Font Size: a A A

The Performance Optimization Techniques Of Coherent Detection Based Optical OFDM Transmission System

Posted on:2019-10-06Degree:DoctorType:Dissertation
Country:ChinaCandidate:L LiuFull Text:PDF
GTID:1368330590970404Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
With the continuous development of Internet technology,bandwidth-hungry services such as cloud services,virtual/augmented reality,network video surveillance,and live video streaming are emerging.The demand for global communication capacity has exploded.As the cornerstone of modern communication network,the optical communication system has the urgent need to increase its transmission capacity.The optical orthogonal frequency division multiplexing(OFDM)transmission system based on coherent detection has been selected as one of the effective solutions for optical transmission systems due to its advantages of high spectral efficiency,flexible modulation format,flexible bandwidth allocation,robustness to dispersion,convenient digital signal processing(DSP)for impairment compensation,high received sensitivity and long relay distance.This dissertation focuses on research of the performance optimization techniques for coherent detection based optical OFDM(CO-OFDM)transmission system.In terms of the implementation methods,the performance optimization techniques can be divided into DSP post-processing based performance optimization technique and channel coding based performance optimization technique.The former one compensates the impairment-deteriorated signal at the receiver(RX)by using suitable DSP algorithms for different system impairments,while the latter one uses the redundant information added at the transmitter(TX)to correct the error,improving the robustness to system impairment by providing additional coding gain.These two techniques can work together to improve the performance of high-speed long-distance CO-OFDM system.Aimed at these key techniques,researches are carried out as follows: 1.The research on the principle and impairments of CO-OFDM systemFirst,from the theoretical point of view,this dissertation theoretically and simulative analyzes the performance and the impairment of CO-OFDM system,which lay the foundation for studying its performance optimization techniques.(1)The simulation and experimental platforms used in this dissertation is built according to the principle and model of CO-OFDM system.(2)The impacts of system main impairments,namely the channel linear impairment,laser phase noise and fiber nonlinearity,on the OFDM signal are deduced by theory.(3)The necessity of studying the above impairments is verified by comparing the tolerance of coherent detection based and direct detection based optical OFDM transmission systems to channel linear impairment,laser phase noise and fiber nonlinearity by simulation.2.The DSP post-processing performance optimization technique in CO-OFDM systemTo compensate the three main system impairments,the channel linear impairment,the laser phase noise and the fiber nonlinearity,this dissertation proposes the corresponding DSP post-processing performance optimization algorithms respectively.The experiments and simulations are carried out to verify the performance.Also,comparison with the existing algorithms are made.(1)For the channel linear impairment,firstly,the affine projection based sign modified constant modulus algorithm(AP-SMCMA)is proposed.To make up for the shortcomings of the existing modified constant modulus algorithm(MCMA),APSMCMA introduces the signum function to reduce the complexity when calculating the error function,and introduces the affine projection(AP)algorithm to improve the convergence speed when updating the weight.The experiment and comparison are conducted in single-channel 10 Gb/s system.Then,in order to further enhance the the anti-noise ability of conventional OFDM structure which is limited by its sidelobe suppression ratio(SSR),we use the filter bank-based OFDM structure to improve the signal's robustness to noise.For the improved structure,the interference approximation-comb pilot based frequency domain intra-symbol averaging and intrasubcarrier linear interpolation(IA-CFAL)algorithm to compensate channel linear impairment is proposed with higher accuracy than the existing interference approximation(IA)algorithm and lower overhead than the existing IA-block pilot based frequency domain intra-symbol averaging(IAFA)algorithm.The experiment and comparison are conducted in single-channel 20 Gb/s system.(2)For the laser phase noise,the Lanrange interpolated extended Kalman filter(LRI-EKF)for laser phase noise compensation is proposed.By analyzing the characteristics of laser phase noise,a more suitable Lagrange interpolation than the existing linear interpolation(LI)is used to fit inter-carrier interference(ICI),and a modified iteration-free blind(MIFB)algorithm is used to estimate common phase error(CPE),which can improve the accuracy of phase noise estimation and system spectral efficiency.The simulation and comparison are conducted in single-channel 25 Gb/s system.(3)For the fiber nonlinearity,the radial basis function neural network(RBF-NN)algorithm for fiber nonlinear compensation is proposed.Using the idea of machine learning,RBF-NN nonlinear equalizer with faster convergence rate and lower complexity than the existing multi-layer perceptron neural network(MLP-NN)is adopted.This equalizer uses lower-complexity easier-operation orthogonal least square(OLS)learning strategy compared with the existing K-means clustering strategy.This scheme achieves better performance than the existing MLP-NN with lower training overhead,and can simultaneously compensate for channel linear impairment and laser phase noise.The experiment and comparison are conducted in single-channel 40 Gb/s system.3.The channel coding performance optimization technique in CO-OFDM systemBased on the key technology of channel coding performance optimization,the performance of CO-OFDM system is further optimized,especially for the long distance and high speed case.Aiming at the problems of high complexity and large delay of existing low density parity check(LDPC)coding scheme caused by the concatenated outer algebraic code which is used to restrain the inherent shortcoming brought by its iterative decoding,this dissertation introduce the polar code into CO-OFDM system for the first time.As the only soft decision-forward error correction(SD-FEC)that has been proven to achieve the Shannon's capacity so far,polar code has the advantages of explicit structure,strict analysis range of error correction performance,low decoding complexity and error floor-free.This dissertation studies the polar coding performance optimization technique for CO-OFDM systems from the following aspects:(1)The principle of polar coded CO-OFDM system and the suitable coding/decoding algorithms are expounded.The system performance is experimentally verified and compared with the existing LDPC scheme.(2)For the first time,the robustness of the polar code to nonlinear impairment of high speed CO-OFDM system is discussed to comprehensively evaluate its robustness to all system impairments.Theoretical analysis and simulation/experiment are carried out.(3)A robust filter bank-based polar coded CO-OFDM system is proposed for the first time to further optimize the performance of polar coded CO-OFDM system.The simulation and experimental verifications are carried out.
Keywords/Search Tags:optical transmission system, coherent detection, optical orthogonal frequency division multiplexing, digital signal processing, channel coding
PDF Full Text Request
Related items