Researcher: Pei Zhang
Research Area: Mobility
Poor posture, incorrect muscle usage and excessive muscle fatigue are leading causes of many injuries and inefficiencies in sports and fitness. As a result, real-time feedback regarding human motion, posture and muscular activity could be used by individual athletes both professional and amateur, to mitigate these injuries and refine athletic performance so as to gain a competitive advantage. For this reason, sensing and monitoring muscles, as well as human motion, is important. Many single-point on-body sensors and camera systems have been proposed but none have successfully gathered enough data to gain a holistic understanding of musculoskeletal activity. These approaches often are limited to laboratory environments and sometimes require intrusive sensors that are difficult to deploy. Toward this end, we developed the Physiological Activity Recognition System (PARS). The PARS system utilizes a network of small inertial sensors distributed inside clothing to measure vibration that are caused by physiological activity such as muscle activity or cardiovascular functions. We then use these vibrations to deduce body motion, individual muscle fatigue and other physiological activity by identifying features through a combination of machine learning and activity identification algorithms. The full spectrum of sensed data are used. We propose to use the high frequency component, with machine learning techniques, to yield muscular exertion and state and the lower components for motion detection and other cardiovascular features. The system uses the sensed data and an animated human body model to display the motion of the subject. Initial experiments in this area show immense promise of this approach.
Outcomes: Papers, wearable prototype suit