Designing sensor sets for capturing energy events in buildings

We study the problem of designing sensor sets for capturing energy events in buildings. In addition to direct energy sensing methods, e.g. electricity and gas, it is often desirable to monitor energy use and occupant activity through other sensors such as temperature and motion. However, practical constraints such as cost and deployment requirements can limit the choice, quantity and quality of sensors that can be distributed within each building, especially for large-scale deployments. In this paper, we present an approach to select a set of sensors for capturing energy events, using a measure of each candidate sensor's ability to predict energy events within a building. We use constrained optimisation -- specifically, a bounded knapsack problem (BKP) -- to choose the best sensors for the set given each sensor's predictive value and specified cost constraints. We present the results from a field study of 4 UK homes with temperature, light, motion, humidity, sound and CO2 sensors, showing how valuable yet expensive sensors are often not chosen in the optimal set.