Research On Power Allocation Based On Non-Orthogonal Multiple Access In Indoor Visible Light Communications

Visible light communication(VLC)is a promising technology of optical wireless communication for advanced green-oriented networks,which can supplement traditional radio frequency communication.On the one hand,due to the limitation of lighting material manufacture,the practicable system bandwidth of the common VLC hardware components such as the light-emitting diode(LED)might be narrow,which has the influence on the spectral efficiency.On the other hand,the channel interference based on visible light is primarily involved with the optical cell deployment,which may cause the different performances of spectral efficiency.In order to improve the spectral efficiency,this thesis applies non-orthgonal multiple access(NOMA)for VLC networks,and focuses on the relationship between the spectral efficiency and NOMA-VLC resource allocation.Especially,we theoretically discuss the NOMA-VLC power allocation with visible-light channel characteristic and cell deployment in closed-form solutions.The contributions of this paper are as follows.1.Derivation and analysis of probability distribution of channel gain in NOMA-VLC.Bacause the channel gain distribution with cell deployment is a primary factor for the performance of spectral efficiency,this study investigates the probability distribution of visible-light channel gain based on Lambertian radiation and explores the relevant performance property in the indoor scenario.For the single cell,when the user distribution is uniform and the approximate expression based on Taylor series expansion is used,we derive the probability density function(PDF)and cumulative distribution function(CDF)of the channel gain difference based on visible-light channel,and indicate the intrinsic impact of channel gain on the cell deployment and NOMA-VLC spectral efficiency.For the multiple cells,considering the inter-cell interference(ICI)in the overlapping areas,we quantitatively analyze the performance of user association strategies and derive a closed-form sufficient condition to improve spectral efficiency of overlap-area NOMA-VLC users.Afterwards,we perform the strategic analysis of user association in the scenario probability involved with different numbers of users and the probability of the channel gain ratio involved with a user in the overlapping area.Eventually,the theoretical analysis shows the local extremum property of half-power semi-angle and the aforementioned performance probability in the cell deployment with NOMA-VLC,and the numerical results show that the influence of visible-light channel gain on the spectral efficiency of NOMA-VLC is consistent with the theoretical analysis.2.A NOMA-VLC power allocation algorithm based on gain ratio power allocation(GRPA)constraints.Considering the multi-user interference(MUI)in the single cell,this study proposes a revised GRPA strategy with modified power counting and channel proportions to enhance the spectral efficiency of NOMA-VLC.First,owing to difficulties of directly analyzing NOMA-VLC spectral efficiency,an alternative lower bound based on visible-light channel model to facilitate the performance comparisons is proposed.It is proved in a certain and proper condition that the proposed alternative lower bound is asymptotic and compact with respect to NOMA-VLC spectral efficiency.Second,when the numbers of users are 2 and 3,we prove that the spectral efficiency of our proposed GRPA strategy based on visible-light channels is better than that in the previous study.When the number of users is greater than 3,numerical simulation analysis shows the advantage of our proposed strategy.Third,for the case of LED heights greater than 1 meter,a local extremum solution using the proposed strategy is analytically demonstrated based on the alternative bound.For the case of LED heights smaller than 1 meter,the necessary but not sufficient conditions under which the proposed strategy outperforms the previous study is provided in VLC scenarios.Finally,the theoretical analysis shows the local extremum property of half-power semi-angle for maximizing NOMA-VLC spectral efficiency,and the numerical results show that the proposed alternative lower bound can fit the original performance target and the advantage of the proposed strategy with the analytical conditions is demonstrated.3.A NOMA-VLC power allocation algorithm based on overlapping area constraints.Considering the ICI in the overlapping area of multiple cells,this study presents a further theoretical derivation in NOMA-VLC spectral efficiency,based on the assumption of previous studies.First,we arrange the performance evaluation of spectral efficiency for overlapping-cell and non-overlapping-cell users.Based on visible-light channels,we analyze the performance gap between the alternative bound and the original performance target.Second,based on illumination parameters,we derive the compact sufficient condition and relaxed necessary condition to achieve the upper bound for measuring VLC multi-cell spectral efficiency.Third,we derive the power allocation strategy based on the aforementioned bound and conditions,and analyze the influence of power allocation on the cell coverage.Finally,numerical results show that the constructed bound has a favorable trend for fitting spectral efficiency,and demonstrate the closed-form derivation including the compact sufficient condition and the relaxed necessary conditions.The relationship between spectral efficiency and cell deployment is revealed by the theoretical analysis,which is consistent with the numerical results.4.A NOMA-VLC power allocation algorithm based on quality of service(QoS)constraints.Considering MUI and ICI,this study explores the relationship between the enhancement of spectral efficiency and the guarantee of QoS.Especially,in QoS constrains,we primarily discuss the dynamic priority constraint and the max-min achievable data rate constraint.For the dynamic priority constraint,we propose a NOMA-VLC power allocation strategy based on fuzzy logic algorithm to achieve the QoS satisfaction.We introduce fuzzy logic to flexibly analyze user priority.Afterwards,we derive the PDF of NOMA-VLC outage probability and assign signal power using multi-dimensional user features.Finally,the numerical results verify that the previous study with static user priority cannot guarantee the QoS satisfaction,and our proposed strategy indicates the advantages of fairness and satisfaction.For the max-min achievable data rate constraint,we introduce the rate-splitting multiple access(RSMA)method into multiple cells of VLC and analyze the performance gain of RSMA better than the power-domain NOMA.Subsequently,we develop a linear precoding containing 1-layer RSMA-VLC to achieve spectral efficiency having constrained the non-linear effect of an LED,and the transmitted power allocation is constrained by the optical signal power.Furthermore,we optimize the system spectral efficiency with the guarantee of minimum common stream data rate.Finally,numerical results show the effectiveness of our proposal in system spectral efficiency and VLC constraints,and RSMA as a more flexible interference management scheme generalizes and outperforms NOMA.