Human occupancy recognition with multivariate ambient sensors
Buildings account for roughly 40% of all U.S. energy use, and HVAC systems are a major culprit. The goal of this research is to reduce power consumption without sacrificing human comfort. This paper presents a cooling demand estimation from heat generation to assess the quantity of cooling supply, which helps diagnose potential problems in the HVAC system. A negotiation-based approach is proposed to balance power consumption, cooling for human comfort, and smooth operation for equipment health. Experiments were conducted with the NTU CSIE July 2012 dataset  as well as online live experiments in the computer science building on campus. The experiments demonstrated that the proposed method reduced 3.81% to 5.96% of power consumption with consideration of smoothness.
Unfortunately, ACM prohibits us from displaying non-influential references for this paper.
To see the full reference list, please visit http://dl.acm.org/citation.cfm?id=2516918.