A semantically grounded Digital Twin Architecture for occupant-centric indoor lighting optimisation using a multimodal framework

Kavuthimadathil, Sreelakshmi, BK, Satish and Rezgui, Yacine ORCID: https://orcid.org/0000-0002-5711-8400 2026. A semantically grounded Digital Twin Architecture for occupant-centric indoor lighting optimisation using a multimodal framework. Presented at: ECPPM 2026, Cardiff, UK, 9-11 September 2026. ECPPM 2026 – European Conference on Product and Process Modelling. Zenodo, 10.5281/zenodo.22680140

PDF - Published Version Download (468kB)

Abstract

Indoor lighting is a major contributor to occupant health issues in the built environment, as occupants spend 90% of their time indoors. Poor lighting conditions can cause inefficiency, discomfort, and health problems, significantly contributing to the costs of absenteeism in countries. Lighting significantly influences mood, circadian rhythms, alertness, and productivity, and poor lighting can cause problems such as eyestrain, headaches, and sleep loss. The current building management systems treat building lighting as a static, schedule-driven utility rather than a dynamic, occupant-responsive system. Digital twin technology can shift lighting control from rule- based to adaptive, intelligence-driven when the essential layers, including the semantic, simulation, and learning layers, are properly integrated. This paper proposes a bottom-up, layered approach to a semantically grounded lighting DT architecture that can optimise lighting and close the control loop through rich, multimodal human feedback. The architecture proceeds from the input data and simulation upward through a semantic layer to a surrogate model trained on outputs from lighting simulation engines and ultimately to an adaptive machine learning control layer that closes the loop with occupant feedback. The paper presents a thorough literature review of existing gaps in the effects of lighting on human physiology and cognition, cognitive and subjective assessment instruments, and digital and semantic integration for human-centric buildings. The paper later analyses simulation engine data structures and their semantic mappings, identifies ontological gaps in the photometric and human behavioural domains, and proposes ways to address them. Key occupant parameters, illuminance and CCT preference, glare, luminous contrast of the screen, demographic profile and task context are grounded as typed inputs and outputs within the semantic layer. The framework is positioned as the foundation for experimental validation in an office cubicle environment.

Item Type:	Conference or Workshop Item - published (Paper)
Date Type:	Publication
Status:	Published
Schools:	Schools > Architecture Schools > Engineering
Publisher:	Zenodo
Date of First Compliant Deposit:	19 October 2026
Last Modified:	19 Oct 2026 10:41
URI:	https://orca.cardiff.ac.uk/id/eprint/190284

Actions (repository staff only)

Edit Item