Implementation and Assessment of a Bayesian-Fuzzy Wireless Daylighting Systems: Results from a two-person office space. Jessica Granderson Yao-Jung Wen Professor Alice Agogino Department of Mechanical Engineering UC Berkeley, Berkeley, CA 94720 jgrander, rio}@me.berkeley.edu agogino@berkeley.edu Abstract: Recent research in daylighting systems has brought new focus to sensing and actuation technologies, and increasingly sophisticated control that accommodates specific user needs and increased efficiency. In keeping with these efforts, this paper presents the implementation and assessment of an intelligent wireless daylighting system, within a two-person office space. The intelligent system comprises a powerful combination of three primary elements: fuzzy-logic sensor validation and fusion, wireless sensing and actuation from a single bi-functional platform, and an influence diagram decision algorithm that balances the needs of individual occupants and facilities management. The intelligent system was implemented in tandem with a WattStopper 1 system, and comparative user testing was conducted to determine: quality and feasibility of system design in terms of user satisfaction with sensor location, method of manual override, preference-based decision making, and cost. Results from this initial iteration of user testing indicated that the method of manual override and preference-based decision making components of the intelligent system were preferred with respect to the commercial system. The results were inconclusive with regard to comfort with desktop placement, however preferred physical arrangements upon the desktop were successfully identified. Finally, it was determined that the intelligent system can be cost-competitive with existing commercial daylighting systems. Keywords: Daylighting, Bayesian and Fuzzy intelligence, wireless sensing and actuation 1 Model No.: LS-301Dimming Photosensor