Optical performance monitoring in short reach optical communication systems

With the development of advanced optical communication systems and networks, the critical role of optical performance monitoring (OPM) in system and network operation has emerged. OPM can help service operators to provide users with guaranteed signal transmission, and it can be conducted from both optical domain and electrical domain. In short-reach communication systems, intensity modulation/direct detection (IM/DD) is typically used, and channel parameters such as optical signal-to-noise ratio (OSNR) and chromatic dispersion (CD) are concerned by system management. In high-speed optical fiber communications, CD estimation is essential for the monitoring and compensation of CD-induced impairments. The existing dispersion estimation techniques usually require training signals or a large amount of data for curve fitting or machine learning. However, these methods cannot provide in-service estimation. In addition, compared with electronic monitoring, optical domain OPM has shown competitive advantages because it does not require opto-electronic conversion and can be detected from each node of the system without interrupting the normal operation. OSNR is a widely used optical performance indicator since it suggests the impairments of the system. However, OSNR cannot reflect the performance of systems as accurate as the bit error rate (BER).  Therefore, the objective of this research is to establish a theoretical framework of dispersion, OSNR and SNR/BER analysis for short-reach optical communication systems, considering different types of impairments and limitations. The key questions to answer are the theoretical modelling of dispersion, OSNR and the nonlinear relationship between OSNR and SNR/BER, to provide an easily interpretable optical performance monitoring parameter for short-reach optical communication systems and networks. In this thesis, we first developed a novel technique for in-service CD estimation for IM/DD channel with the random input signal using power spectrum analysis. In addition, we proposed a complete and accurate analytical model which was verified with numerical simulations. Unlike previous studies that only work with single-frequency sinusoidal signals, the practical random data is considered in our model. Therefore, our method can be applied to the dispersion estimation of various pulse amplitude modulation (PAM) based systems. results show that better than 0.5% error rate in dispersion estimation can be achieved in optical communication systems with up to 40 GBaud data rate and 100 km transmission distance. In addition, for the modelling of OSNR and SNR monitoring, a theoretical framework for OSNR and SNR/BER analysis in IM/DD based short-reach optical communication systems under various limitations is developed. The theoretical model is based on power spectrum analysis and can link OSNR with SNR/BER for accurate data transmission quality estimation. Different from previous models, which only apply to amplified spontaneous emission (ASE) noise limited systems, our model can be applied to non-ASE limited IM/DD systems. The proposed analytical model is verified and investigated via numerical simulations, where results show a maximum 0.3% error rate. The work above can be applied in static IM/DD systems for accurate data transmission quality estimation, where the dispersion is assumed to be fully compensated. However, in dynamic short-reach optical communication networks with reconfigurable network topologies, CD is difficult to be fully compensated and it affects the SNR performance as well. One example is the elastic optical networks (EON), where both the wavelength and the transmission distance change dynamically, causing varying dispersion at the receiver side. Therefore, we further proposed a theoretical framework of OSNR and SNR/BER analysis for IM/DD modulated optical communication systems with fibre dispersion. Results consistently match the predictions of the system simulation, with a maximum estimation error rate 1.5%.