Towards an Occupancy Count Functionality for Smart Buildings - An Industrial Perspective

This paper focuses on an overview and evaluation of existing approaches for occupancy detection and people counting within commercial buildings forming the input for a variety of optimization functions for increasing energy efficiency as well as improving the safety within buildings. Existing approaches are compared among each other using a Pugh Matrix approach by looking to which degree each technology fulfills the most important functional and non-functional requirements related to various occupancy detection use cases. Based on the outcome of the Pugh Matrix, we select two different approaches for prototypical evaluation. All of them are privacy-preserving and are easy to implement and deploy. For the most suitable approaches, namely passive infrared (PIR) and millimeter waves (mmWaves) technology, a detailed evaluation is conducted. Based on our results, we demonstrate that for scenarios with up to five people in a room the number of occupants can be determined with an accuracy of +/− one within the time frame of a few seconds, thereby delivering a comparable performance to vision-based approaches.