SVM-Assisted Realization and Demonstration of Indoor 4 Mb/s Non-Line-of-Sight Visible Light Communication With Second-Order Reflection

Abstract

Non-line of sight (NLOS) visible light communication (VLC) from reflections can expand the application of VLC beyond the line-of-sight (LOS) link. We demonstrate the experimental results on such indoor NLOS VLC, up to the second-order reflection. The avalanche photodiode (APD) receiver is employed due to the strong ambient light background. Manchester coding is adopted at the transmitter, and digital filtering is adopted at the receiver. We propose an Support Vector Machine (SVM) detection method to solve the significant distortion issue due to weak signal of the NLOS link. Experimental results on the second-order reflection-based VLC demonstrate that the detection error probability of the proposed SVM detection can be reduced up to at least two orders of magnitude compared with the Gaussian approximation-based detection, and can reach 10 −4 at bit rate 4 Mbps. Thus, low-data rate applications can be achieved by VLC with only the second-order reflection, which implies the potential in speech/image/control message transmission and Internet of Things (IoT) applications, for example the mobile robots that may operate in the shadow or even under a table. Experimental results on the second-order reflection at different transmission rates and different sampling rates are also demonstrated.

Breaking the limitations of visible light communication through its side channel

“…Therefore, it is always challenging to achieve a reliable VLC system because surrounding mobile/stationary objects can easily block the LOS links. To tackle this problem, researchers have proposed methods in three dimensions: at the transmitter side [4,5], the propagation channel [49], and the receiver/user side [6]. To reduce the probability of a VLC link being blocked, the authors in [4,5] propose to deploy dense VLC transmitters to avoid the blockages.…”

“…Such a process is not straightforward due to the complicated reflection channels and weak amplitude of the received light signal. Although advanced techniques such as machine learning could be adopted to infer the data from the reflected light [49], the training cost is very high. People who host the VLC receivers could rotate their bodies or the receivers to avoid blockages [6].…”

A Manchester-OOK Visible Light Communication System for Patient Monitoring in Intensive Care Units

“…Experimental results demonstrate the applicability of their solution in links that reach ≈ 6 m with relatively good performance (BER ≈ 1 × 10 −3 ). In [33], this line code was adopted to suppress low-frequency components of transmitted signals to alleviate impairments introduced by the bias-tee. The authors demonstrate 4 Mbps indoor communication in non line-of-sight VLC links slightly farther that 1 m. The 12 kb/s OCC system evaluated in [34] also employed Manchester line code.…”