LiGen: GAN-Augmented Spectral Fingerprinting for Indoor Positioning
Jie Lin, Hsun-Yu Lee, Ho-Ming Li, Fang-Jing Wu
TL;DR
LiGen tackles indoor localization by replacing fragile radio fingerprints with stable ambient-light spectral fingerprints and strengthening data scarcity through GAN-based augmentation. It introduces two augmentation variants: PointGAN, which conditions synthesis on coordinates, and FreeGAN, which uses a weak localization model to pseudo-label unlabeled samples; together they train a compact MLP that maps 10-channel spectral fingerprints to 2D coordinates. Empirical results show submeter accuracy, outperforming Wi‑Fi baselines by over 50%, and robust performance in cluttered environments. This infrastructure-free, sensor-based approach offers a practical pathway toward scalable, accurate indoor positioning with minimal deployment cost.
Abstract
Accurate and robust indoor localization is critical for smart building applications, yet existing Wi-Fi-based systems are often vulnerable to environmental conditions. This work presents a novel indoor localization system, called LiGen, that leverages the spectral intensity patterns of ambient light as fingerprints, offering a more stable and infrastructure-free alternative to radio signals. To address the limited spectral data, we design a data augmentation framework based on generative adversarial networks (GANs), featuring two variants: PointGAN, which generates fingerprints conditioned on coordinates, and FreeGAN, which uses a weak localization model to label unconditioned samples. Our positioning model, leveraging a Multi-Layer Perceptron (MLP) architecture to train on synthesized data, achieves submeter-level accuracy, outperforming Wi-Fi-based baselines by over 50\%. LiGen also demonstrates strong robustness in cluttered environments. To the best of our knowledge, this is the first system to combine spectral fingerprints with GAN-based data augmentation for indoor localization.