A simple and non-blocking polarisation-mode dispersion (PMD) monitoring technique using coherent detection is demonstrated and applied to a 40 Gbit=s live fibre-optic system. The results demonstrated a clear correlation between the instantaneous differential group delay and the receiver Q-margin. It was also demonstrated that PMDinduced outage can be eliminated by an adaptive PMD compensator.Introduction: In high-speed fibre-optic communication systems, polarisation-mode dispersion (PMD) is one of the most important factors of performance degradation. In-situ monitoring of PMD in live, multichannel WDM systems will be a key requirement for future dynamic optical networks to ensure quality of operation. Especially, for optical links with exceptionally high PMD values, an in-situ monitoring of PMD during operation will help network engineers to determine the impact of PMD and to select wavelength bands where PMD appears less damaging [1]. Many PMD measurement techniques have been proposed and demonstrated during the last two decades; the most popular include the fixed analyser method, the Jones matrix method, the Poincare arc method and the pulse delay method [2]. All these methods were developed for PMD characterisation instrumentation, which usually require access to both ends of an optical fibre cable to be measured. In live optical networks with installed optical fibres, the source and receiver are at distance and not accessible at the same time. We have recently introduced a novel and simpler method to evaluate differential group delay (DGD) using coherent detection and RF signal processing. This method utilises the spectral characteristics of the digital signal carried by each wavelength channel and measures the PMD-induced polarisation walk-off between the carrier and the clock components [3]. By tuning the wavelength of the local oscillator, the measurement can be performed across various channels in a WDM optical network.In [3], the PMD monitoring technique was briefly introduced and tested with a fixed PMD emulator. In this Letter, we provide a more detailed signal processing algorithm and report the results of our field measurement applying this technique to a 200 km OC768 SONET optical link. Most importantly, we show that this PMD monitoring technique does not rely on the existence of clock components of the signal and therefore it is independent of the optical modulation format. The results demonstrate clear correlation between the instantaneous DGD and the Q-margin at the receiver. We also demonstrate that PMDinduced outage can be eliminated by an adaptive PMD compensator.

Abstract-A fiber-optic low-coherent reflectometer was developed to accurately monitor fiber length variation. A large lengthcoverage range was obtained by using a fiber Bragg grating array in a wavelength-division-multiplexing configuration. The polarization mismatch-induced signal fading was eliminated by applying a polarization spreading technique at the optical receiver and, therefore, no active polarization adjustment was necessary for long-term measurement.Index Terms-Optical delay lines, optical distance measurement, optical fiber measurement, optical interferometry, optical polarization. PRECISE MEASUREMENT of the length of an optical fiber is useful in optic communications as well as in fiber-optic sensors. A popular way to measure fiber length is to use a time-domain reflectometer (OTDR). By evaluating the propagation delay of optical pulses, an OTDR can be used to measure fibers as long as several hundred kilometers. However, the length resolution of a typical OTDR is on the order of meters due to pulsewidth limitations. On the other hand, for high precision measurements, optical low-coherence reflectometers (OLCRs) have been used extensively to measure the distribution of reflections in optical components [1], [2]. The OLCR concept is illustrated in Fig. 1, where a wide-band low-coherent light source is used. Coherent interference happens only when the length difference between the two interferometer arms is shorter than the coherence length of the light source. The length resolution of an OLCR is related to the bandwidth of the light source by [1] ( 1) where is the center wavelength and is the refractive index. A resolution as fine as 10 m can be achieved [2]. The maximum measurement range of an OLCR is usually limited to a few centimeters, which is mainly determined by the length coverage of the scanning optical delay line. Range extension has been proposed in [3] using a pair of retroreflectors in the optical delay line. By letting the light bounce back and forth between the two retroreflectors for times, the length coverage range of the delay line is increased by times. In addition to extra insertion loss in the delay line, another disadvantage of this method is that both motor step size and mechanical errors in the translation stage (which controls the position of the retroreflectors in the scanning delay line) will be amplified by times, degrading the length resolution of the measurement. Furthermore, since an OLCR is based on coherent interference, it obviously requires the matching of polarization states of the signals reflected from both interferometer arms. Due to the random nature of polarization-mode coupling in an optical fiber, polarization-state mismatch may occur, which causes temporary fading of the coherent beating signal at optical receiver.A polarization-independent OLCR has been demonstrated [4], where the low coherent light source was first polarized and then launched into an interferometer composed of polarization maintaining (PM) fibers, thus to eliminate the polarization sensi...

No abstract

Abstract-We propose and experimentally demonstrate a nondestructive method to monitor chromatic dispersion (CD) and polarization-mode dispersion (PMD) in traffic-carrying wavelength-division-multiplexing optical systems. Coherent heterodyne detection is used to down convert the spectrum of digitally modulated signal from optical domain into radio-frequency (RF) domain. By analyzing group delay difference and polarization walkoff between different frequency components through proper RF signal processing, both CD and PMD can be precisely determined. Good agreement between experimental results and theoretical values has been obtained.Index Terms-Chromatic dispersion monitoring, coherent detection, heterodyne, optical communication, polarization-mode dispersion (PMD) monitoring, radio-frequency (RF) signal processing. FIBER chromatic dispersion (CD) and polarization-mode dispersion (PMD) are two major sources of transmission performance degradation in high-speed long-distance optical communication systems [1].So far, several CD and PMD monitoring methods have been proposed. The measurements of eye-opening penalty, -factor, or BER have been proposed earlier to monitor CD [2], [3], but these electrical domain measurements are data rate dependent and are performed on per-channel basis. The phase-modulation-amplitude-modulation (AM) conversion method proposed in [4] and the AM pilot tones method proposed in [5] and [6] require nonstandard transmitters and receivers. The CD can also be determined by monitoring the power of the clock component after photodetection [7]. However, the measured clock power is influenced not only by CD but also by PMD, and these two effects cannot be separated. Sideband optical filtering technique was also proposed recently to monitor CD [8]. In this case, the measurement accuracy may depend on the optical filter's parameters such as shape, bandwidth, and detuning as well as the optical signal modulation format. In practice, the bandwidth and the shape of an optical filter is difficult to control especially when the required bandwidth is very narrow.As far as the PMD monitoring is concerned, the effect of PMD on an optical system can be evaluated by monitoring the degree of polarization of the optical signal [9], but the results may be sensitive to the variation of signal optical spectrum and the modulation bandwidth. Vestigial sideband optical filtering has been recently proposed to evaluate PMD through measuring the strength reduction of the beating signal between the optical . Although this method is insensitive to effect of CD, the absolute signal strength at the beating frequency depends not only on PMD, but also on the modulation format and the spectral distribution of the optical signal. Therefore, a quantitative and reliable evaluation of PMD might not be possible. Very recently, coherent frequency-selective polarimeter was proposed which has the ability to evaluate PMD [12]. In this method, the measurement of different frequency components within the signal bandwidth is accomplished b...