The Study Of Comfort And Accuracy In Smart Clothing Design For Babies

Smart clothing design is a new research field on the edge of fashion design, computerscience, material science and etc. In the current research, sensors are usually tightly attachedto the skin so as to get an accurate signal, but it tends to degrade the comfort and safetyquality especially for baby clothes. This study focuses on skin-loose sensors where there isspace between clothes and skin. This design eliminates the discomfort caused by tighteningthe skin but tends to degrade signal accuracy of sensors. Therefore, the trade-off betweenaccuracy and comfort becomes one of the biggest challenges in smart clothing design.Addressing this issue, we propose a novel approach to balance the accuracy and comfortquality, which utilizes the prevailing techniques in artificial intelligence and machine learningto combine signals from multiple skin-loose sensors in different locations to improve theaccuracy without sacrificing the comfort quality. Combining the methodology of babyclothing design, principle of comfort quality in fashion design, regression and classificationmodel in machine learning, we designed an experiment for body temperature monitoring withsmart clothing. We found the accuracy performance was substantially improved viacombination of features from multiple sensors. Various comparative experiments wereperformed to justify the effectiveness of the proposed method.Different from popular techniques to improve the comfort quality, e.g., textile sensors,our approach focuses on using the technique of artificial intelligence to get a better trade-offbetween accuracy and comfort without changing material of hardware, which is the advantageover textile sensors.In Section1, we introduce the background knowledge of smart clothing. In Section2wediscuss the trade-off between accuracy and comfort quality in smart clothing design. InSection3, the methodology was described in detail. Section4presents the experimentalsetting and result analysis. In Section5, we summarize the conclusion and future work.