Predictive modeling of occupancy patterns in smart buildings

The intensive research for developing smart environments is based on building energy management systems. There exist demonstrated results that indicate a significant correlation between measured occupancy status and environmental conditions. In this context, the occupancy behaviour plays a key role to illustrate the consequent energy impact in energy savings HVAC control strategies, security, energy demand and interactions with space equipment. This paper presents the modeling of occupant presence and prediction of status - vacant or occupied - in smart buildings. For this objective Markov chain models are used and the results suggest potential energy savings for the input that they bring in HVAC control strategies. Forecasting results based on realistic data sets are discussed.

I. Introduction

The majority of the world standards for buildings consider empiric examples for occupancy scenarios when low consumption is discussed. In smart buildings, the human activity should always be the predominant factor in scenarios of HVAC control, interaction with physical equipment, etc. The human behaviour and occupancy prediction become a determinant consumption factor with high impact in building's performance [1].