CSR: Small: RF-Wear: Enabling RF Sensing on Wearable Devices for Non-Intrusive Human Activity, Vital Sign and Context Monitoring

Michigan State University, East Lansing MI

Investigators

Abstract

Wearable devices are growing in popularity at a tremendous rate. Today, millions of people wear fitness trackers and smart watches on a daily basis to track steps, heart rates and sleep. However, the constraints of the built-in sensors in existing wearable devices act as key bottlenecks that limit the broader impact of wearable devices. In particular, accelerometers and gyroscopes can only sense motion and rotation of body parts to which they are attached. Microphones and cameras can capture rich information but come with severe privacy concerns. Many physiological sensors such as respiration and electrocardiogram sensors are intrusive in that they must be placed at certain body locations and require tight skin contact to function. This project focuses on pushing the boundaries of wearable devices by exploring wireless radio as a new sensing modality on wearable devices for human physical activity, vital sign, and context sensing. Compared to existing sensing modalities in wearable devices, radio provides a contactless and privacy-preserving approach to capture rich information about users and their context. The underpinning principle is that movements of different parts of human body caused by different human activities (physical, physiological and contextual) generate different changes on radio signals. By analyzing these changes, the activity that causes the changes can be recognized. Based on this principle, this project aims to pursue five tightly coupled research tasks associated with the goal of exploring radio based sensing on wearable devices: (1) investigating radio as a whole-body motion sensor for recognizing activities of daily living (ADLs); (2) investigating radio as a contact-less physiological sensor for monitoring vital signs; (3) investigating radio as a context sensor for identifying the object a user is interacting with; (4) putting everything together to use a single radio sensing modality for recognizing ADLs, monitoring vital signs, and identifying context simultaneously; and (5) developing power-aware schemes to minimize the power consumption of the wearable device. This project will advance the knowledge and understanding of radio based sensing on wearable devices. It has the potential to fundamentally transform wearable devices as they are currently known. If successful, this project will not only make a theoretical contribution to radio based sensing on wearable devices but also yield a prototype platform, opening up a new world of applications in domains such as healthcare, entertainment, social computing and education.