| This paper presents a device-free visible light positioning framework for indoor human localization using passive optical sensors mounted on room walls. The proposed method exploits time-of-arrival and angle-of-arrival information extracted from optical signals that are altered by the presence of a human in the environment. Measurements collected from multiple sensors are organized into structured signal representations and used to train a convolutional neural network regressor that estimates spatial coordinates. To produce realistic training and evaluation data, a physics-based Monte Carlo light propagation simulator is employed, accurately modeling multipath reflections and atmospheric scattering phenomena. Experimental results demonstrate that the proposed approach provides accurate and robust localization performance, clearly outperforming a conventional fingerprinting-based method. |
*** Title, author list and abstract as submitted during Camera-Ready version delivery. Small changes that may have occurred during processing by Springer may not appear in this window.