Research On Key Techniques Of Nonlinear Distortion Mitigation In Visible Light Communication

As an important complementary technology of wireless radio frequency commu-nication,visible light communication(VLC)has many unique advantages.However,VLC is a nonlinear system with limited dynamic range.At the same time,the memory effect of LED and channel is superimposed on LED static nonlinearity to further degrade system performance.In order to improve the communication performance,generally consider increasing the dynamic range of input signal,but this will introduce further nonlinear distortion,therefore we need to study the corresponding nonlinear distortion mitigation technology.The main content of this theisi includes the nonlinear model of VLC and nonlinear distortion effects.At the same time,three kinds of equalizers are designed to mitigate nonlinear distortion.In this work,we propose a Wiener Hammerstein nonlinear model.The static non-linearity and memory effect are considered comprehensively,and the relevant param-eters are determined according to the simulation of the indoor channel.Based on this nonlinear model,the impact of constellation amplitude distortion and phase distortion is investigated.Moreover,the nonlinear harmonic distortion phenomenon in the frequency spectrum of the received signals is verified by an offline organic VLC system.We design three types dynamic range input signals.Volterra series simplification model and Chebyshev polynomial expansion of memory effect are analyzed,and their corresponding nonlinear equalizers are designed based on RLS algorithm.The simu-lation explores the performance under different dynamic range signals,as well as the impact of parameters such as nonlinear term order and training sequence length.At the same time,we verify the effectiveness of the proposed nonlinear equalizer in mit-igating nonlinearity and memory effects by the offline VLC system,which achieve a communication distance of 3m and refresh relevant research records.Moreover,we study the nonlinear equalizer based on artificial neural network(ANN).According to the nonlinear model,the structure of ANN is determined,the network ro-bustness is improved by using multiple repeated training methods,and learning is per-formed using the LM algorithm.The simulation results show that compared to the other two types of nonlinear equalizers,performance gains of two orders of magnitude can be achieved for all dynamic range signals at high SNR region.