Researcher: Anthony Rowe
Scope: This project will study the degree to which user activity information can be inferred from residential smart grid energy metering. Using live data from heavily metered and sensed living spaces we will attempt to estimate occupancy and classify user activities. We will analyze the impact of sensing data, update rate and resolution on classification performance.
Outcomes: Determine how much contextual information about user activities can be inferred from smart grid sensors. Analyze how classification performance changes with respect to sensor fidelity. Evaluate how these approaches generalize by cross-validating on multiple test-beds.