Moment based optimization of optically preamplified systems in the presence of jitter

Abstract

Optically preamplified direct detection receivers offer sensitivities that approach the quantum limit at very high bit-rates [1, 2]. Several alternative approaches to the modelling of these systems have been reported in the literature: Yamamoto [3] used a Gaussian approximation (GA) based on the mean and variance of the sampled received data to estimate the performance of an SLA based system. As noted by Fyath [4], although the GA provides a useful first approximation a more detailed knowledge of the PDF or moment generating function (MGF) of the signal following amplification is required if a more rigorous performance assessment is desired. Personick [5] developed an MGF for the number of primary counts produced by an optically preamplified receiver and used the Chernoff bound to evaluate an upper bound on the error rate. Although powerful, the method was not developed to the state were it could be applied to practical systems. Fyath [4] developed an MGF for an optically preamplified receiver and then applied the much tighter modified Chernoff bound (MCB) to assess system performance [6].