Reconciling Approaches to SNR Analysis in Optical Wireless Communications

Abstract

Optical communications, both fiber-optics and optical wireless (OWC), have a proven potential for high speed data transmission due to the vast optical spectrum. Given the ever increasing wireless data demand, we look towards OWC as a supplemental medium in heterogeneous networks (HetNets). In such networks, fair comparison of available connections is required for optimal allocation of resources. This comparison is essential for provisioning and analysis as well as in the decision metrics for handover and aggregation. As the optical communications field has developed, Signal to Noise Ratio (SNR) has been defined in various ways. Instantaneous optical power constraints have led the fiber- optics community to define SNR relative to peak- to-peak optical signal power. Additional average optical power constraints have led the OWC community to observe SNR relative to average optical signal power rather than average electrical signal power. Accordingly, the relationship between SNR and link performance has also changed. While the various definitions offer fair comparison under similar constraints, relative analysis is necessary when comparing the quality of multiple links in a HetNet or performance of networks operating under different constraints. In this paper, we reconcile three common SNR definitions and use symbol error rate (SER) analysis for pulse amplitude modulation in order to exemplify how different definitions affect the SER versus SNR curves.