CN113128031A - Measuring device and evaluation method for comfort level of seat - Google Patents
Measuring device and evaluation method for comfort level of seat Download PDFInfo
- Publication number
- CN113128031A CN113128031A CN202110345532.9A CN202110345532A CN113128031A CN 113128031 A CN113128031 A CN 113128031A CN 202110345532 A CN202110345532 A CN 202110345532A CN 113128031 A CN113128031 A CN 113128031A
- Authority
- CN
- China
- Prior art keywords
- seat
- comfort level
- human
- model
- pressure distribution
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000011156 evaluation Methods 0.000 title claims abstract description 49
- 238000000034 method Methods 0.000 claims abstract description 35
- 238000003062 neural network model Methods 0.000 claims abstract description 11
- 210000001624 hip Anatomy 0.000 claims description 22
- 210000002414 leg Anatomy 0.000 claims description 18
- 238000013528 artificial neural network Methods 0.000 claims description 17
- 239000000463 material Substances 0.000 claims description 16
- 210000000689 upper leg Anatomy 0.000 claims description 15
- 238000004088 simulation Methods 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 11
- 210000001519 tissue Anatomy 0.000 claims description 7
- 101000911772 Homo sapiens Hsc70-interacting protein Proteins 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 5
- 210000003205 muscle Anatomy 0.000 claims description 5
- 210000000115 thoracic cavity Anatomy 0.000 claims description 4
- 210000000988 bone and bone Anatomy 0.000 claims description 3
- 238000000513 principal component analysis Methods 0.000 claims description 3
- 210000001991 scapula Anatomy 0.000 claims description 3
- 238000013210 evaluation model Methods 0.000 claims description 2
- 210000004197 pelvis Anatomy 0.000 claims description 2
- 239000002861 polymer material Substances 0.000 claims description 2
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 210000001694 thigh bone Anatomy 0.000 claims description 2
- 239000000284 extract Substances 0.000 claims 2
- 238000012847 principal component analysis method Methods 0.000 claims 1
- 238000013461 design Methods 0.000 abstract description 7
- 238000007621 cluster analysis Methods 0.000 abstract description 2
- 230000000875 corresponding effect Effects 0.000 description 13
- 238000004458 analytical method Methods 0.000 description 11
- 238000005259 measurement Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000012827 research and development Methods 0.000 description 4
- 239000002023 wood Substances 0.000 description 4
- 238000012360 testing method Methods 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000037237 body shape Effects 0.000 description 1
- 230000037396 body weight Effects 0.000 description 1
- 210000001217 buttock Anatomy 0.000 description 1
- 239000011248 coating agent Substances 0.000 description 1
- 238000000576 coating method Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000003703 image analysis method Methods 0.000 description 1
- 239000010985 leather Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 210000001747 pupil Anatomy 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 210000002027 skeletal muscle Anatomy 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
Abstract
The invention relates to a measuring device and an evaluation method for the comfort level of a seat, which mainly comprise a human body model device design and a seat use comfort level evaluation prediction method and a model. The human body model device is characterized by simulating real persons with different percentiles, and is used for measuring body pressure distribution signals on the surface of the seat and reducing the influence of individual difference. The method for evaluating the comfort level of the seat is completed through a statistical principle and a neural network model, the comfort level of the seat is obtained through assigning values to cluster clusters obtained through system cluster analysis of body pressure distribution characteristic values of different seat surfaces, the classification comprises 90% of Chinese human percentile population, and the method has universality and does not depend on subjective evaluation. The result of the model for predicting the comfort level of the seat is consistent with the result regularity of the body pressure distribution characteristic value, and the result shows that the neural network model can predict the use comfort level of the seat through the body pressure distribution characteristic value and anthropometric data of a user.
Description
Technical Field
The invention belongs to a device and a method for evaluating comfort level of a seat, and particularly relates to a human hip, leg and backrest model for replacing a real person in seat surface body pressure distribution test and a comfort level evaluation method independent of subjective evaluation.
Background
At present, two comfort evaluation methods for seat products are mainly used, and the comfort evaluation method is based on the combination of subjective and objective indexes and pure objective evaluation indexes. The objective measurement and evaluation indexes comprise means such as biomechanics of a skeletal muscle system, electroencephalogram signals, pupil detection and analysis, sitting posture, waist and back electromyogram signal analysis, multi-modal physiological signal joint analysis and the like. In addition, the posture of the user is researched by a clustering and image analysis method, so that the use comfort of the seat is analyzed and judged according to the relevant angles of all parts of the sitting posture. The evaluation method is limited by experimental conditions and professional equipment in the using process, so that the evaluation method cannot be generally popularized, and biomechanics analysis is generally simulation research (such as finite element simulation analysis), so that the requirement on model accuracy is high, and the boundary condition is difficult to determine. Therefore, the evaluation method mainly used is to measure indirect objective indexes and combine the indirect objective indexes with subjective evaluation, such as body pressure distribution signals. Studies have shown a link between pressure distribution and comfort, so it is theoretically possible to use this index to evaluate the comfort of use of the seat. However, two problems which are not negligible in the using process of the method are that 1) the body pressure distribution has a complex nonlinear relation with the sitting posture, the body shape and the seat, and the direct quantification is difficult; 2) the subjective evaluation has errors such as randomness, individual difference and the like. To solve this problem, there are two common methods. One is to establish the relationship between the body pressure distribution signals and subjective evaluation through strong generalization ability of machine learning. However, the output signal used for modeling by the method is still the subjective comfort evaluation value, and the accuracy of the subjective evaluation cannot be widely accepted. And secondly, errors caused by randomness of subjective evaluation are reduced as much as possible, and an analytic hierarchy process, a fuzzy comprehensive evaluation method and the like are adopted. The method is based on experts and a statistical probability principle, the randomness of subjective evaluation is reduced theoretically, but the method needs to be evaluated by the experts every time the method is used, the method and the established model cannot be directly applied to the research and development and detection of users and seats, and more subjective evaluation values need to be collected for judgment according to a membership degree principle.
In addition, the distribution signals of the surface pressure of the seat are more dependent on real testees, the signal acquisition process is inconvenient, the testees need to cooperate, and the individual difference of the real testees is larger, so that the difference between the distribution signals of the surface pressure of the seat is also larger. Therefore, the problems can be avoided by adopting the human body model when the body pressure distribution signals of the surface of the seat are collected. The existing physical model is an H-point device used in the field of automobiles and aims at simulating the body pressure of seats such as office chairs, sofas and the like. The research results of the distribution device are relatively few, the shape of the H-point device is not similar to that of a real person, and the adjustment mode is not involved in the size adjustment part
Disclosure of Invention
The invention aims to provide a seat comfort evaluation method based on seat surface pressure distribution index system clustering and neural network prediction, which does not depend on subjective evaluation. The method is completely based on the pressure distribution index, and human body measurement parameters are comprehensively considered, subjective comfort level evaluation is not relied on, the influences of subjective randomness and perception difference can be removed, and the model has universality. And the human body model device is used for replacing a real person to measure the body pressure distribution on the surface of the seat, is used for quickly measuring, reduces the measurement cost and shortens the measurement time.
The purpose of the invention is realized by the following technical scheme:
the invention relates to a device and a method for measuring comfort level of a seat, which mainly comprise a human ischial tuberosity model device (1), a human hip and leg model device (2), a human backrest model device (3), a human skin tissue simulation material (4), a seat surface pressure extraction method (5), a seat surface pressure classification number and classification method (6) and a comfort level evaluation algorithm model (7). The human body ischial tuberosity device (1) is assembled in the human body hip and leg model device, the human body model backrest device is assembled with the human body hip and leg model device (2), the human body skin tissue simulation material (4) is coated on the surfaces of the human body hip and leg model device (2) and the human body backrest model device (3), and weights with different weights are configured on the human body hip and leg model device (2) and the human body backrest model device (3) to simulate different weights.
The human ischial tuberosity model device (1) is a frustum pyramid-shaped geometric body made of two hard incompressible materials. The distance between the two ischial tuberosities is adjustable to simulate different percentile adult human ischial tuberosities.
The geometric volume of the human ischial tuberosity model device (1) is in various specifications so as to simulate human skeleton data of different percentiles.
Thigh bones and pelvis bones of the human body hip and leg model device (2) are made of incompressible hard materials, human body thighs and hips are simulated by coating human body skin tissue simulation materials (4) on the surface of the device, and human bodies with different weights are simulated by using different counter weights in thigh areas and hip areas of the device.
The human body backrest model device (3) is made of an incompressible hard material, the back curved surface is made according to the back of a human body, and the side surface is S-shaped; wherein the thoracic vertebra part is backward convex, the lumbar vertebra part is forward convex, and the shape of the thoracic vertebra part comprises the shape of a scapula. The device is used for simulating the back form of a human body.
The human skin tissue simulation material (4) is divided into two parts of muscles and skins, is a structure which is made of soft and hard high polymer materials and has different densities and thicknesses, is used for simulating the muscle tissues of the waist, the back, the hip and the leg of a human body, and realizes the simulation of the human body of a real person with different percentiles by changing the material collocation of different thicknesses and densities.
The seat comfort level assessment method comprises the steps of seat surface body pressure distribution index extraction, comfort level assessment grade data set construction and comfort level assessment prediction model modeling.
The extraction of the body pressure distribution index of the surface of the seat and the construction of the comfort evaluation grade data set are completed by the following steps. Firstly, carrying out complete experiments by various percentile true persons (including m persons for both men and women) and a plurality of types and a plurality of numbers of seats (n) as much as possible, acquiring surface body pressure distribution signals of the seats to obtain mxn groups of data, dividing the seats into a plurality of clusters according to body pressure distribution index clustering results by extracting surface body pressure distribution indexes and system clustering analysis, and endowing the seats contained in different clusters with different comfort levels according to the quantity relation of the body pressure distribution index range of each cluster, namely considering that the seats contained in a certain cluster are in certain level comfort for most persons; secondly, extracting body pressure distribution signal indexes of the surfaces of the seats in each cluster within a 95% confidence interval, including corresponding anthropometric parameters (height, weight and sex), constructing a new comfort level evaluation index through principal component analysis, wherein the comfort level evaluation principal component index, the anthropometric parameters and the corresponding comfort level grade matrix are a seat comfort level evaluation data set.
The comfort level evaluation prediction model is modeled by designing a neural network and training the neural network. Dividing the obtained seat comfort evaluation principal component index and the matrix corresponding to the comfort level into an input signal and an output signal, wherein the output signal (teacher signal) is the comfort level given according to the clustering analysis result. And obtaining a seat comfort level evaluation prediction model by training a neural network. And an experimental seat is replaced, comfort level rating data is extracted by measuring surface body pressure distribution signals, the trained neural network model is used for predicting the comfort level ranking of the experimental seat, the prediction result is combined with the body pressure distribution characteristic value, and the reliability of the result is analyzed.
After the human body model device and the comfort level evaluation model are completed, the human body model is used for simulating different percentile true persons and measuring the body pressure distribution signals of the surface of the seat of any seat, the characteristic values are extracted to construct the input vector of the neural network model and input into the neural network model for simulation, and the use comfort level of any seat is predicted.
Compared with the existing seat comfort degree prediction method and the designed measuring device, the invention has the following advantages:
firstly, the shape and structure of the contact part between the human body model device and the surface of the seat are similar to those of a real person, so that errors caused by large difference between the device and the real person are reduced. And the human model device can simulate different percentile human bodies by adjusting the sizes of the ischial tuberosities, the distances of the ischial tuberosities, the thicknesses and the densities of the muscle and skin simulation materials and the weights of the thigh areas and the backrest areas, and is convenient to adjust.
Secondly, the processing of the body pressure distribution indexes on the surface of the seat is realized on the basis of the statistical principle, the classification of different comfort levels is completed through system clustering analysis, subjective classification is not relied on, and the classification result is objective.
Thirdly, the comfort level of the seat is obtained through neural network training, the seat comfort level corresponding to the characteristic value of the body pressure distribution signal measurement of the human body model is estimated through the seat comfort level estimation prediction model. The method is simple to operate, saves cost, and can be used in different stages of product research and development, finished product evaluation and the like.
Drawings
FIG. 1 is a schematic view of the ischial tuberosity structure;
figure 2 is a schematic view of the ischial tuberosity apparatus;
FIG. 3 is a schematic view of the hip-leg device configuration;
fig. 4 is a schematic view of the structure of the manikin device.
FIG. 5 is a schematic view of seating body pressure distribution signal seating surface partition
FIG. 6 is a system clustering tree diagram
FIG. 7 is a graph of neural network training results
FIG. 8(a) is a graph showing a body pressure distribution of the seat 1 in a checking manner
FIG. 8(b) is a diagram showing the body pressure distribution of the backrest of the verification seat 1
FIG. 9(a) is a graph showing a body pressure distribution of a seat 2 on which a body pressure is verified
FIG. 9(b) is a diagram showing verification of the body pressure of the backrest of the seat 2
FIG. 10(a) is a graph showing a body pressure distribution of the seat 3 on which the body pressure of the seat is verified
FIG. 10(b) is a diagram showing the body pressure distribution of the backrest of the verification seat 3
Detailed Description
The invention is further described below with reference to the accompanying drawings.
The human body model is implemented as follows:
as shown in fig. 1, is a single ischial tuberosity a. The maximum pressure of the normal human body and the chair surface is concentrated at two ischial tuberosities of the human body, and the total pressure is about 25cm2. Different ischial contact areas are achieved by setting different sizes of the ischial tuberosities a of the bottom contact surface. The length and the width of the single ischial tuberosity A are respectively provided with three groups of 3 x 3cm, 3 x 4cm and 3 x 5cm, so that the contact area of the two ischial tuberosities is respectively 18cm2，24cm2，30cm2Three groups in total. The distance between the normal two human ischial tuberosities is 8.5-9.5 cm. The distance between the centers of the bottom surfaces of the two ischial tuberosities is 9cm, and the tuberosities are properly changed by properly rotating the angle of the ischial tuberosities A (0-45 degrees). When the weight is applied consistently, the overall contact area and the maximum pressure of the seat surface are unchanged by changing the sizes of the ischial tuberosities, thereby changing the average pressure value of the seat surface. Therefore, the influence research of different human body bone sizes on the body pressure distribution of the sitting body is realized.
As shown in fig. 2, the ischial tuberosity arrangement 1 comprises two ischial tuberosities a, a square plank B and four tenons C. The square wood board B is used as a main body, and hole positions are arranged on the wood board, so that the replacement of the ischial tuberosities A with different sizes is realized. Wherein the ischial tuberosities A can be sleeved in the square wood board B and fixed through hole positions; the square wood board B can be sleeved into the hip-leg device 2 and fixed through the tenon C. As shown in fig. 3, the hip-leg device 2 comprises a hip-leg inner core D, a filling layer E, a wrapping layer F and a lower leg support G, which are independent of each other. The weight adjusting position is that five grooves are arranged on the inner core D of the hip and the leg, and the effect of human body weight is achieved by arranging different weights;
as shown in fig. 4, the hip-leg device is connected to the backrest trim by a hinge H and corresponding screws. The hip-leg device and the backrest device can be mutually detached, and pressure distribution data of a local seat surface or the back can be respectively and independently acquired. And can also simultaneously collect the pressure distribution of the seat surface and the backrest when the human body is in a backward leaning posture.
The included angle between the seat surface and the backrest is adjusted by fixing the hip legs on the seat surface and correspondingly adjusting the angle of the backrest device according to the angle of the backrest, so that the upright and the backward leaning can be realized.
The comfort level assessment method and the model based on the body pressure distribution characteristics and the anthropometric characteristics are realized by the following three steps.
The first step is as follows: and (3) acquiring body pressure distribution signals of different individuals on the surfaces of different types of seating units and extracting characteristic values.
Surface pressure distribution signals of the seat surface and the backrest are acquired by a Tactilus 4.0(Sensor Products LLC in Madison, NJ) body pressure distribution measuring device. The 26 office chairs with different types and different hardness for the experiment comprise common office chairs for staff, office chairs for office, conference chairs, electronic contest chairs and the like, wherein the seat surface and the backrest comprise net surfaces, leather surfaces, cloth surfaces and the like, and the types of the office chairs basically cover the types of the office chairs used in office places.
16 tested subjects (8 male subjects) were selected, the average height was 166.53 + -7.61 Cm, the average weight was 64 + -11.663 Kg, the maximum weight was 87Kg, and the maximum height was 182Cm, which basically covered 90% of the anthropometric sizes of Chinese adults in GB/T10000-1988 "Standard for human sizes of Chinese adults".
The objective evaluation index used in the study is the related index of the body pressure distribution on the surface of the seat, and the related parameters of the human body, such as sex, height and weight, are considered. Wherein, the distribution index of the surface pressure of the seat respectively considers two parts of the seat surface and the backrest. Since the front edge of the seat surface affects the sitting feeling and the leg is pressed if the front edge is protruded, the comfortable feeling is lowered, and thus the contact part of the seat surface with the human body is divided into a thigh region and a hip region as shown in fig. 5.
The surface pressure distribution evaluation index is shown in table 1.
TABLE 1
The 16 subjects were combined with 26 office chairs in an arrangement of 416 groups of experiments, and each index in table 1 was 416 groups of experimental data. Each chair was used as a subject, the comfort index is listed in Table 1, the index values were the mean of the experimental data of 16 subjects, and 26 chairs were clustered systematically by Ward method and squared Euclidean distance. And calculating the mean value of all indexes of the seats in each cluster, and comprehensively analyzing the pressure distribution characteristics of the seat surface and the backrest to assign values to the comfort level grades of the office chairs represented by each cluster.
Five comfort levels are defined in advance in the research, and the comfort levels are 1-5 from low to high, so that 5 classes of the seat are finally obtained through clustering. After the experiment, the body pressure distribution information of 16 persons is collected for each chair, and the surface pressure distribution condition of 26 experimental chairs under the condition of the same batch of testees can be obtained by averaging the pressure distribution indexes of each chair.
Office chairs are systematically clustered according to individual cases, and a dynamic clustering chart is shown in fig. 6. The left side is the office chair sample number. Selecting different clustering numbers corresponding to different clustering results, and selecting 5 clustering classification numbers for this time by referring to the semantics of a commonly used five-level scale in perceptual engineering in order to achieve the purpose of multi-level comfort level classification. The office chair number corresponding to each cluster and the average value of the office chair pressure distribution indexes of all clusters are shown in table 2,
table 2.
As can be seen from Table 2, each pressure distribution index of the office seat surface of cluster 2 is the maximum value in 5 clusters, and the contact area of the seat surface is the minimum value (< 1000 cm) in 5 clusters2)。
The surface of the office chair like the cluster 2 is mainly a training chair and a conference chair, is made of a net surface or hard materials, has a hard seat surface and poor fitting degree, shows maximum pressure and average pressure in body pressure distribution indexes, and has a small contact area of the seat surface.
All the pressure distribution indexes of the cluster-like 3 office chair are the minimum values in 5 clusters, and the contact area of the seat surface is the maximum value (1193 cm) in 5 clusters2)。
The office chair like the cluster 2 is mainly an electric competition chair and a soft-packaged high-back chair, has a softer surface, good pressure dispersing effect and good attaching degree with the buttocks and the thighs.
Pressure index value difference between class cluster 1 and class cluster 4 office chair seat surfacesSmaller, but greater variation in back pressure distribution index values, mainly manifested as greater contact area variation (85 cm)2) And the seating feeling of the office chair like the cluster 1 is considered to be better than that of the office chair like the cluster 4 comprehensively. The cluster 5 office chair seat hip area maximum pressure (23.23Kpa) and thigh area maximum pressure (10.18Kpa) are lower than the cluster 2 office chair identity index values (31.12Kpa, 13.10Kpa) and higher than the other cluster office chairs identity index values.
The maximum pressure of the seat surface is in a negative correlation with the use comfort level of the seat, so that the various office chairs in the cluster are sorted from low to high according to the use comfort level by using the pressure distribution index of the seat surface as follows: the cluster 2 is more than the cluster 5 is more than the cluster 4 is more than the cluster 1 is more than the cluster 3.
In each pressure distribution index of the backrest, the difference between the maximum pressure and the average pressure is small, and the difference between the contact areas is large.
The contact area can reflect the degree of fitting, and in the case of a short sitting posture, the influence of the contact area on the tactile sensation should be most important. Sorting various office chairs in clusters from low to high according to the use comfort degree by using the backrest pressure distribution indexes as follows: the cluster 5 is more than the cluster 4 is more than the cluster 2 is more than the cluster 3 is more than the cluster 1.
The office chair seat face bears the vast majority of weight of the human body, the sitting feeling is mainly influenced by the touch feeling of the seat face, and when the touch feeling of the backrest is similar, the touch feeling of the seat face has decisive influence on the whole use comfort level.
And (3) integrating the surface pressure distribution index values of all clusters in the table 2, and finally sequencing the use comfort levels of all clusters from low to high as: the cluster 2 is more than the cluster 5 is more than the cluster 4 is more than the cluster 1 is more than the cluster 3.
The comfort level is assigned to each class of office chair, and the assignment results and meanings are shown in table 3.
Table 3.
| Clustering results | Comfort level awarding | Means of | Subjective evaluation |
| Cluster | |||
| 3 | 5 | (Comfort) | 75.34 |
| |
4 | Is more comfortable | 73.61 |
| |
3 | In general | 71.70 |
| |
2 | Is less comfortable | 69.57. |
| |
1 | Discomfort | 63.96 |
The second step is that: and constructing a matrix containing anthropometric indexes, body pressure distribution characteristic values of the surface of the corresponding seat and corresponding comfort level.
And according to the assignment results in the table 3, assigning the same comfort level to all office chairs in each cluster respectively, and establishing the mapping relation between each pressure index and the comfort level of the office chair.
Because complex nonlinear relations exist between each pressure index of the office chair and the defined comfort level, the method is suitable for establishing a relation model between the pressure index and the comfort level by using a neural network with strong generalization capability.
The comfort level is a teacher signal of the neural network, each pressure index is an input signal of the neural network, and the pressure indexes of various office chairs in clusters and the corresponding comfort levels are training data sets of the neural network.
Firstly, correlation analysis is carried out on the pressure indexes. The matrix of correlation coefficients for each pressure index is shown in table 4, and the pressure indexes are correlated with each other, and are suitable for principal component analysis.
Table 4.
And extracting the principal components according to the principle that the corresponding characteristic value of the principal component is greater than 1 and the accumulated contribution rate of the principal component is greater than 85%. Principal component characteristic root and cumulative contribution rate results are shown in table 5.
Table 5.
The component load versus component score coefficients are shown in tables 6 and 7.
Table 6.
Table 7.
As can be seen from table 5, the original 9 indexes are represented by 3 new indexes after the extraction of the principal component, and include information of 85% or more of the original indexes.
As can be seen from the analysis of table 6, the hip area maximum pressure and average pressure, the seat surface average pressure, and the seat surface contact area have a higher load in the first principal component, the backrest maximum pressure and backrest contact area have a higher load in the second principal component, and the thigh area maximum pressure and average pressure have a higher load in the third principal component.
The 3 new indexes are extracted to replace the original 8 indexes, and the analytical expressions of the new indexes are obtained according to the component scoring coefficients in the table 7.
F1＝0.174X1+0.131X2+0.198X3+0.153X4+0.199X5+-0.19X6+0.086X7+0.142X8+0.032X9。
F2＝-0.023X1-0.223X2-0.072X3-0.169X4-0.101X5+0.008X6+0.454X7+0.307X8+0.391X9。
F3＝-0.332X1+0.462X2-0.124X30.484X4-0.051X5+0.295X6+0.042X7-0.007X8+0.368X9。
In the analytical expression, X1-X8 are 9 indexes corresponding to the normalized original.
The third step: and (4) modeling a neural network.
The model selects a BP neural network with better generalization capability and is designed into a three-layer structure, namely an input layer, a hidden layer and an output layer. The number of input elements is 3, and the number of output elements is 1-5, wherein the output elements are in comfort level. The comfort level is determined by 5 categories obtained by cluster analysis. The training effect is best when the number of the neurons in the hidden layer is found to be 13 through multiple times of training, and the credibility R of the training result, the prediction result and the test result is more than 90 percent. So that the hidden layer of the neural network is finally determined to be 13 neurons.
The training results are shown in fig. 7. Alternatively, 3 experimental chairs and 1 male subject (height 175cm, weight 60kg) were collected, and the body pressure distribution information of the seat surface and the backrest thereof was collected, and each surface pressure distribution index was extracted.
And (4) predicting according to the surface pressure distribution index and the human body parameter by using the trained PCA-BP neural network model to finally obtain the comfort evaluation predicted value of the three seats.
The maximum pressure and average pressure of the seat and back are measured as shown in table 8, and the corresponding cloud charts of the body pressure distribution of the seat and back are shown in fig. 8-10.
Table 8.
From the final prediction result, the highest comfort prediction score of the three test seats is the No. 3 seat, the comfort level is close to 5, the difference between the No. 2 seat and the No. 1 seat prediction score is small, the difference between the comfort level close to 2 and the No. 3 seat prediction score is relatively large.
From the cloud figures of the seat surface and backrest body pressure distribution of the seat and the design analysis of the seat, the front edges of the seat surfaces of the No. 1 and No. 2 chairs, namely the parts contacting with thighs, are darker, which indicates that the pressure at the front edges is larger than that at other parts of the thighs, and the design is related to the seat surface shape design of the two seats.
The front edges of the seat surfaces of the No. 1 chair and the No. 2 chair are slightly raised, so that the contact parts of the legs are stressed, the front edges of the seat surfaces of the No. 3 chair are designed to be smooth, the legs are not contacted with obvious stress, and therefore the body pressure distribution of the seat surfaces in the leg areas is more uniform, and the color is uniform in the seat surface pressure distribution cloud chart.
From analysis of maximum hip and thigh pressures, the maximum hip area, thigh area and seat surface average pressure of the No. 3 chair are the minimum values of the 3 chairs, and the difference between the three indexes of the seat surfaces of the No. 2 chair and the No. 1 chair is not large, so that the seat surface sensing of the No. 3 chair is the best, and the No. 2 chair is close to the No. 1 chair. From the backrest body pressure distribution cloud picture, the lumbar part of the No. 3 chair is supported, the maximum pressure is the maximum of the three parts, so the lumbar support effect is the best, the maximum pressures of the No. 1 chair backrest and the No. 2 chair backrest are both positioned at the scapula part, and the cloud picture of the lumbar part is light in color, so compared with the No. 3 chair, the No. 1 chair and the No. 2 chair backrest have larger difference in use comfort, and the difference between the No. 1 chair and the No. 2 chair is not obvious. The objective evaluation indexes of the seat surface and the backrest consider that a neural network model has random errors, the sitting feeling difference between the No. 1 chair and the No. 2 chair is possibly not obvious in actual use, and the seat comfort level classification can be carried out in the same classification.
For the selected testee, the overall use comfort of the chair No. 3 is relatively highest, the chair No. 2 and the chair No. 1 are relatively low, and the objective evaluation index comprehensive judgment result is relatively consistent with the prediction result of the PCA-BP neural network model.
The model can accurately predict the comfort levels of the seats used by different users through objective evaluation indexes without depending on a subjective comfort level evaluation method.
The significance in practical application is that the comfort degree prediction of different seats can be accurately and quickly realized; for the research and development of the seat, the use comfort of different designs can be predicted in advance in the design stage, so that the optimal design scheme is obtained, the research and development efficiency is improved, and the cost is reduced.
Claims (9)
1. The invention relates to a device and a method for measuring comfort level of a seat, which mainly comprise a human ischial tuberosity model device (1), a human hip and leg model device (2), a human backrest model device (3), a human skin tissue simulation material (4), a seat surface pressure extraction method (5), a seat surface pressure classification number and classification method (6) and a comfort level evaluation algorithm model (7). The method is characterized in that pressure distribution signals of different human percentile testees on the surfaces of different types of seats are collected, surface pressure signal indexes of the seats are extracted, and the seats are clustered systematically according to the indexes to obtain classification clusters of different seats. And assigning the comfort level represented by each cluster according to the characteristics of the pressure distribution indexes of the surfaces of the seats of each cluster. And extracting a new comfort evaluation index through principal component analysis, and constructing a training data set of the neural network by using the new index and the corresponding comfort value to train the neural network model. The successfully trained neural network model is used for predicting the use comfort level evaluation of different users on different seats. The human ischial tuberosity model device is assembled with a human hip model device, a human leg model device and a human backrest model device, adult human bodies with different percentiles are simulated, real people are replaced to collect body pressure distribution signals on the surface of the seat and extract characteristic values, and the characteristic values of the body pressure distribution signals of the human body model device on the surface of the seat and anthropometric data are input into a trained neural network evaluation model, so that the use comfort evaluation grades of different percentiles of users on different seats are obtained.
2. A seat comfort level measuring device and assessment method according to claim 1, characterized in that the human ischial tuberosity model device (1) is a frustum of a pyramid geometry made of two hard incompressible materials. The distance between the two ischial tuberosities is adjustable to simulate different percentile adult human ischial tuberosities.
3. A measuring device and an assessment method for the comfort level of a sitting means according to claim 1, characterized in that the geometric volume of the model device (1) for ischial tuberosity of a human body is in various specifications to simulate different percentile human skeleton data. And in cooperation with the features of claim 2, achieve the purpose of simulating data of various percentile human ischial tuberosities.
4. The device for measuring and assessing the comfort level of a sitting device as claimed in claim 1, wherein the thigh bones and pelvis bones of the human hip and leg model device (2) are made of incompressible hard materials, the surface of the device is coated with human skin tissue simulation materials (4) to simulate the thighs and the hips of the human body, and different weights are used for simulating the human bodies with different weights in the thigh area and the hip area of the device.
5. The device for measuring and assessing the comfort level of a sitting means according to claim 1, wherein the back model device (3) is made of an incompressible hard material, the back curve is made according to the back of the human body, and the side surface is in the shape of an "S"; wherein the thoracic vertebra part is backward convex, the lumbar vertebra part is forward convex, and the shape of the thoracic vertebra part comprises the shape of a scapula. The device is used for simulating the back form of a human body.
6. The measuring device and the evaluating method for the comfort level of the seat according to claim 1, wherein the human skin tissue simulation material (4) is divided into two parts of muscle and skin and is made of soft and hard high polymer materials; and the human body back rest model device is assembled with the human body hip and leg model device (2) and the human body back rest model device (3) for use, and the human body thigh, the human body hip, the back muscle and the skin are simulated.
7. The seat comfort level measuring device and the seat comfort level evaluating method according to claim 1, wherein the seat surface pressure extracting method (5) is to extract the maximum value, the average value and the contact area of body pressure distribution from the seat surface area and the backrest area of the seat surface; the seat surface area is divided into a hip area and a thigh area, and the maximum pressure and the average pressure of the two areas are respectively extracted.
8. The device and the method for measuring and evaluating the comfort level of a seat according to claim 1, wherein the classification number and the classification method (6) of the surface pressure of the seat are implemented by clustering seat individual patterns according to the distribution index of the seat pressure through a Euclidean squared distance algorithm, and determining the number of classification clusters and the seats included in each classification cluster. And extracting the mean values of body pressure distribution indexes in different areas in each classification cluster, and sequencing and assigning the seats according to the comfort level of various cluster body pressure distribution indexes.
9. The device and the method for measuring the comfort level of the seat according to claim 1, wherein the comfort level evaluation algorithm model (7) uses a neural network model with teacher signals, extracts the principal components of the original pressure distribution indexes by a principal component analysis method, constructs new indexes to be used as input signals of the neural network, and outputs signals of the neural network as the evaluation level of the comfort level of the seat. The teacher signal is the comfort level assigned by claim 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110345532.9A CN113128031A (en) | 2021-03-27 | 2021-03-27 | Measuring device and evaluation method for comfort level of seat |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110345532.9A CN113128031A (en) | 2021-03-27 | 2021-03-27 | Measuring device and evaluation method for comfort level of seat |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113128031A true CN113128031A (en) | 2021-07-16 |
Family
ID=76774941
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110345532.9A Pending CN113128031A (en) | 2021-03-27 | 2021-03-27 | Measuring device and evaluation method for comfort level of seat |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113128031A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113936335A (en) * | 2021-10-11 | 2022-01-14 | 苏州爱果乐智能家居有限公司 | Intelligent sitting posture reminding method and device |
| CN114878122A (en) * | 2022-07-12 | 2022-08-09 | 中国飞机强度研究所 | Vibration evaluation method for aircraft cabin environment test |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003111646A (en) * | 2001-10-03 | 2003-04-15 | Ergo Seating Kk | System for measuring and analyzing characteristic of body pressure distribution and system for predicting and evaluating change with time of body pressure distribution |
| CN102110243A (en) * | 2009-12-23 | 2011-06-29 | 新奥特（北京）视频技术有限公司 | Method for predicting human comfort |
| CN110443318A (en) * | 2019-08-09 | 2019-11-12 | 武汉烽火普天信息技术有限公司 | A kind of deep neural network method based on principal component analysis and clustering |
| CN110441048A (en) * | 2019-09-16 | 2019-11-12 | 一汽-大众汽车有限公司 | A kind of seat comfort evaluation method based on the distribution of objective pressure |
| CN110889440A (en) * | 2019-11-15 | 2020-03-17 | 山东大学 | Rockburst grade prediction method and system based on principal component analysis and BP neural network |
| CN112465347A (en) * | 2020-11-26 | 2021-03-09 | 湖南科技大学 | Method for cooperatively predicting roof stability based on cluster analysis and improved neural network |
-
2021
- 2021-03-27 CN CN202110345532.9A patent/CN113128031A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003111646A (en) * | 2001-10-03 | 2003-04-15 | Ergo Seating Kk | System for measuring and analyzing characteristic of body pressure distribution and system for predicting and evaluating change with time of body pressure distribution |
| CN102110243A (en) * | 2009-12-23 | 2011-06-29 | 新奥特（北京）视频技术有限公司 | Method for predicting human comfort |
| CN110443318A (en) * | 2019-08-09 | 2019-11-12 | 武汉烽火普天信息技术有限公司 | A kind of deep neural network method based on principal component analysis and clustering |
| CN110441048A (en) * | 2019-09-16 | 2019-11-12 | 一汽-大众汽车有限公司 | A kind of seat comfort evaluation method based on the distribution of objective pressure |
| CN110889440A (en) * | 2019-11-15 | 2020-03-17 | 山东大学 | Rockburst grade prediction method and system based on principal component analysis and BP neural network |
| CN112465347A (en) * | 2020-11-26 | 2021-03-09 | 湖南科技大学 | Method for cooperatively predicting roof stability based on cluster analysis and improved neural network |
Non-Patent Citations (1)
| Title |
|---|
| 贾明兴等: "基于ＲＢＦ神经网络的非线性主元分析方法", 《仪器仪表学报》, vol. 29, no. 3, 31 March 2008 (2008-03-31), pages 453 - 457 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113936335A (en) * | 2021-10-11 | 2022-01-14 | 苏州爱果乐智能家居有限公司 | Intelligent sitting posture reminding method and device |
| CN114878122A (en) * | 2022-07-12 | 2022-08-09 | 中国飞机强度研究所 | Vibration evaluation method for aircraft cabin environment test |
| CN114878122B (en) * | 2022-07-12 | 2022-09-13 | 中国飞机强度研究所 | Vibration evaluation method for aircraft cabin environment test |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fazlollahtabar | A subjective framework for seat comfort based on a heuristic multi criteria decision making technique and anthropometry | |
| CN113128031A (en) | Measuring device and evaluation method for comfort level of seat | |
| Naddeo et al. | The effect of posture, pressure and load distribution on (dis) comfort perceived by students seated on school chairs | |
| JP2003111646A (en) | System for measuring and analyzing characteristic of body pressure distribution and system for predicting and evaluating change with time of body pressure distribution | |
| CN106092635A (en) | A kind of mattress pressure comfort evaluation methodology | |
| CN109760691B (en) | It is a kind of consider electromyography signal driver comfort drive posture under the matched method of man-machine arrangement parameter | |
| Franz et al. | A light weight car seat shaped by human body contour | |
| US20070143927A1 (en) | Cushion and method of evaluating the cushion | |
| Vermander et al. | Intelligent sitting posture classifier for wheelchair users | |
| Sousa et al. | University classrooms and students' compatibility | |
| Andrade | An ergonomic evaluation of aircraft pilot seats | |
| Lyu et al. | Influence of body mass index on comfort and parametric optimization design of seat | |
| Silva et al. | Classification algorithms for body posture | |
| Dénes et al. | Comparative study of body pressure distribution on the user-cushion interfaces with various support elasticities | |
| Wang et al. | A method of aircraft seat dimension design for long-term use by passengers with different body types | |
| Carter et al. | Physique: phenotype, somatotype and 3D scanning | |
| Adler | The relation between long-term seating comfort and driver movement | |
| Rahayu et al. | Designing A Reading Chair using Kansei Engineering Approach | |
| CN112158353B (en) | Method for researching design of ejection seat backrest surface by using three-dimensional scanner | |
| Savonnet | Development of a customizable digital tool for assessing aircraft passenger discomfort and fatigue | |
| Hu et al. | Research on pressure comfort of sofa based on body pressure distribution and subjective experience | |
| Al-Harkan et al. | Designing a new school furniture suitable for Saudi Students | |
| Lee | Anthropometric Design and Ergonomic Posture Assessment based on Intelligent Algorithms for Seated Work | |
| Liu et al. | PIPER adult comfort: an open-source full body human body model for seating comfort assessment and its validation under static loading conditions | |
| CAPPETTI et al. | Posture prediction of a human on a chair: model validation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |