Optical IRS-Enhanced Indoor Visible Light Positioning: Maximizing the Minimum Euclidean Distance of RSS Fingerprints

Among various indoor positioning technologies, visible light positioning (VLP) has emerged as a promising solution owing to its high accuracy and compatibility with existing infrastructures. Besides, the received signal strength (RSS) fingerprint method is commonly utilized in VLP, but suffers from the noise-induced confusion owing to the similarity of RSS values. Recent advancements in optical intelligent reflecting surface (OIRS) present new opportunities to enhance the performance of VLP systems. In this paper, we investigate the integration of OIRS with RSS fingerprint-based VLP systems to improve the indoor positioning accuracy through dynamic configuration of the RSS distribution. Specifically, the approach involves a two-stage scheme that initially narrows down the positioning range in a coarse localization stage, followed by a fine localization stage that pinpoints the position employing OIRS to enhance the difference among RSS fingerprints at nearby locations, thereby improving the positioning accuracy. Then, the problem is formulated to maximize the minimum Euclidean distance of the RSS fingerprints by jointly optimizing the OIRS alignment and power allocation. Subsequently, a block coordinate descent-based optimization algorithm including successive convex approximation and semi-definite relaxation methods is proposed to solve the problem iteratively. Moreover, simulations demonstrate that the proposed OIRS-assisted VLP system substantially outperforms conventional benchmarks, achieving superior positioning accuracy in indoor environments, and also highlight the potential of OIRS in supporting VLP under adverse conditions.