The ITR project "Human Activity Monitoring Using Simple Sensors" explores the use of binary sensors typically found in home security systems to monitor human behavior. We envision systems that might remind a user that the stove was left on, detect situations that might lead to a fall or injury, or help a caregiver diagnose a treatable mental condition. Science advances include new machine learning techniques to extract more information from simple sensors. The project focuses on Bayesian or probabilistic methods to analyse the sensor data. One implication of this work is that it may help the elderly live in their own homes longer than they might otherwise. The significance of this work is that it focuses on developing the potential of inexpensive sensing and minimizing the invasion of privacy of behavior monitoring systems.
