OFDM Optical Mimo Visible Light Communication Using ML DL Approach

Abstract

ABSTRACT newlineVisible Light Communication (VLC) is emerging as a promising technology to offer newlineand support ubiquitous communication. In this thesis, the performance Of Orthogonal newlineFrequency Division Multiplexing (OFDM)-based optical-Multiple Input Multiple Output newline(MIMO) VLC system is analyzed. The MIMO VLC system with 4 × 4 configuration is newlinedeveloped with four different transmission techniques namely, Repetition Code (RC), Spatial newlineMultiplexing (SMP), Spatial Modulation (SM), and Generalized SM. The performance of all newlinethese schemes are analyzed. Furthermore, OFDM with MIMO for VLC system is developed newlineand analyzed. Additionally, flip-OFDM is used as a unipolar technique to support VLC for newlinemany applications like multiuser scenario. The performance of their two classes, namely, newlineAsymmetrically Clipped Optical OFDM (ACO-OFDM) and Direct-Current Biased Optical newlineOFDM (DCO-OFDM) is compared. Extending the work towards multiuser VLC system, a newlinecomparative analysis for RC, SM, generalized SM and SMP is analyzed with and without the newlineuse of flip-OFDM. To enhance the multiuser performance in VLC system, Non-Orthogonal newlineMultiple Access (NOMA) model is developed and analyzed. Also, Machine Learning newline(ML)/Deep Learning (DL) model developed for VLC system. Investigation is extended on newlineVLC receiver using Convolution Neural Network (CNN), Long Short Term Memory (LSTM), newlineand Deep Neural Network (DNN) algorithms and analyzed the performance. NOMA uses newlinesuperposition in power domain at the transmitter and Successive Interference Cancellation newline(SIC) at the receiver. SIC operation is expected to perform perfect cancellation to avoid errors newlinein the received signal. NN methods are used to overcome imperfect SIC in a NOMA VLC newlinesystem. These trained and optimized algorithms are also used to simulate and analyze the newlineperformance of flip OFDM based VLC system. The VLC system using conventional flip- newlineOFDM is compared with the VLC system using ML/DL based receiver. newlineAll the simulation in this work is performed using MATLAB (2017b and 2021a) tool and newlineresults obtained has shown the merits of our approach. newlineIt is found that overall performance has increased. While flip-OFDM using generalized newlineSM outperforms DCO-OFDM by 2dB for a BER of 10-6, NOMA achieves average rate greater newlinethan 10Mbps. It is also observed that NN methods outperforms the conventional NOMA VLC newlinesystem to a perfect SIC. Thus, NN-based receiver using ML/DL approach will be a better newlinealternative for imperfect SIC. newlineKeywords: Visible Light Communication (VLC), Orthogonal Frequency Division newlineMultiplexing (OFDM), Multiple Input Multiple Output (MIMO),Non-Orthogonal Multiple newlineAccess (NOMA), Repetition Code (RC), Spatial Multiplexing (SMP), Spatial Modulation newline(SM), Generalized SM, Machine Learning (ML), Deep Learning (DL), Convolution Neural newlineNetwork (CNN), Long Short Term Memory (LSTM), and Deep Neural Network (DNN). newline