Master of Engineering Science
Electrical and Computer Engineering
The explosive proliferation of smart devices in the 5-th generation (5G) network expects 1,000-fold capacity enhancement, leading to the urgent need of highly resource-efficient technologies. Non-orthogonal multiple access (NOMA), a promising spectral efficient technology for 5G to serve multiple users concurrently, can be combined with massive multiple input multiple output (MIMO) and relaying technology, to achieve highly efficient communications. Hence, this thesis studies the design and resource allocation of NOMA-based massive MIMO and relaying systems.
Due to hardware constraints and channel condition variation, the first topic of the thesis develops efficient antenna selection and user scheduling algorithms for sum rate maximization in two MIMO-NOMA scenarios. In the single-band scenario, the proposed algorithm improves antenna search efficiency by limiting the candidate antennas to those are beneficial to the relevant users. In the multi-band scenario, the proposed algorithm selects the antennas and users with the highest contribution total channel gain. Numerical results show that our proposed algorithms achieve similar performance to other algorithms with reduced complexity.
The second part of the thesis proposes the relaying and power allocation scheme for the NOMA-assisted relaying system to serve multiple cell-edge users. The relay node decodes its own message from the source NOMA signal and transmits the remaining part of signal to cell-edge users. The power allocation scheme is developed by minimizing the system outage probability. To further evaluate the system performance, the ergodic capacity is approximated by analyzing the interference at cell-edge users. Numerical results proves the performance improvement of the proposed system over conventional orthogonal multiple access mechanism.
Liu, Xin, "Highly Efficient Resource Allocation Techniques in 5G for NOMA-based Massive MIMO and Relaying Systems" (2016). Electronic Thesis and Dissertation Repository. 4136.