Tracking and Positioning Dynamic Targets in VLC: Signal Dependent Noises and Bayesian Bound Analysis

This paper investigates the problem of position estimation and tracking in indoor visible light communication (VLC) systems under a multiple-input single-output (MISO) configuration. The study specifically addresses the impact of signal-dependent shot noise (SDSN) and relative intensity noise (RIN), which significantly degrade localization accuracy in practical VLC scenarios. To establish a theoretical performance benchmark, a novel closed-form Bayesian Cramér-Rao lower bound (BCRLB) is derived for two-dimensional (2D) position estimation. The proposed extended Kalman filter (EKF) demonstrates superior localization accuracy by effectively accounting for motion dynamics through its recursive updates, unlike estimation methods that rely solely on noisy distance measurements. Our findings show that increasing the number of light sources improves spatial diversity, thereby enhancing positioning precision. Furthermore, analysis indicates that RIN has a more pronounced negative effect on tracking accuracy than SDSN. The EKF’s performance is validated against the proposed BCRLB, confirming its suitability for real-time and reliable indoor localization. Overall, this work provides meaningful insights into the design of robust and efficient VLC-based positioning systems.