Signal processing techniques for optical performance monitoring in dynamic fiber-optic networks

Abstract

Future fiber-optic transmission systems are envisioned to evolve from static point-to-point links to completely reconfigurable optical networks in order to meet the high data rate requirements and to promote flexibility and efficient utilization of available network resources. The dynamic nature of such networks will impose stringent requirements on the control and management of these complex networks. Optical performance monitoring (OPM) is an enabling technology for the reliable and stable operation of emerging dynamic optical networks. The installation of monitoring mechanisms across the optical network, which could incessantly monitor the health of network as well as data signals, may enable several advanced network functionalities such as automatic fault-identification and management, dynamic resources allocation, adaptive impairments compensation and impairment-aware routing. This thesis focuses on the development of cost-effective OPM techniques capable of monitoring two critical network performance parameters: optical signal-to-noise ratio (OSNR) and fiber chromatic dispersion (CD). Digital signal processing (DSP) techniques exploiting the statistical properties of the data signals have been applied. These techniques may facilitate low-cost monitoring of multiple physical parameters of the channel simultaneously and can be applied to monitor several data rates and modulation formats without necessitating modifications of monitoring devices. In particular, two cost-effective in-band OSNR monitoring techniques have been proposed and experimentally demonstrated for intensity and phase-modulated systems. Both techniques employ asynchronous delay-tap sampling (DTS) and DSP and hence enable OSNR monitoring at various data rates without requiring clock information. Using simple analytical models for the statistics of signal and noise samples, these techniques facilitate calibration-free OSNR monitoring with good accuracies and monitoring ranges. The adverse effects of dispersive impairments on the accuracies of both OSNR monitoring techniques have been probed through numerical simulations and the tolerances have been ascertained. In addition, a technique for monitoring CD in multiple return-to-zero (RZ) intensity and phase-modulated systems is proposed and experimentally validated using asynchronous sampling and subsequent processing of two vestigial sideband (VSB) signals. This technique utilizes simple hardware and signal processing and demonstrates CD monitoring for several RZ modulation formats and different data rates with good monitoring ranges and sensitivities. The proposed technique is shown to be resilient against deleterious noise effects thus enabling OSNR-independent CD monitoring. Finally, a joint OSNR and CD monitoring technique using the empirical moments of asynchronously sampled signal amplitudes is proposed and validated through numerical simulations. The first three raw moments of asynchronously sampled signal are analytically formulated and it is shown that the solutions of derived equations can successfully decouple the effects of OSNR and CD thus enabling simultaneous and independent monitoring of these two parameters. The proposed technique employs simple hardware and signal processing to facilitate OSNR and CD monitoring for various modulation formats and different data rates with good accuracies and monitoring ranges. Furthermore, the CD monitoring ranges demonstrated by this technique are several times larger than the ones exhibited by most of the existing CD monitoring techniques.