Font Size: a A A

Research Of Super-Nyquist Optical Transmission Systems And Digital Signal Processing Techniques

Posted on:2016-06-22Degree:DoctorType:Dissertation
Country:ChinaCandidate:S ChenFull Text:PDF
GTID:1108330482957713Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
Internet has been constantly growing. The convergence with industry economy may trigger another round of data traffic increase, which ultimately shifts the traffic burden to underground optical fiber networks. Commercially available class of single-band (C-or L-band) optical amplifiers provides about 5 THz bandwidth. To pack more information into the limited bandwidth has been the goal of optical transport technologies. Nowadays, optical orthogonal frequency division multiplexing (O-OFDM) and Nyquist wavelength division multiplexing (Nyquist-WDM) techniques have been extensively studied due to the high spectral efficiency (SE). However, Nyquist’s theorem points out the SE limit for orthogonal transmission.The super-Nyquist optical transmission technique introduces inter-symbol interferences (ISI) to overcome the constraint of Nyqusit’s theorem and provide higher SE. ISI is handled by digital signal processing (DSP). However, with the existence of ISI, the whole system becomes more vulnerable to various transmission impairments. The DSP detection scheme is facing critical problems such as interactional impairments, high complexity, and low ISI-tolerance, which have been the limitations of super-Nyquist systems. A comprehensive evaluation of performance and computation complexity is needed. This paper investigates the combined optimization mechanism of channel equalization, carrier recovery and ISI compensation, acquires suitable detection schemes for 4 bit/s/Hz PDM-QPSK WDM systems, develops a method that can compress the duration of ISI via delay-&-add operation to reduce computation complexity, and finally proposes an adaptive detection scheme with high ISI-tolerance. Some results are first demonstrated. In this paper, the main work and contributions are listed as follows:1) Aiming at non-orthogonal transmission with the existence of ISI, we investigate transmission modelling, optimal detection theory and the advantage of a super-Nyquist system. The analysis shows that a super-Nyquist system have higher capacity than a conventional Nyquist system, the super-Nyquist technique can use physically realizable pulses to reach or surpass 2 Baud/s/Hz SE, and a super-Nyquist system can use low-level modulation formats to achieve high SE. The analysis of optimal detection theory shows that the super-Nyquist technique is basically using DSP power to acquire higher SE.2) Aiming at problems of interactional impairments and high complexity of current DSP detection scheme, we investigate combined optimization mechanism of channel equalization, carrier recovery and ISI compensation, compare the performance of maximum likelihood sequence estimation (MLSE) and maximum a posteriori (MAP) in a super-Nyquist system, and evaluate a method that can compress the duration of ISI via delay-&-add operation to reduce computation complexity. Simulation results show that, the performance improvement does not necessarily require the increase of DSP’s computation complexity. An optimized detection scheme with 2 tap MLSE may have better performance than non-optimized detection scheme with multi-tap MLSE. In an experiment study,112 Gbit/s PDM-QPSK WDM signals are multiplexed on a 25 GHz channel spacing, QDB detection+2 tap MLSE has preferable performance with low complexity, which is suitable for 4 bit/s/Hz PDM-QPSK WDM systems.3) Aiming at the problem of low ISI-toleration of current DSP detection scheme, we propose an adaptive detection scheme with high ISI-tolerance, to further increase the spectral efficiency of a super-Nyquist WDM system. The detection scheme utilizes polybinary modulation combined with MLSE equalization, and the receiver side DSP defines and adjusts the polybinary modulation levels according to channel spacing in order to acquire a better result. Simulations realize the detection of 112 Gbit/s PDM-QPSK signals that are multiplexed on a channel spacing ranging from 15 GHz to 28 GHz. In the experiments,112 Gbit/s PDM-QPSK signals multiplexed on a 20 GHz channel spacing are transmitted over 960 km standard single mode fiber (SSMF) at hard-decision forward-error-correction (HD-FEC) threshold (3.8×10-3), achieving 5 bit/s/Hz spectral efficiency. To the best of our knowledge, this is the highest SE for a QPSK system with HD-FEC.4) Aiming at the ISI issue in directly detected wavelength division multiplexing passive optical networks (WDM-PON), we propose a frequency domain equalization scheme with time-clipping method to improve noise tolerance. With the analysis of the relationship between time-domain ISI response and frequency-domain ISI response, we find that the update of filter coefficients in time domain with reduced length may decrease the effects of noise corruption. Simulation results show that the time-clipping method with optimized clipping position and symbol number can provide better results than conventional time domain equalization and frequency domain equalization. In a 10 Gbit/s NRZ-OOK experiment study, the modified frequency domain equalization has about 1 dB performance improvement than time domain equalization.
Keywords/Search Tags:Super-Nyquist wavelength division multiplexing systems, spectral efficiency, digital signal processing, maximum likelihood sequence estimarion, maximum a posteriori
PDF Full Text Request
Related items