Genetic Optimization of Liquid Crystal Matrix Based Interference Suppression for VLC MIMO Transmissions

Abstract

Optical channel decorrelation is a key aspect in designing a multi-user visible light communication system. As light-emitting diodes (LEDs) are commonly used for illumination and data transmission in such a system simultaneously, the interference between them is a major problem for suitable modulation schemes. We have proposed to position liquid-crystal (LC) matrix elements in front of the receivers in order to block unwanted light-so effectively acting as filter and reducing the interference. To do so, the appropriate pixels of the LC filter are switched either to a transmissive state or to an opaque state. The selection of the correct pixel patterns is a high-order optimization problem, which makes it challenging for classic algorithms. In this paper, we suggest applying a genetic optimization to find suitable pixel patterns. We combine this kind of receiver-side optimization with conventional convex precoding at the transmitter side in an iterative manner. The overall optimization goal is to maximize the sum rate over all users. To prove the effectiveness, we use a numeric ray-tracing simulation with different settings. Also we analyze the impact of the main parameters for the genetic algorithm.