Postprocessing Technology Based On SVM For Linearizaition Of Microwave Photonic Link

Microwave photonics combines photonic technology with the microwave system.By exploiting the low loss,light weight,large bandwidth and electromagnetic isolation of optical fiber,it can fulfill many powerful functions which are usually tough or even impossible in microwave or electrical domain.Microwave photonics system has been widely used in signal processing,civil communications,military defense,aerospace and other fields.A detailed theoretical analysis on the performance of microwave photonics link is given in this thesis.And the nonlinear phenomena due to the baseband data-dependent modulation is also analyzed both in simulation and experiment.An M-ary SVM algorithm based linearization post-processing scheme is proposed,which can effectively compensate the system nonlinear impairment and increase the system dynamic range.The works in this dissertation include:1.Theoretical and simulative analysis of microwave photonic link performanceIn this paper,the performance index of microwave photonic link is discussed.After giving the definition of the parameters including link noise,link nonlinearity,link dynamic range,the theoretical derivation and simulation analysis are carried out.From these analysis results,the general principle and performance of microwave photonic link can be understood more clearly.The basic concept of vector signal modulation in microwave photonic link is also introduced in this paper.2.Analysis of data-dependent nonlinear phenomena for IMDD linkWan Jing et al.from Georgia Institute of Technology for the first time investigated the data-dependent baseband cross modulation(XM)in multiband RoF systems.They consider the nonlinear modulation produced by the crosstalk between different channels of data modulation and analyzed the nonlinear phenomenon of signal eye-diagram and constellation changes caused by XM both in simulation and experiment.Then a general conclusion is deduced.In this thesis,we take a similar analysis on the intensity modulation direct detection(IMDD)link and make a similar conclusion.3 linearization postprocessing scheme based on SVM M-aryAn optimized decision detection method based on SVM M-ary algorithm is proposed for optimizing the received signal constellation.This is the innovation point of this paper.In this method,the distortion vector signal is optimally decided and recovered by the machine learning SVM classification scheme.It can greatly reduce the system bit error rate(BER)with no specific link parameters is required.Comparing with the traditional method of base band decision and recovery,it can increase the operable power range by 3dB in the high nonlinear regime.