Using Many Simple Sensors for Automatic Monitoring in the Home

About 1 in 5 Americans have some kind of disability, and 1 in 10 have a severe disability [16]. Instrumented environments can be used for automatic monitoring for people with physical and cognitive disabilities. We describe a system that uses a handful of motion detectors to predict light switch usage in an instrumented home. We use a naive Bayes learner with an automatically extracted feature set. We demonstrate results over two months of data collected from an instrumented house occupied by students. We explain why our methods are independent of this environment and describe specific applications to home health care.