Wearable textile strain sensors for measurement of spinal flexion

Wearable sensors have the potential to enable continuous real-time health monitoring of people during normal daily activities. In contrast, the current paradigm requires patients to devote a period of time to attend for tests in specialist facilities and under conditions that, at best, are not representative of their normal life patterns, and at worst, may induce considerable stress, leading to significantly biased data. Physical therapy, training technique, rehabilitation, respiration monitoring and diagnostics could all be improved by implementing wearable sensors and data acquisition software [1]. This will not only improve the accuracy of the measurements, but the ability to analyze the data over time. Herein, we present an alternative to the current clinician measurement of spinal flexion; the modified Schober’s test. The accuracy of the test is determined by each clinician causing a large tendency towards error [2]. By implementing a strain sensor in place of the measuring tape currently used, it is proposed that inter-observer error would be reduced and more consistent measurements would be provided over time. Using a textile sensor for this application would allow for the wearable system to be integrated into a garment. This would be more favorable than the use of traditional rigid electronics encumbering the user. In this work, two textile based sensors were tested for use in this application; a knitted spandex cylindrical structure with integrated carbon nanotubes (CNT) and a flat, knitted piezoresistive fabric (KPF) knitted with Lycra®. The sensors were tested for resistance changes versus strain both in the laboratory with a linear actuator and on-body using the protocol mandated in the Modified Schober’s test. Each sensor was stretched from 0% to 100% in 10% increases with a 5 second hold at each interval. Varying CNT sensor lengths and core sizes were chosen, and after preliminary testing, a 16.5 cm 8 core sensor was chosen to have the best separation between consecutive percent strains. The same method of preliminary testing was followed for the KPF sensors. Several major features emerged as priorities on comparison of the sensors. These factors were health, robustness, and repeatability. Although the CNT sensors have a much greater elasticity than the KPF sensors, the relaxation time at room temperature is much longer than desired for clinical applications. In addition, because of the high recovery force of the KPF sensors, they show strength over time as well as the consistent readings necessary for such precise measurements. Finally, given the health risks associated with CNTs [3] it is a priority to assure that the sensors are safe to use on-body. This would require complete encapsulation of the sensor as well as additional testing to ensure no CNTs emerged over time. Taking these factors into account, the KPF sensors emerged as the optimum choice for this application