Channel Capacity Analysis Of Visible Light Communication System

Visible light communication(VLC)refers to the use of visible light at 430-790THz as a carrier to transport information,because there is no need of spectrum license,can take into account lighting,and have no electromagnetic pollution,and become one of the candidate schemes of 6G indoor commnuication.Because VLC uses light emphasis on signals,the estimation of its channel capacity is more difficult than the estimation of traditional wireless channel capacity.This thesis analyzes and estimates its channel capacity for SISO-VLC plus Gaussian noise channels under the constraints with and without average light power constraints,and MIMO-VLC plus Gaussian channels.The main work is as follows:1.The channel capacity problem of SISO-VLC system without the constraint of average optical power is studied.The main work is as follows:1)Under the assumption that the input source is discretely distributed,the mathematical optimization problem corresponding to the capacity problem is given;2)When the number of discrete points N is the optimization variable,the following three conclusions are proved:(1)The best matching source must satisfy that the average mutual information of each symbol to the output signal is equal;(2)The best matching source must satisfy the distribution symmetry;(3)The boundary of the signal interval must be covered by constellation points;3)When the number of discrete points is determined,the following conclusions are obtained:(1)When N is less than the optimal number of points N~*,the conclusion is similar to 2);(2)When N is greater thanN~*,there areN-N~*points where the probability tends to 0.Finally,the above conclusions are verified by simulation.2.The channel capacity problem of SISO-VLC system with the constraint of average optical power is studied.The main work is as follows:1)Under the assumption that the input source is discretely distributed,the mathematical optimization problem corresponding to the capacity problem is given;2)When the number of discrete points N is the optimization variable,the following four conclusions are proved:(1)The best matching source no longer satisfy that the average mutual information of each symbol to the output signal is equal,the mutual information should be related to the power of the input symbol;(2)The same channel capacity can be obtained under the condition that the dimming coefficients are and 1-,and the optimal constellation points are symmetrical about A/2;(3)The optimal constellation points can no longer cover the start and end points of the input signal amplitude range;(4)Under the condition that the dimming coefficient is 0.5,a larger capacity can be obtained than other dimming coefficients;3)When the number of discrete points is determined,the following conclusions are obtained:(1)When N is less than the optimal number of points N~*,the conclusion is similar to 2);(2)When N is greater than N~*,there areN-N~*points where the probability tends to 0.Finally,the above conclusions are verified by simulation.3.the channel capacity of MIMO-VLC system is studied.The main work is as follows:1)Through the SVD theory,the coupled system is decomposed into multiple equivalent independent SISO systems,and the corresponding optimization problem of the equivalent system is given;2)The feasible region change of the system after SVD is explained,and the algorithm for solving the capacity sum is given;3)For the problem of the upper and lower bounds of the capacity,an idea is given through the combination of numbers and shapes.Finally,We simulate the 2×2 system and give the result under the condition that the channel matrix is a real symmetric matrix,and the optimal constellation point design in this situation is given too.The simulation results show that:(1)When the channel matrix correlation is reduced,a larger channel capacity can be obtained;(2)In the optimal constellation diagram,the probability of the constellation points on both sides is always greater than the probability of the constellation point in the middle.