Agent based control in the Smart Home

The term Smart Home promises an energy efficient and comfort oriented operation of buildings. While the technical foundations have mostly been laid (e.g. integration of entertainment devices), also the complexity of the systems reached a peak level. Users find it hard to understand their systems and, worse, to tailor them to their individual needs. At the bottom line, smart homes therefore often fail to exploit all their potentials, among others due to non-optimized control strategies, frequent manual comfort overrides or simple user dissatisfaction. In this paper agent based systems are applied to control a smart home, with the main reason to exploit built-in intelligence and autonomous operation capabilities to tackle the existing shortcomings. In particular, it is shown how agent characteristics such as goal drivenness, robustness and learning capable behavior complement a smart home control system. Further focus is put on describing the complete process from selection of the most suited agent approach for smart homes to detailing the transition from system specification to implementation.