Initial situation: In Europe, buildings are responsible for a significant proportion of primary energy use. Targeted control of building management systems, based on automatic detection of people in the building or in sub-areas of the building, can achieve energy savings of up to 30% without reducing user comfort. Currently implemented systems are based on different technologies and methods. In most cases, the installation of these systems requires significant effort, both through placement of sensors and the associated cabling effort. Some technologies are also problematic, raising legitimate privacy concerns. The expected high installation cost, especially in existing buildings, prevents implementation due to economic concerns. Other approaches that build on existing infrastructure have little or no additional installation effort, but due to the main task being focused elsewhere, the results to be achieved are not reliable. Recognizing people based on the reflection of visible light, caused by the person themselves is known by the synonym Visible Light Sensing. Visible Light Sensing represents an innovative way to use existing lighting infrastructure for people recognition without privacy concerns. However, currently described systems do not include the communication interfaces to be used and the aspect of networking several such innovative lighting fixtures in their investigations.
Aims and innovation content: The aim of the project is to test the economic and technical feasibility of an approach for a network of lighting fixtures that perform person recognition based on the technology of Visible Light Sensing. One of the main aspects is the differentiation of persons and groups, as well as the recognition of parameters such as speed of persons and, if necessary, the detection of abnormal conditions by networking the lighting fixtures. In order to make the planned system economically feasible for use in existing buildings, another focus of the project is to enable communication between the lighting fixtures and with a higher-level system without the need for any additional installation of cabling. Machine learning algorithms are to be used as a basis, both for the detection and the classification, to be performed.
Results and Findings: The result of the project shall be the evaluation of a concept for a network of lighting fixtures, which determines with a high success rate the number of persons in a subarea of a building. In addition, parameters such as the speed of the people shall be determined. The installation effort shall be minimally small and therefore economically feasible. Likewise, the system shall be highly modifiable and a SWOT analysis shall be performed to assess the technical and economic advantages and disadvantages of the approach. Based on the SWOT analysis, a Stop/Go decision will be made regarding a follow-up projects.