Design and Modeling of a Textile Pressure Sensor for Sitting Posture Classification

Abstract

A textile pressure sensor has been designed for measuring pressure distribution on the human body. Electrodes built with conductive textiles are arranged on both sides of a compressible spacer, forming a variable capacitor. The use of textiles makes unobtrusive, comfortable, lightweight, and washable sensors possible. This simplifies the goal to integrate such sensors into clothing in the future, to be simple, and fast to mount just as getting dressed. Hysteresis induced by the spacer of the sensor has been modeled with the Preisach model to reduce the measurement error from 24% to 5% on average and the maximal error from above 50% to below 10%. Standard textiles that are not specially optimized for low hysteresis can be used for designing the sensor due to the modeling. The model may also be used for other pressure or even strain sensors to reduce their hysteresis. The modeling enhances the accuracy of the textile sensor to those of commercial, nontextile pressure sensing mats. Furthermore, we used the sensor system for the classification of sitting postures in a chair. The data of nine subjects have been classified with Naive Bayes classifier, achieving an average recognition rate of 82%. We show that the textile sensor performs similarly to a commercial, nontextile pressure sensing mat for this application.