Power Division Multiple Access (PDMA) System Performance Analysis And Application Design

Power division multiple access(PDMA),which exploits the channel gain difference among users,allocates unequal power to the different users at the transmitter and adopts successive interference cancellation(SIC)at the receiver,is a promising non-orthogonal multiple access scheme for improving the system capacity and enhancing the user fairness.This paper investigates the performance gain of PDMA in the downlink wireless communication,and proposes solutions to the application of the PDMA technology.The main research achievements are as follows:Firstly,the capacity gain over the OMA scheme is theoretically analyzed and numerically evaluated in the SISO/MIMO/massive MIMO scenarios.After establishing a unified channel model,the paper derives the distribution of the users' large-scale fading and received signal to interference and noise ratio,and obtains the pairing proportion of the PDMA scheme.On this basis,the paper evaluates the throughput gain over the OMA scheme.Next,the novel user pairing schemes,named "grouping-random pairing searching","grouping-fairness criteria" and "grouping-channel correlation maximization criteria",are proposed,which take the large-scale fading and channel correlation among users into consideration.Among them,the first scheme strikes a balance between fairness and effectiveness with the least complexity,the second focuses on fairness,and the last enjoys the best effectiveness with the most complexity.Thirdly,for the power allocation scheme,in order to exploit the channel fading characteristic of the paired users,the paper puts forward the "no loss of the single users' capacity" criteria,which means for each of the PDMA paired users,the capacity is not less than that of the OMA scheme.The value range of the power allocation is calculated by the aforementioned criteria,and the specific value can be selected according to the system requirements for the effectiveness and fairness.At last,for the SIC algorithm,three implementation methods are presented.The first is symbol level IC(SLIC),which means no decoding operation in the interference estimation process;the second is codeword level "hard" IC,which decodes the interference to the bit level,reconstructs it as the constellation symbol,and then removes it;and the last one is codeword level "soft" IC,which decodes the interference to the logarithmic likelihood-ratio,reconstructs it as the"soft" symbol,and then removes it.The paper also gives the link abstraction method to estimate the block error rate of the last SIC algorithm.In this thesis,the methods to the system performance analysis are worthy for consideration,and the simulation results reveal that the application design for PDMA can achieve better performance in terms of the system fairness or effectiveness.