Section: New Software and Platforms

THEGAME: data fusion for Smart Home and Smart Building

Keywords: Smart home - Smart building

• Participants: Yoann Maurel and Frédéric Weis

• Partner: Université de Rennes 1

• Contact: Frédéric Weis

• URL: https://github.com/bpietropaoli/THEGAME/ (C-version)

• URL: https://bitbucket.org/TACOMA-irisa/java-bft (Java-version)

Scientific Description

Context-aware applications have to sense the environment in order to adapt themselves and provide with contextual services. This is the case of Smart Homes equipped with sensors and augmented appliances. However, sensors can be numerous, heterogeneous and unreliable. Thus the data fusion is complex and requires a solid theory to handle those problems. The aim of the data fusion, in our case, is to compute small pieces of context we call context attributes. Those context attributes are diverse and could be for example the presence in a room, the number of people in a room or even that someone may be sleeping in a room. For this purpose, we developed an implementation of the belief functions theory (BFT). THE GAME (THeory of Evidence in a lanGuage Adapted for Many Embedded systems) is made of a set of Libraries. It provides the basics of belief functions theory, computations are optimized for an embedded environment (binary representation of sets, conditional compilation and diverse algorithmic optimizations).

THE GAME is published under apache licence. It is maintained and experimented within a sensor network platform developed by Tacoma since June 2013.