The value of smart environments in understanding and monitoring human behavior has become increasingly obvious in the past few years. Using data collected from sensors in these environments, scientists have been able to recognize activities that residents perform and use the information to provide context-aware services and information. However, less attention has been paid to monitoring and analyzing energy usage in smart homes, despite the fact that electricity consumption in homes has grown dramatically. In this chapter, the authors demonstrate how energy consumption relates to human activity through verifying that energy consumption can be predicted based on the activity that is being performed. The authors then automatically identify novelties in human behavior by recognizing outliers in energy consumption generated by the residents in a smart environment. To validate these approaches, they use real energy data collected in their CASAS smart apartment testbed and analyze the results for two different data sets collected in this smart home.
TopBackground
Given the recent progress in computing power, networking, and sensor technology, we are steadily moving into the world of ubiquitous computing where technology recedes into the background of our lives. Using sensor technology combined with the power of data mining and machine learning, many researchers are now working on smart environments which can discover and recognize residents’ activities and respond to resident needs in a context-aware way.
A core technology component in this research is the ability to automatically recognize and identify activities performed by residents in smart environments. A variety of approaches have been used to achieve this goal. For example, Hu et al. (Hu, Pan, Zheng, Liu, & Yang, 2008) find common trends in Activities of Daily Living (ADLs) to see whether the inhabitants perform multiple concurrent and interleaved activities or single activities. Gao et al. (Gao, Hauptmann, Bharucha, & Wactlar, 2004) use hidden Markov models to characterize different stages in dining activities. The smart hospital project (Sánchez, Tentori, & Favela, 2008) develops a robust approach for recognizing user’s activities and estimating hospital-staff activities by employing a hidden Markov model with contextual information in the smart hospital environment. The Georgia Tech Aware Home (Orr & Abowd, 2000) identifies people based on pressure sensors embedded into the smart floor in strategic locations. The CASAS smart home project (Singla, Cook, & Schmitter-Edgecombe, 2010) builds probabilistic models of activities and uses them to recognize activities in complex situations where multiple residents perform activities in parallel in the same environment. A new idea of transfer learning (Rashidi & Cook, 2010) is gaining popularity in smart home research due to its ability to use the knowledge gained from one domain to a different but related domain, making the learning problem more generalized for similar environments, activities, or inhabitants.