations, W is the convolutional kernel parameter in the residual block, and y is the output feature.
[0067] 2) Forward propagation. Calculate the network output and calculate the loss value based on the objective function. The cross-entropy loss function is selected as the objective function for training: where y i is the one-hot encoding of the true class, is the class probability predicted by the model.
[0068] 3) Backward propagation. Calculate the gradient and use AdamW for parameter update. The specific rule is: where θ t represents the grid parameters in the t-th round, η is the learning rate, m t and v t are the first moment (momentum) and the second moment (exponential moving average of the gradient square) of the gradient respectively, and λ is the regularization term.
[0069] 4) Adjust the learning rate. Adopt the cosine annealing learning rate adjustment strategy to gradually reduce the learning rate in the later stage of training. The specific rule is: where η t is the learning rate in the t-th round, η max is the initial maximum learning rate, T cur is the current number of steps, and T max is the total number of training steps.
[0070] 5) Save the optimal model weights.
[0071] The comfort heatmap drawing unit constructs a scalar field for local comfort evaluation based on the capacitance value and visualizes it into a heatmap to achieve a refined analysis of the comfort of different parts of the human body. The specific steps are as follows:
[0072] 1) Since there are differences in the perception of comfort in different parts of the human body, the same capacitance value may correspond to different comfort levels in different parts. Therefore, in the above Model-1, the capacitance threshold ranges of each part related to comfort are obtained.
[0073] 2) In the process of comfort evaluation, an adaptive normalization strategy is adopted for the capacitance values of different parts to improve the comparability of capacitance values of different parts. Specifically, the normalization formula is as follows: where C thresh,max and Cthresh,min It is determined by the capacitance threshold obtained through training, so as to ensure that the normalized capacitance value can accurately reflect the local comfort level.
[0074] 3) The designed scalar field includes: the normalized capacitance value C norm , which is used to quantitatively evaluate the local comfort level; the capacitance value gradient information, which is used to characterize the capacitance change trend of this point and the surrounding area. Using this scalar field, a whole-body comfort heat map can be constructed to intuitively display the comfort distribution of each part of the human body. This heat map uses color gradient to map the comfort level, so as to realize refined and quantitative comfort evaluation.
[0075] When this system is deployed in hardware, it can be composed of four main parts: a seat, a sensor module, a capacitance measurement module, and a host computer. The sensor module is fixed on the surface of the seat and is connected to the capacitance measurement module. Structure 4 in the capacitance measurement module is connected to the host computer to achieve two-way communication. The functions include receiving and storing data from the capacitance sensor, processing and analyzing the data, training and deploying the seat comfort detection model, and using a visualization interface to display the measured sensor values and the corresponding seat comfort evaluation in real time; among them, data processing includes the normalization of the capacitance array data, and then the comfort evaluation is carried out through the machine learning seat comfort evaluation model.
[0076] The present invention also provides a method for evaluating the comfort of a seat based on a capacitance sensor. The steps include:
[0077] S1. Obtain the capacitance numerical information corresponding to the seat to be measured in the manned state through the sensor module and the capacitance measurement module;
[0078] S2. Obtain the individual characteristic information of the subject, such as height, weight, BMI, limb length, trunk length, etc.;
[0079] S3. Calibrate the data measured in S1 and use it as the recognition capacitance value;
[0080] S4. Input the recognition capacitance value obtained in step S3 and the individual characteristic information of the subject such as height, weight, BMI, limb length, trunk length, etc. obtained in S2 into the pre-deployed machine learning recognition model for recognition, and output the comfort evaluation result.
[0081] The above is only a preferred embodiment of the present invention, and it is not any other form of limitation to the present invention. Any modification or equivalent change made according to the technical essence of the present invention still belongs to the scope protected by the present invention.
Claims
1. A seat comfort detection system based on capacitive sensing, characterized in that The system includes: A sensor module, which is a five-layer capacitance structure including two shielding layers, two conductive layers and one dielectric layer. The dielectric layer is in the middle position, the two conductive layers are separated on both sides of the dielectric layer, and the outermost part is wrapped by two shielding layers. At the same time, the five layers are pasted after being placed in order to achieve fixation; A capacitance measurement module, which includes a mutual capacitance measurement circuit, a microcontroller Arduino and a control switch circuit. The mutual capacitance measurement circuit has the function of sending an excitation signal at the Tx end and receiving and processing signals at the Rx end, and supports data interaction with the microcontroller Arduino through the I2C communication protocol. The microcontroller Arduino is responsible for providing power and controlling the mutual capacitance measurement circuit to complete the measurement of capacitance values; A data processing module, which is used to process the measured capacitance values and map the obtained sensor readings into a two-dimensional matrix, that is, the capacitance values corresponding to positions; A comfort evaluation module, which is used to comprehensively evaluate the comfort of the seat. This module receives the identified capacitance values and the user's human characteristic information, and inputs them into a pre-deployed machine learning seat comfort evaluation model for calculation; A model deployment module, which is used to pre-train and deploy a machine learning seat comfort evaluation model, including five units: a user perception unit, a signal reception unit, a normalization unit, a key point annotation unit, and a training unit.
2. The seat comfort detection system based on capacitive sensing according to claim 1, wherein The shielding layer and dielectric layer of the sensor module are made of insulating cloth, and the conductive layer is a conductive material in the shape of a rhombus with a side length of 5 mm and a spacing of 2 mm.
3. The seat comfort detection system based on capacitive sensing according to claim 1, characterized in that, The evaluation results of the comfort evaluation module include: 1) Seat overall comfort score: Calculate the overall comfort index based on the whole-body capacitance distribution characteristics and user individual characteristics; 2) Seat local comfort score: Calculate the local comfort index based on the whole-body capacitance distribution characteristics and user individual characteristics. The local areas include the head, upper back, middle back, lower back, buttocks, thighs, and calves; 3) Comfort heat map: Based on the local comfort of the seat, convert the capacitance distribution situation into a comfort situation and generate a heat map to visually present the comfort distribution of each area of the seat.
4. The seat comfort detection system based on capacitive sensing according to claim 1, characterized in that, In the model deployment module; The user perception unit is used to subjectively evaluate the comfort of the seat. For a specific seat, multiple subjects with different body types are invited to take a test ride, and their individual characteristic information such as height, weight, BMI, limb length, and trunk length is collected. The subjects use a standardized comfort assessment scale to score the comfort of the seat. The scale covers multiple evaluation dimensions such as overall comfort and local comfort. Among them, the local comfort is divided into the head, upper back, middle back, lower back, buttocks, thighs, and calves; The signal reception unit is used to receive the capacitance values of the target seat during use obtained by the sensor module and the comfort scores obtained by the corresponding user perception unit; The normalization unit is used to normalize the capacitance values processed by the calibration matrix; The key point annotation unit collects capacitance sensor data, combines individual characteristic information and an external reference device to obtain human key point information, and establishes the corresponding relationship between the capacitance distribution and human key points; The training unit includes at least two sub-models, which are respectively used for human key point recognition, overall comfort evaluation, and local comfort evaluation. Each model uses the deep convolutional neural network ResNet for feature extraction and learning; The comfort heat map drawing unit constructs a scalar field for local comfort evaluation based on the capacitance value and visually converts it into a heat map to achieve refined analysis of the comfort of different parts of the human body.
5. The seat comfort detection system based on capacitive sensing according to claim 1, characterized in that, The external reference device of the key point annotation unit is an optical camera, a pressure sensor or an IMU.
6. The seat comfort detection system based on capacitive sensing according to claim 1, characterized in that, The specific functions of each sub-model of the training unit are as follows: 1) The human body structure recognition model is used to analyze the mapping relationship between the two-dimensional capacitance distribution map and the human anatomical structure. The normalized capacitance value, the individual feature information collected by the user perception unit, and the labeled human key point data are used as inputs to train the human key point recognition model; 2) The overall comfort evaluation model is used to learn the relationship between the capacitance data and the overall comfort score. The normalized capacitance value, the individual feature information collected by the user perception unit, and the overall comfort obtained based on the standardized comfort evaluation scale are used as inputs to train the overall comfort score model; 3) The local comfort evaluation model combines the human body structure information and the capacitance value to evaluate the comfort of the local area. The normalized capacitance value, the individual feature information collected by the user perception unit, the local comfort obtained based on the standardized comfort evaluation scale, and the human body contour information provided by the human body structure recognition model are used as inputs to train the local comfort score model.
7. The comfort evaluation method of the seat comfort detection system based on capacitive sensing according to any one of claims 1-6, characterized in that the steps It includes: S1. Obtain the capacitance numerical information corresponding to the seat to be measured in the manned state through the sensor module and the capacitance measurement module; S2. Obtain individual feature information of the subject, such as height, weight, BMI, limb length, trunk length, etc.; S3. Calibrate the data measured in S1 and use it as the recognized capacitance value; S4. Input the recognized capacitance value obtained in step S3 and the individual feature information of the subject such as height, weight, BMI, limb length, trunk length, etc. obtained in S2 into the pre-deployed machine learning recognition model for recognition, and output the comfort evaluation result.