Human activity tracking using wearable sensors in loose-fitting clothes: survey

Abstract

Human activity tracking using wearable technology is an important research field of study and is employed in many applications such as healthcare, aged care, fitness, and tracking activity. Although this field has been well addressed by researchers in the last decade that most of the research is done with tightly fixed sensors (either to the body or to the tight garment), while in contrast, the majority of our daily clothes are loosely fitting so we must be interested in this fields but the main challenge in tracking/recognizing activities in loose clothing is the higher level of noise generated by the movement of clothing compared fixed sensor. This paper reviews recent research on human activity tracking techniques that use sensors that are attached or embedded in loose-fitted clothes. It is predicted from recent studies that many low-cost, reliable, and socially acceptable wearable devices will be available to aid activity tracking. Survey also predicts an increase in the interest for wearables in loose-fitting clothes due to their wide real-time applications. Most of the reviewed work based heavily on off-shelf microcontroller board and sensor, which are characterized by their moderate performance. Hence most of the reviewed work used machine learning to enhance results classification.