The multiplicative SPAD Noise effects on the performance of adaptive RLS incremental networks implemented with the VLC technology

Noise is the main distorting parameter in the performance of all communication systems including adaptive networks. The problem of distributed estimation with adaptive networks has been greatly related with the presence of noise in the nodes and in the links that are connecting these nodes. There are various types of noises according to their different sources. The coefficients of these noise types follow certain distributions that are determined with numerous trials. Recently, with the development of the visible light communication (VLC) systems, the implementation of wireless adaptive networks has been recommended with this technology. The VLC systems have their own noise type regarding the usage of single photon avalanche diodes (SP ADs). The usage of these SP ADs will lead to the generation of the multiplicative sub-Poisson noise which will show its effects in the links between the connecting nodes. In this paper, we examined the effects of this unique type of noise on the performance of adaptive recursive least squares (RLS) incremental networks. Furthermore, the effect of the combined Gaussian, sub-Poisson noise model is examined and compared with the only Gaussian noise model.