An efficient scheduling mechanism for building energy management systems

The continuous growth of energy consumption has become a critical issue globally. According to the international energy agency, the energy consumed in residential and commercial buildings contributed about 40% of the total energy consumption in 2014. In recent years, building energy management has become a popular research topic. The techniques used in automatic control have evolved from static schedules specified manually into real-time sensing through wireless sensor network. The demand for a dedicated platform for building energy management systems has emerged later. Many proposed platforms focused on the high-level architecture and were based on existing standards and technologies. However, there was little research about the fundamental aspects of the infrastructure. In this paper, we propose a scheduling mechanism for the fundamental infrastructure based on the distributed pinwheel model. It addresses the issues of distributed real-time computing, and provides a more predictable system behavior for potential improvement of energy consumption of the infrastructure itself. Unlike the original distributed pinwheel model, we also take network delay into account, and present an adaptive method to mitigate the effects of variable network delay. The simulation results suggest that the total end-to-end delay can be reduced to 20% or lower for more than half of the input task sets without the presence of variable network delay. The adaptive method shows an improvement of 10.06% to 61.16% with the presence of variable network delay.