A framework for distributed intelligence for energy efficient operation of smart homes

Abstract:
The world-wide energy demand and with it the greenhouse gas emissions are rising constantly. One of the main energy consumers are buildings, both in the commercial and residential area. Regardless if viewed from an international, European or Austrian vantage point, the consumption figures call for immediate actions also in the building sector to reach emission goals such as the European Union's 20-20-20 targets. The requirements concerning energy efficiency in households are considerably different to those in commercial environments, hence a differentiated strategy must be developed for both of them. The huge number of households, the ever increasing count of energy consuming devices in the homes and the complex relations between energy efforts and the expected environmental effects make them a promising candidate for significant improvements.
In this context smart homes become increasingly popular as they offer improvements, such as increased comfort for their inhabitants at promised energy reductions. Nevertheless, in many cases smart homes cannot fully realize their promises, their systems are still complex to use, seldom tailored to pervasive energy savings and often not characterized by real smartness.
Thus, a novel system concept for smart homes is developed from scratch in this dissertation. The residential area has special characteristics and introduces specific requirements for any control approaches. Most obvious, comfort plays a main role for humans, and mostly energy efficiency comes only second after it. Hence, focus in the system design is given to the realization of energy savings while fully preserving comfort parameters of the users. For this purpose a comprehensive system concept for energy efficiency and comfort in residential homes is developed. It is realized as a multiagent system that implements the main smart home characteristics such as adaptivity, reactivity and proactivity in a dedicated and extensible software framework. It is capable of controlling its environment by exploiting automation technology, adapts to its users, alleviates the users of routine tasks and supports the overall energy efficient operation of the smart home.
The design starts with an extensive requirements analysis, follows an established design methodology and results in a detailed specification of a multiagent based smart home control system. Finally, this system concept is prototypically implemented, tested and evaluated by means of simulation.