Research On Channel Modeling And Performance Optimization Of Visible Light Communication

Visible light communication(VLC),as an promising,efficient and secure wireless communication technology,can not only effectively alleviate the current shortage of wireless spectrum resources,but also meet the demand for communication rates in the era of high-speed data information.However,there are still some problems in the VLC field that need to be further resolved,which mainly includes:lack of basic research on channel modeling of VLC in different scenarios,how to use other key technologies to optimize the performance of VLC systems,the performance of traditional channel estimation methods is limited and etc.Therefore,in order to realize a more efficient and reliable VLC system,this paper mainly starts from the visible light channel modeling,and summarizes and analyzes the channel model,channel characteristics and performance indexes of the VLC system.Then optimize the performance of the VLC system from two aspects:channel coding and channel estimation.On the one hand,in order to improve the spectrum efficiency and transmission delay of the VLC system,non-orthogonal multiple access(NOMA),which is more suitable for a small number of receivers,is adopted as the access technology of the system.In addition,in order to improve the reliability of the VLC system,polar code with better error correction performance and lower coding complexity is adopted as the channel coding scheme of the system.Therefore,the NOMA-enabled VLC system model based on polar code is proposed.Moreover,the architecture of the transmitter and receiver of the system is designed,and a joint decoding algorithm based on successive cancellation and successive interference cancellation is proposed at the receiver to decode the received signal.Finally,the simulation results demonstrate that the NOMA-enabled VLC system based on polar codes not only has better error correction performance than the system based on turbo codes,but also the overall complexity of the system is lower.On the other hand,in order to further improve the communication capacity of the VLC system,combined with massive multiple-input multiple-output(MIMO)technology,by deploying light-emitting diodes(LEDs)and photodiodes(PDs)in the form of arrays at the transmitter and receiver,a massive MIMO visible light channel model based on LEDs and PDs array is established.In order to achieve a more efficient and reliable massive MIMO VLC system,it is essential to ensure accurate channel estimation.However,due to the limited estimation performance of traditional channel estimation methods in massive MIMO scenarios,this paper utilizes the sparse characteristics of massive MIMO visible light channel to treat the channel matrix as a two-dimensional natural image.Then,through the method of image denoising in the deep learning,a fast and flexible denoising convolutional neural network(FFDNet)is designed and trained to estimate the channel.Finally,the simulation results show that the channel estimation method based on FFDNet can obtain better estimation performance than the traditional channel estimation methods.