Research On Robust Transmission Technology Of Visible Light Communications

With the explosive growth of access devices and data traffic in 5G,radio frequency communication is facing an increasingly serious crisis of lack of spectrum resources.Due to its advantages of rich spectrum resources and electromagnetic immunity,visible light communication(VLC)has attracted the research attention of academia and industry in recent years,and is regarded as one of the key technologies of B5G/6G.However,in practice,due to the existence of quantization errors,channel estimation errors,and pointing errors of mobile VLC,channel state information(CSI)is always imperfect,and CSI errors easily lead to the deterioration or even interruption of VLC performance.To make up for the defect that VLC is sensitive to CSI errors in practice,this thesis mainly investigates robust transmission technology of VLC.The specific research contents and innovations are as follows1.Aiming at the CSI errors caused by quantization errors and channel estimation errors in multi-user scenarios,this thesis proposes robust power allocation algorithms of Rate Splitting Multiple Access(RSMA)-VLC system.Specifically,by analyzing the signal processing process at the transmitter and receiver,this thesis first derives the lower bound of the achievable rate closed expression for RSMA-VLC system.On this basis,considering the practical power constraints of LEDs and service quality of users,the transmit power minimization algorithms are respectively designed under perfect CSI and imperfect CSI based on semidefinite relaxation and constrained convex concave programming.Numerical results show that the proposed algorithm outperforms non-Orthogonal nultiple access,spatial division multiple access and non-robust algorithms in terms of power consumption and robustness.2.Aiming at the CSI errors caused by receiver pointing errors in mobile user scenario,this thesis proposes robust power allocation algorithms of integrated visible light communication and position(VLCP)system.Specifically,by designing an integrated frame structure of communication and positioning,real-time value of CSI is estimated by using positioning information.Then,the Cramer-Rao lower bound(CRLB)of positioning error and the lower bound expression of communication rate are respectively derived,and the relationship between communication power and positioning power is established.On this basis,considering the practical power constraints of LEDs and the CRLB threshold,the problem of maximizing the average communication rate of integrated VLCP system under imperfect CSI is studied,and two power allocation algorithms based on deep reinforcement learning are proposed.Numerical results show that the proposed algorithm outperforms the equal power allocation algorithm in terms of average communication rate and robustness.This thesis contains 24 figures,8 tables and 100 references in total.