Research On Channel Estimation Algorithm Of Visible Light OFDM System Based On Compressed Sensing

In view of the problem that the bit error rate of visible light communication(VLC)is very high under non line of sight(NLOS)link,this paper uses channel estimation technology to reduce the bit error rate of system communication.Channel estimation is a key link to ensure the reliability of the communication system.The purpose of channel estimation is to get the impulse response of the transmission channel,so that the transmitted signal can be accurately demodulated at the receiving end,so as to improve the performance of communication.Since the bipolar signals modulated by traditional Orthogonal Frequency Division Multiplexing(OFDM)systems are not suitable for visible light communication systems,Asymmetrically Clipped Optical Orthogonal Frequency Division Multiplexing(ACO-OFDM)and DC Biased Optical Orthogonal Frequency Division Multiplexing(DCO-OFDM)are usually used.This paper will use ACO-OFDM modulation.Due to the particularity of ACO-OFDM subcarrier location,traditional channel estimation algorithms such as least square method need a large number of pilots as the premise to ensure that the system communication has a low bit error rate,which is not suitable for ACO-OFDM system.In this paper,the reconstruction algorithm based o-n compressed sensing is used to estimate the channel of ACO-OFDM system.In this paper,firstly,the model of indoor visible light in NLOS is built,and the radiation mode of indoor visible light and the cause of multipath effect are introduced;Then,the theoretical basis of CS is introduced.At the same time,the time domain signal of VLC is sparsely represented,and the channel estimation is combined with the signal reconstruction algorithm of CS.Secondly,this paper proposes a Modified-Tree-based Stochastic Search(M-TSS)algorithm to optimize the pilot position index position,so that the BER of VLC system will be further reduced when the number of pilots is lower.Secondly,a channel estimation algorithm based on Modified-Sparsity Adaptive Multipath Matching Pursuit(M-SAMMP)is proposed to solve the problem of high communication error rate without channel estimation under NLOS link.This algorithm combines the advantages of Sparsity Adaptive Matching Pursuit(SAMP)and Multipath Matching Pursuit(MMP),implements a Sparsity Adaptive channel estimation algorithm with multiple candidate sets,and uses a threshold strategy to clip the sub-tree to reduce the complexity of the algorithm.Simulation results show that when the pilot frequency is 16 and SNR is 30dB,the bit error rate of the algorithm is 2.70×10^-3.When the pilot frequency is 32 and SNR is 30dB,the bit error rate of the algorithm is 3.23×10^-4.Compared with MMP-BF algorithm,the computational complexity is reduced by about 4 times,and the bit error rate and computational complexity are greatly reduced.Finally,the system is simulated by combining the two improved algorithms.The simulation results show that the error rate of M-SAMMP algorithm is still lower than that of MMP and SAMP.This paper summarizes the research content and points out the next research direction.