Study of knit-based comfortable and scalable sensors for human motion monitoring

Abstract

In the current era, human motion monitoring is a crucial research area with profound implications, playing significant roles in healthcare, sports, and daily life. In healthcare, it facilitates disease diagnosis and rehabilitation progress monitoring by assessing patients' physical activity levels. In sports, it optimizes training regimens through analyzing athletes' movements. In daily life, it oversees individuals' well-being and activity patterns. Given its importance, there has been a remarkable increase in research on flexible sensors, aiming to overcome the limitations of traditional rigid sensors like poor adaptability to body contours and user discomfort. To address these issues, researchers are utilizing innovative materials and designs that can easily adapt to body movements and contours. Textile-based sensors, in particular, hold a prominent position in human motion sensing, mainly due to their comfort-oriented characteristics. However, there are still many challenges in current research. For example, the production process is often complex, there exists a persistent contradiction between achieving high sensitivity and maintaining comfort, and the associated costs remain relatively high. To address these hurdles, the present thesis focuses on leveraging knitting technology for the development of human motion sensors. Firstly, the knitted structure imparts the sensor with remarkable flexibility and softness. Concurrently, it not only facilitates direct customization but also enables seamless integration, thereby significantly enhancing the practicality and comfort levels of the sensor. Additionally, advanced knitting machines are capable of realizing intricate sensor structures by integrating yarns possessing diverse functions in multifarious ways. This, in turn, augments the diversity and flexibility in the design process. Furthermore, the well-established knitting technology plays a pivotal role in promoting the scale expansion of flexible sensors as well as in reducing their production costs. Collectively, these distinctive characteristics firmly establish the importance and necessity of developing knitted-based sensors. In this research, three distinct types of knitted-based sensors have been successfully developed. First, a soft, warm, and mass-producible triboelectric carpet fabric is developed for motion posture monitoring and user recognition. A specially prepared conductive chenille yarn consisting of core-Ag-coated nylon filaments and shell-acrylic staple fibers is used as the main raw material. And knitting weft insertion technology is employed to bundle conductive chenille yarns in 1×1 rib courses formed by nylon elastic yarns to make the chenille TENG carpet fabric. The chenille TENG fabric, which exhibits flexibility, warmth, low cost, ease of manufacturing, and compatibility with the living environment, has demonstrated a maximum power density of approximately 2942 μW/m² in the contact-separation mode. Through simple circuit management for energy harvesting, it is capable of powering small electronic devices. Moreover, with the assistance of machine learning, the chenille TENG can be applied to behavior recognition and user identification. After training the time-current data, four distinct behaviors, namely slow walking, regular walking, jogging, and jumping, have been successfully classified, and four different subjects have been recognized. This carpet fabric thus shows significant potential in smart monitoring systems for home security, owing to its aforementioned properties and capabilities. In the second work, an innovative method for efficiently manufacturing flexible sensing fabrics is proposed, leveraging three-thread fleecy knitting technology. This approach achieves a relatively high production rate of around 11.53 m²/h. The produced fleecy sensing fabric functions as a triboelectric nanogenerator, capable of generating electrical signals and attaining a peak power density of about 2446 μW/m² upon rubbing against cotton fabric. Composed entirely of commercially available yarn materials, the fabric possesses outstanding flexibility, plumpness, and breathability. Remarkably, it maintains stable output performance even after undergoing multiple machine wash cycles. This fabric can be freely cut and tailored into self-powered flexible sensors for various applications, including insoles for movement pattern monitoring and carpets for movement posture tracking. With the reinforcement of machine learning algorithms, the fleecy sensing fabric exhibits strong recognition capabilities. The combination of these features creates opportunities for the development of flexible sensors that are cost-efficient, comfortable, and widely applicable, thus expanding their potential for practical use in a wide range of scenarios. The third work introduces a novel auxetic braided strain yarn sensor (ABSYS) capable of achieving industrialized automatic production. The highly stretchable ABSYS is constructed by wrapping rigid conductive multifilament and highly elastic nylon/spandex-covered yarn around an elastic core yarn in a mesh pattern using a circular braiding machine. It exhibits a pronounced auxetic effect and exceptional sensing performance, including good structural stability, a wide working range, and high sensitivity. Based on the developed ABSYS, a highly stretchable sensing fabric is created by seamlessly integrating it into rib fabric using knitting technology. The fabric demonstrates a broad working range of 2% to 60%, a rapid response time of 0.018 s, and dependable stability, effectively addressing the majority of human motion sensing requirements. This type of sensor enables effective motion monitoring through reliable bending detection at various joints, all while maintaining wearing comfort and aesthetics. This work offers a promising new approach for strain-sensing fibers and fabrics in wearable applications.