Energy Prediction Based on Resident's Activity

In smart home environment research, little attention has been given to monitoring, analyzing, and predicting energy usage, despite the fact that electricity consumption in homes has grown dramatically in the last few decades. We envision that a potential application of this smart environment technology is predicting the energy would be used to support specific daily activities. The purpose of this paper is thus to validate our hypothesis that energy usage can be predicted based on sensor data that can be collected and generated by the residents in a smart home environment, including recognized activities, resident movement in the space, and frequency of classes of sensor. In this paper, we extract useful features from sensor data collected in a smart home environment and utilize several machine learning algorithms to predict energy usage given these features. To validate these algorithms, we use real sensor data collected in our CASAS smart apartment testbed. We also compare the performance between different learning algorithms and analyze the prediction results for two different experiments performed in the smart home.