Design And Simulated Implementation Of PAPR Reduction Algorithm In CO-OFDM System

Recently,the information society have witnessed the rapid development of heavy flow services,such as multi-media,interactive games and real-time video.Along with the development comes the ever growing demand for transmission network with high speed and large capacity.This particular kind of demand greatly promotes the development of the optical transmission network.Orthogonal Frequency Division Multiplexing(OFDM)has become one of the hot research spot in the communication field due to advantages such as high spectral efficiency and high resistance towards fiber chromatic dispersion.Compared with direct detection,coherent detection has the advantages of high receive sensitivity thus can prolong the transmission distance of optical signal.Combining the merits of coherent detection and OFDM modulation technology,Coherent Optical OFDM(CO-OFDM)has become a promising candidate for optical transmission system with information rate over 100Gbit/s.However,the high Peak-to-average Power Ratio(PAPR)of the OFDM signal has always been one of the inherent defects that limits its performance.The previous works regarding the PAPR reduction of OFDM are mostly confined within the wireless transmission field and focus on the energy efficiency and working range of high power amplifier(HPA).In CO-OFDM transmission system,signal with high PAPR will drive linear devices including mach-zehnder modulator,amplifier and analog-to-digital converter out of their linear range.What's more,high PAPR will cause severe nonlinear effects in fiber channel.The resulting nonlinear distortion and extra nonlinear noise will degrade the transmission performance as well as raise the bit error rate.This thesis analyzes the PAPR character of optical OFDM signal in a CO-OFDM system.The theoretical explanations of the CO-OFDM system architecture and mathematical reasoning of the high PAPR effects are presented.The main work of this thesis is to propose effective power adjustment scheme in order to achieve PAPR reduction.First,we propose an adaptive power adjustment algorithm based on the Least Mean Square(LMS)criterion to reduce the PAPR character of OFDM signal.Then,we propose a variable step-size LMS(VSLMS)to deal with the contradiction between the convergence rate and steady-state error.Furthermore,in view of the bigger influence that the large signal has on system performance,we modify the cost function of LMS to readjust the weight assignment of the signal.Second,we propose an adaptive power adjustment algorithm based on the Recursive Least Squares(RLS)criterion to reduce the PAPR character of the OFDM signal.Thorough analysis of the algorithm including output power and convergence performance are presented.Finally,in consideration of the complementary properties of the LMS and RLS regarding the convergence rate and computation complexity,we further propose a RLS-LMS joint PAPR reduction algorithm.By analyzing the calculation of PAPR reduction algorithms.We prove that the RLS-LMS can achieve ideal trade-off between PAPR reduction and computational complexity.The VPItransmissionmaker and Matlab software are employed to develop the simulation platform in order to evaluate the performance of proposed algorithms.Performance comparisons are made between the proposed algorithms and the previous PAPR reduction approach.Simulation results demonstrate that the proposed PAPR reduction algorithms gain obvious advantages in reducing the PAPR of OFDM signal and improving the BER performance of the transmission system.