VIOLAS: A vision-based sensing system for sentient building models

Abstract We describe the design, implementation, and test of VIOLAS, a vision-based system for object location and occupancy sensing in sentient buildings. Sentient building operations require the existence of a dynamic and self-updating model of building context, components, spaces, systems, processes, and occupancy. Such a model can support applications in building and facility management as well as indoor–environmental controls. Specifically, comprehensive self-updating models can facilitate the implementation of simulation-based building systems control strategies (e.g. for heating, cooling, ventilation, lighting, etc.). Since the underlying model for such operations must possess the capability to autonomously update itself, a versatile sensing mechanism is required that provides context awareness, i.e., real-time facility state information. The research described in this paper aims to examine and demonstrate the potential of vision-based sensing solutions to meet this requirement. For the generation of a comprehensive, self-updating space model, the prototype system particularly requires object identification and location sensing as well as occupancy detection. Toward this end, VIOLAS offers a flexible and scalable arrangement of hardware and software components (tied together via Internet), which is generally well suited to the requirements of sentient buildings.