Yang Cai (Carnegie Mellon University, USA)
Abstract: Analyzing human motion in a building has been an active subject in Ambient Intelligence and Universal Design for many years. In this study, we present a rapid prototype of a mobile and interactive sensing platform for smart buildings. The biologically inspired robot can follow the moving person around, memorize the motion patterns in form of sequences of symbols, and detect surprising events, based on similarity between the priori and posteriori probability distributions. The key modules in this study have been prototyped and tested with real-world data, such as the twomonth sensory data in a building. Furthermore, the author believes that simple and recursive algorithms would enable mobile robots to simulate natural ethological intelligence.
Keywords: anomaly detection, instinctive computing, motion, sensor network, smart environment
Categories: H.1, J.0, L.7