Genetic Algorithm Optimisation of the SNR for Indoor Optical Wireless Communication Systems

A genetic algorithm controlled multispot transmitter is proposed as an alternative approach to optimising the SNR for single element receivers in fully diffuse mobile indoor optical wireless communication systems. Results are presented that show by dynamically controlling the powers of individual diffusing spots, a consistent SNR distribution can be created in a deployment environment, independent of the room illumination sources, reflection characteristics and user movement patterns. The advantageous adaptability of the genetic algorithm approach removes the need for bespoke system design and provides a more cost effective system, capable of deployment in multiple locations, whilst improving end user friendliness. The case study scenarios considered in this paper are based around the use of a single element receiver with a FOV = 55 degrees in two non-identically illuminated rooms. A genetic algorithm is then utilised to show the concepts capability to reduce the dynamic range of the SNR, referenced against the peak SNR, by up to 13% when empty. Furthermore, the algorithm is then shown to be capable of reducing the user movement induced SNR perturbation by up to 11%, and maintaining to within 21% of this optimised case, the SNR distribution.