Wearable Textile-Based Contact Sensing for Functional Fit Assessment

Abstract

Fit of a wearable system influences many human factors, including comfort, performance, and risk of injury. Sensors can provide objective and quantitative measures of mechanical interactions between the body and the wearable system for functional fit assessment. However, accurate on-body sensing is a challenge due to contaminating variables that can affect accuracy, including forces introduced by garment/body interactions such as stretching and folding. Contact sensing is a simpler sensing approach that is less susceptible to on-body contaminating variables. However, there is currently no gold-standard reference measure for on-body contact measurement. Here, multiple imperfect sources of information are compared and their respective limitations contrasted. This research focuses on evaluating methods to quantify functional fit of a wearable system by measuring contact between the body and a spacesuit component mockup during controlled robotic manikin testing through a wearable contact and force sensing e-textile garment. This study compared two sensor-based fit quantification methods (contact and force sensors) with a non-wearable reference (optical Motion Capture (MoCap)). Garment-integrated sensors were characterized in a bench test apparatus (Instron) under controlled loading conditions. The translation of these methods to the wearable environment was investigated using a robotic manikin that performs repeatable dynamic movements for a controlled on-body sensing scenario. Two different manikin conditions were evaluated to simulate effects of anthropometric differences. Under controlled conditions, contact sensors showed some hysteresis and generally exhibited higher closing forces compared to opening forces. Using the threshold calibration model, contact sensors accurately measured contacts above about 0.5 N, but recorded intermittent false negative contacts between approximately 0-0.5 N. Force sensors reliably measured contacts above 0.15 N and comparatively recorded a smaller range of false negatives between 0-0.15 N, but a much larger proportion of false positives. However, under on-body conditions, the contact-threshold calibration did not accurately translate for force sensors. There were no strong similarities found between contact sensor, force sensor, and MoCap marker data. Force sensors were difficult to calibrate and sensitive to factors like donning forces, movement, and wrinkling. Contact sensors were influenced by fewer and more resolvable contaminating variables, and were found to be better suited for on-body applications.