Investigation Of Key Techniques For Digital Coherent Fiber Optical Superchannel Systems

With the increment of the high-bandwidth-consumed network services such as video conference,interactive game,cloud computing and so on,there exists a severe demand on the transmission capacity for the optical fiber backbone network,which is assumed to be the core of national network infrastructure development and construction.Therefore,it is essential to increase the transmission capacity for the optical backbone network.Given the limited spectrum resource,spectral efficiency(SE)must be a vital factor in designing future optical netwoks.In order to overcome the problem of limited bandwidths originated from optical and electrical components when increasing the single channel baudrate,with the help of coherent detection,the superchannel technique can greatly increases the SE so as to meet the increasing demand on capacity.Therefore,it has been extensively investigated recently.Moreover,future networks must be capable of adapting to different bandwidth requirements in order to conduct dynamic bandwidth allocation.As for superchannel,by manipulating the subchannel baudrate,the use of modulation formats and the number of subchannels,it can achieve variable bit rates transmissions and meet the requirements for future agile networks.In this thesis,we aim to build up superchannel transmission systems and investigate corresponding digital signal processing techniques.The inovative contributions of this thesis are summarized as follows:(1)We propose to use advanced modulation format offset-16 QAM and receiver-side digital spectral shaping(RS-DSS)for the purpose of building up a Nyquist-WDM superchannel.Considering 7% hard decision forward error correction(HD-FEC)overhead,a spectral efficiency up to 7.44 bit/s/Hz can be obtained.Compared with the case using 16 QAM and RS-DSS,1.4 dB improvement of required OSNR at BER = 10-3 udner the scenario of back-to-back(B2B)transmission is obtained.Moreover,over a 960 km SSMF transmission link with EDFA-only,the launched power from-6 dBm to 1.2 dBm is allowed for our proposed system considering HD-FEC threshold.In the end,less than 1.8 dB required OSNR penalty is secured even with the phase difference between channels varying from 0 to 360 degree for the proposed system.(2)We comprehensively investigate the performance of Nyquist-WDM superchannel with transmitter-side digital spectral shaping(TS-DSS),using 16 QAM and offset-16 QAM modulation formats,respectively.Since the tap number of 35 and 81 is essencial for the required finite impulse response(FIR)filter using offset-16 QAM and 16 QAM,respectively,using offset-16 QAM can greatly relax the transmitter-side implementation complexity.Moreover,about 1.3 dB improvement of required OSNR at BER = 10-3 udner the scenario of back-to-back(B2B)transmission is obtained using offset-16 QAM.In particular,using offset-16 QAM can increase the tolerance of DACs' constraints,including the untracked jitter,sampling rate and resolution.(3)We propose a joint digital signal processing(DSP)technique for 28 GBaud 16 QAM Nyquist-WDM superchannel.Along with the use of 3 channel MIMO equalizer,digital spectral shaping(DSS)together with corresponding maximum likelihood sequence detection is implemented.Compared with traditional joint DSP with only 3 channel MIMO equalizer,we can obtain 1.8-dB required-OSNR improvement at BER of 10-3 for the back-to-back(B2B)measurement.Moreover,0.9-dB Q2-factor improvement is also secured after 960-km SSMF transmission with EDFA-only,considering the optimal launched power.In the end,using the proposed joint DSP technique,linewidth tolerance is increased from 80 to 230 kHz given 1-dB required-OSNR penalty.(4)We propose a low complexity feed-forward carrier phase recovery(CPR)scheme.Firstly,the quadratic relationship between the test angle and corresponding distance matric is theoretically obtained during the BPS implementation.Afterwards,we propose a CPR based on a two-stage BPS and quadratic approximation(QA).Since QA can significantly accelerate the speed of phase searching using the quadratic relationship,a group factor of 2.96/3.05,4.55/4.67 and 2.27/2.3(in the form of multipliers/adders)reduction of computation complexity(CC)is achieved for 16 QAM,64QAM and 256 QAM,respectively,compared with the traditional single-stage BPS scheme.Meanwhile,a guideline for determining the summing filter block length is put forward during performance optimization.Under the condition of optimal filter block length,our proposed scheme shows same performance as traditional single-stage BPS scheme.(5)We propose the design of a hitless coherent transceiver with the function of modulation format identification(MFI)for flexible optical networks.In particular,the block-wise based decision-directed least-mean-square(DD-LMS)equalizer and the pilot symbol aided superscalar phase locked loop(PLL)are used for channel tracking and CPR,respectively.For the purpose of MFI,the modulation format information is encoded onto BPSK pilot symbols,which are initially used for the superscalar CPR.Therefore,the proposed MFI does not result in extra overhead.In both simulations and experiments with QPSK,16 QAM,64QAM,Hybrid QPSK/8QAM,and set-partitioning(SP)-512-QAM,we verify that the proposed MFI method induces no performance penalty.Furthermore,our proposed hitless transceiver can support fast block by block modulation format switching.This is demonstrated in a 6400 km SSMF transmission experiment where we show format switching between QPSK and Hybrid QPSK/8QAM,and in a 1600 km SSMF transmission experiment with format switching between 16 QAM and SP-128-QAM.Finally,we evaluate the performance of the proposed MFI.The results show that the proposed MFI covers a wider range of modulation formats and is more reliable especially at low OSNR.