Indoor Intruder Tracking Using Visible Light Communications.

This paper proposes a comprehensive study of indoor intruder tracking using visible light communication (VLC). A realistic indoor VLC channel was developed, taking into consideration reflections, shadowing, and ambient noise. The intruder was considered smart and aiming to escape tracking. This was modelled by adding noise and disturbance to the intruder's trajectory. We propose to extend the application of minimax filtering from state estimation in the radio frequency (RF) domain to intruder tracking using VLC. The performance of the proposed method was examined and compared with Kalman filter for both VLC and RF. The simulation results showed that the minimax filter provided marginally better tracking and was more robust to the adversary behavior of the intruder than Kalman filter, with less than 0.5 cm estimation error. In addition, minimax was significantly better than Kalman filter for RF tracking applications.