Research On MIMO Equalization Technology In Space Division Multiplexing Transmission System

The single-mode transmission capacity is affected by the nonlinear effect of fiber,and there is a bottleneck in the continuous improvement of transmission capacity.At the same time,the increasing demand for transmission capacity makes space division multiplexing technology one of the important directions for the future development of long-distance large-capacity transmission.At present,there are two main types of optical fibers used by researchers to achieve space division multiplexing,namely,few-mode fiber(FMF)and multi-core fiber(MCF).In theory,FMF uses the orthogonality between different modes to increase the transmission capacity exponentially.However,in the actual process,due to the imperfect fiber manufacturing process and other reasons,the orthogonal relationship between modes is destroyed,and then the coupling between modes occurs.In addition,the difference in group velocity between modes leads to the existence of group delay,that is,different pulses interfere with each other to form inter-symbol interference(ISI).The performance impairment caused by mode coupling and group delay requires multiple-input multiple-output(MIMO)equalization at the receiver to compensate.MCF improves the transmission capacity by increasing the number of cores in the cladding,but it is also limited by the manufacturing process,resulting in inter-core delay.At the same time,it is also limited by the diameter of the cladding,resulting in inter-core coupling,and MIMO equalization is also needed at the receiving end.In this thesis,the MIMO equalization technology in space division multiplexing transmission system is studied,and the methods to optimize the complexity and performance of MIMO equalization are explored.This thesis is mainly divided into two aspects.On the one hand,based on the traditional time domain equalization(TDE)and frequency domain equalization(FDE)schemes,different kinds of adaptive algorithms are introduced,and they are applied to the simulation system dealing with 6×6 FMF transmission 200 km and 8×8 MCF transmission 1201 km.The offline experimental system compares and analyzes the differences in convergence speed,performance and complexity of different algorithms.Simulation results show that compared with the frequency domain least mean square(LMS)algorithm commonly used in MIMO equalization,the convergence speed of the frequency domain improved proportional normalized least mean square(IPNLMS)algorithm is about 3 times faster,which reduces the complexity of the training phase by61.51%.The experimental results show that the frequency domain IPNLMS algorithm has better bit error rate(BER)performance than the frequency domain LMS algorithm when the training sequence is short.On the other hand,an improved cyclic prefix(CP)scheme is proposed based on single carrier cyclic prefix(SCCP).Firstly,it and the traditional FDE scheme are applied to the simulation system of 10×10 FMF transmission 80 km,and the computational complexity of the two schemes under the independent equalization structure is compared.The simulation results show that the computational complexity of the improved CP is reduced to 15.53% of the traditional FDE scheme at the expense of 11.11% spectral efficiency.Secondly,both schemes are applied to a simulation system that processes 6×6 FMF transmission for 10 km,and the computational complexity of the two schemes under the joint equalization structure is compared.Simulation results show that the improved CP scheme not only achieves better BER performance at the expense of 20% and 11.11% spectral efficiency,but also reduces the computational complexity to 10.26% and 10.83% of the traditional FDE scheme.