Wearable Analytics Lab
The work in this lab is characterized by development of algorithms to recognize activities, emotions, and physiological parameters by body-worn sensors.
Analytics of the data obtained from wearable devices is an emerging field of research, which includes ubiquitous computing, context-aware computing and multimedia. Rapid popularization of smartphones and wearable devices with powerful sensors, continuous improvements in their form factors, and improving battery lifetime also fuel interest in the field. The focus of our wearable analytics lab is on understanding the motion patterns of subjects wearing the sensors (IMUs in most cases). We are solving problems related to Activity Recognition, Gesture/Action Detection, quantifying fatigue and level of physical activity, estimating speed, trajectory, and other motion characteristics of the subject wearing the sensor. Most of the work focuses on human subjects, however similar techniques are also applied to data collected from wearable sensors on animals.
From our lab team member:
“As a PhD in physics I have always been excited about modeling and capturing the physical aspects of our world. From motions, forces, speed action, and activity recognition: Machine Learning for Wearable Analytics is opening up new ways of understanding our world through physical modelling. At AGT I can apply my knowledge in my role as a Data Scientist in various areas from sports (athlete modelling) to consumer areas.”