| In the past 10 years, we have witnessed the rapid development of the modern wireless communication technologies as well as the tremendous changes caused by it in our daily lives. However, along with the proliferation in the demands for wireless applications, there has an explosive growth in mobile data traffic volume due to the emergence of services like intelligent terminals, internet of things, and multimedia infotainment. To address the challenge of this data explosion, the academia and industry have initiated the research and development on the fifth generation mobile communications(5G). Currently, it has been the consensus that massive MIMO and full-duplex are two main candidate techniques in future 5G networks. Firstly, the massive MIMO can eliminate effectively the small-scale fading and the additive thermal noise when the number of antennas at terminals is large enough. As a result, the spectral and energy efficiencies of systems can be improved greatly. Secondly, under the case where the loop interference is canceled,the double spectral efficiencies can be achieved by the full-duplex systems. In this thesis, we consider a relay system where the massive MIMO and full-duplex are integrated effective. For the full-duplex massive MIMO relay systems the spectral and energy are investigated under the two cases where the amplify-and-forward(AF) and decoder-and-forward(DF) relay protocols are employed, respectively.At the same time, we also investigate the effect of the full-duplex schemes on the loop interference. With the obtained results, the effective power allocation schemes are proposed as well as the corresponding performance comparison.(1) Firstly, based on AF relaying protocol a novel full-duplex massive MIMO scheme are proposed, which consists of multi-user two-hop relay systems. The relay is equipped with massive MIMO antenna array while the user terminals have the single antenna. Due to employing full-duplex mode the massive MIMO antenna array amplifies and forwards the received signals from sources to destinations at the same time-frequency resources. For such full-duplex massive MIMO schemes, we obtain the spectral and energy efficiencies as well as three power scaling power schemes. The result shows that the full-duplex massive MIMO AF relay systems can cancel the loop interference only when the second power scaling scheme is employed so that the significant gains on spectrum and energy efficiencies are achieved.(2) Secondly, We study the full-duplex massive MIMO DF relay systems. By obtaining the closed-form expressions of the spectrum and energy efficiencies, we compare the AF and DF systems in terms of system performance and the effect on loop interference. Our results show that different from AF schemes, the DF schemes can cancel the loop interference under the both cases where the second and third power scaling schemes are employed, respectively. This is due to the fact that in DF schemes the two hops in relay systems are separated while in AF ones the two hops are combined together. Besides this, the AF schemes outperform the DF ones when the gains of uplink are higher than the ones of downlink. On the contrary, we find that gap of performance between AF and DF schemes is small and the effect of loop interference on performance is very little.(3) Finally, to achieve a comprehensive knowledge about the impact of duplex models on loop interference, we present the half-duplex implementation of massive MIMO relay systems and give the comparison in terms of system performance. We find that the full-duplex systems are inferior to the half-duplex ones when the level of loop interference is high, relatively. The reason is that the loop interference is amplified in full-duplex systems. Therefore, based on these considerations, we propose a novel full-half-duplex(or hybrid full-and-half duplex)massive MIMO relay scheme, where the systems work under the full-duplex or half-duplex model according to the system parameters. As a result, the spectrum and energy efficiencies are further improved. |
PDF Full Text Request