Prediction of clothing sensory comfort using neural networks and fuzzy logic

Pao Yue-kong Library Electronic Theses Database

Prediction of clothing sensory comfort using neural networks and fuzzy logic


Author: Wong, Siu-wo Anthony
Title: Prediction of clothing sensory comfort using neural networks and fuzzy logic
Degree: Ph.D.
Year: 2003
Subject: Hong Kong Polytechnic University -- Dissertations
Fuzzy systems -- Industrial applications
Clothing and dress -- Physiological aspects
Textile fabrics -- Physiological aspects
Department: Institute of Textiles and Clothing
Pages: xxiii, 258 leaves : ill. ; 30 cm
Language: English
InnoPac Record:
Abstract: The purpose of this study is to investigate the process of human psychological perception of clothing related sensations and comfort to develop intellectual understanding and methodology to predict clothing comfort performance from fabric physical properties. The framework for this study is to establish theoretical understanding of human sensory perception process, from fabric physical properties to individual sensory perceptions, and then to overall clothing comfort perception. The relationships between individual sensory perceptions and overall comfort perception during exercise are studied by wear trials and model development. Ten individual sensory perceptions are abstracted into three independent sensory factors, which are related to moisture comfort, tactile comfort and thermal-fit comfort by using statistical factor analysis. Mathematical models are developed on the basis of assumption of linear relationship between the sensory factors and overall clothing comfort. Further, feed-forward backpropagation neural networks and fuzzy logic models are developed to predict overall clothing comfort from individual sensory perceptions. Good correlations are found between predicted and experimental comfort rating with all three models. The relationships between subjective perceptions and objectively measured clothing pressure at nine different body locations are studied. It is found that objectively measured pressure values are significantly different amongst different garments and postures. On the other hand, garment has a significant influence on subjective pressure comfort ratings. Clothing pressures and pressure comfort perceptions are not uniformly distributed at different body locations. Therefore, skin pressure sensitivity index is derived. Dynamic garment pressure distribution in tight-fit sportswear is simulated by using computational biomechanical models. Again it is found that clothing pressure distribution is not uniformly distributed and high-pressure zone is concentrated around the waist girth at the end of the wearing process. The predicted pressures are consistent with the experimental pressure measurements. A new series of experiments are conducted to validate the models. Hand evaluation results show that subjects of all ages require a comfortable sportswear being cool, soft and smooth. Three sensory factors associated with thermal-wet comfort, tactile comfort and pressure comfort are also identified, which confirm previous findings. In the investigation of how fabric properties influence subjective judgement on clothing comfort, fabric physical properties associated with fabric hand and clothing pressure comfort are measured. Prediction model for overall comfort during exercise consists of frictional coefficient, compressional energy and maximum wetted radius, which can be described as individual factor that relates to tactile, pressure and moisture respectively. To predict overall clothing comfort from fabric physical properties, various hybrid models are developed by studying human sensory perception and judgement processes using different modelling techniques. By combining the strengths of statistics (data reduction and information summation), neural networks (self-learning capability) and fuzzy logic (fuzzy reasoning capability), hybrid models have been developed using the most suitable technique to simulate different stages of the perception processes. Results show that the TS-TS-NN-FL four stages model has the highest capability to predict the overall comfort performance from fabric physical properties. This indicates that the combination of using statistical method to abstract sensory and fabric physical factors, using neural networks to simulate the sensory perception processes and then using fuzzy logic to predict final judgement of overall clothing comfort is the most suitable hybrid models. To summarise, the three key elements in the prediction of psychological perception of clothing comfort from fabric physical properties are: data reduction and summation, self-learning capability and fuzzy reasoning. This thesis shows that the model, which integrates these three elements, can generate the best predictions comparing with other hybrid models.

Files in this item

Files Size Format
b16866575.pdf 9.984Mb PDF
Copyright Undertaking
As a bona fide Library user, I declare that:
  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.


Quick Search


More Information