2013 · 18 citations · 13 references
ArchitectureAutonomous SystemsBuilding ScienceAutomationEnergy ManagementEnvironmental ControlAgent-based DesignBuilding AutomationBuilding Automation SystemsBuilding ControlsIntelligent AgentsSmart Building TechnologiesAdaptive Building SystemsBuilding Energy OptimizationComfortEnergy ConsumptionEnvironment DynamicsSmart Building SystemsSmart BuildingsUser Preferences
Building automation systems (BAS) are technologies which automatically control building appliances based on specialized algorithms to achieve predefined goals, such as optimising user comfort or regulating energy consumption. Normally, they are configured during deployment as a one-time effort by technicians with expert knowledge and tools. However, the dynamic nature of the environment and the preferences of occupants calls for the need of frequent reconfigurations, which is often impractical and expensive to implement. As a solution, we propose a novel BAS design which considers user-based information such as user preferences and feedback in order to continuously reconfigure itself, as opposed to static configurations. This high-level information is then adopted by a software framework based on multi-agent systems (MAS) that applies the PROSA reference model [9] to dynamically control the building. Under this framework, every room in the building is automated by a dedicated agent, acting as a software mediator between the appliances and its occupants, with the aim of maximising the result of a utility function taking into account power optimisation, user comfort, and environment dynamics. The proposed approach has been implemented in a real office environment, and we discuss our design to improve user experience a BAS deployment.
13
Reference architecture for holonic manufacturing systems: PROSA
Hendrik Van Brussel, Jo Wyns, Paul Valckenaers et al. · Computers in Industry · 1998 · 1.4K citations
Engineering, Manufacturing Engineering, Systems Engineering +12
Communication systems for building automation and control
Wolfgang Kästner, Georg Neugschwandtner, Stefan Soucek et al. · Proceedings of the IEEE · 2005 · 476 citations