Designing smart environments: A paradigm based on learning and prediction

Abstract

We propose a learning and prediction based paradigm for designing smart home environments. The foundation of this paradigm lies in information theory as it manages uncertainties of the inhabitants' contexts (e.g., locations or activities) in daily lives. The idea is to build compressed dictionaries of context-aware data collected from sensors and devices monitoring and/or controlling the smart environment, efficiently learn from these profiles, and finally predict inhabitant's future contexts. Successful prediction helps automate device control operations and tasks within the environment as well as to identify anomalies. Thus, the learning and prediction based paradigm optimizes such goal functions of smart environments as minimizing maintenance cost, manual interactions and energy utilization. After identifying important features of smart environments, we present an overview of our MavHome architecture and apply the proposed paradigm to the inhabitant's location and activity tracking and prediction, and automated decision-making capability. © Springer-Verlag Berlin Heidelberg 2005.