Joint Phase Noise Estimation and Symbol Detection in Optical Communication Systems
Examensarbete för masterexamen
Communication Engineering (MPCOM), MSc
Demands for higher transmission rates are ever increasing as a result of requirements imposed by a number of different applications. To meet these demands, world-wide research efforts for optical communications have led to tremendous growth in performance in the past few decades. Nowadays, optical communication systems find applications in various scenarios, ranging from backbone networks to data centers. In recent years, coherent communication systems in fiber-optical channels have attracted interest in the academic and industrial communities. Coherent optical systems enable the encoding of information in both the amplitude and phase of the signal, which can result in improved performance with respect to spectral and power efficiency. Furthermore, advances in the electronic hardware have facilitated effective signal impairment compensation using digital signal processing, allowing the adoption of algorithms from wireless communications. A particular impairment is laser phase noise, which can severely limit the gains promised by coherent optical systems. In this thesis the problem of optimal symbol detection in the presence of laser phase noise in uncoded optical communication systems is studied. We assume a single carrier transmission on two independent polarizations in a channel that includes phase noise, additive white Gaussian noise, and a random constant phase offset on each polarization component. To this end, the maximum a posteriori (MAP) symbol detector is presented, which is shown to be analytically intractable. Then, a pilot-based algorithm developed using the factor graph framework and the sum product algorithm, that jointly estimates phase and detects symbols for arbitrary quadrature amplitude modulation (QAM) constellations, is studied. Performance is evaluated using Monte Carlo simulations for quadrature phase shift keying, 16-QAM and 64-QAM constellations. Results show that this algorithm is more tolerant to phase noise compared to other algorithms found in optical literature.
Elektroteknik och elektronik , Electrical Engineering, Electronic Engineering, Information Engineering