Full-Duplex Massive MIMO Heterogeneous Network Backhaul Scheme By Exploiting Non-Orthogonal Multiple Access

With the rapid popularization of smart mobile devices such as smart-phones and the continuous emergence of new types of services,wireless networks are facing huge challenges in the exponential growth of data traffic.In order to meet higher network performance requirements,heterogeneous network(HetNet),massive multiple-input multipleoutput(MIMO),full-duplex(FD)and non-orthogonal multiple access(NOMA)have become the popular choice for constructing fifth generation(5G)networks.However,the denseness of HetNet has led to heavy backhaul problems.Therefore,this thesis considers a 5G heterogeneous network wireless backhaul scheme and implementation combining FD and NOMA technologies.Firstly,by investigating the limitations of wired backhaul and the disadvantages of wireless backhaul occupying more spectrum in the existing research,this thesis combines with the advantages of NOMA technology to superimpose the power domain so as not to occupy additional spectrum and constructs a full-duplex massive MIMO-enabled multitier HetNet network model to implement wireless backhaul.The densely deployed single antenna small cell base stations(SBSs)cover macro cell.NOMA technology is used on SBS to superimpose backhaul on the uplink/downlink according to the power allocation coefficient.Macro base stations(MBSs)are equipped with a massive MIMO antenna array.SBS and MBS both work in FD mode.All transmissions on MBS and SBS execute on the same time-frequency resources.Secondly,for the uplink(UL)backhaul network model constructed,this dissertation uses stochastic geometry to evaluate the coverage of UL backhaul transmission of a small cell.On the one hand,the distance between two MUs is approximated as the distance between one MU and an access point(SBS or MBS)that interferes with the MU.Meanwhile,the statistical characteristics of the corresponding distance are obtained.On the other hand,the approximately homogeneous distributed Poisson point process(PPP)is used to derive the statistical description of the signal to interference ratio(SIR)and corresponding coverage performance of the small cell UL and UL backhaul transmission.Simulation results show that the NOMA power allocation coefficient has a large impact on the coverage probability of the UL backhaul transmission in small cells.In addition,the transmit power of the MU also has the different effect on the coverage probability of the UL transmission and the UL backhaul transmission.Finally,in order to further estimate the HetNet performance,this dissertation uses stochastic geometry and then gains the coverage probability of DL transmission in small cells for the constructed downlink(DL)backhaul network model.We gain the statistical description of the downlink SIR and the corresponding coverage by using the user association rule based on the nearest distance.The simulation results verify that the integration of FD and NOMA on SBS is an effective scheme to solve the network backhaul problem and the reasonable power allocation coefficient is critical to the coverage of the network.