NON-ORTHOGONAL MULTIPLE ACCESS SYSTEM BASED ON COMBINATORIAL DESIGN
WU, YUTENG WU
MetadataShow full item record
In this thesis, a new class of multiuser Non-Orthogonal Multiple Access (NOMA) is studied. Overall, the main contributions are listed as follows: Firstly, we study a new class of the space-time code (STC) based on a circulant matrix. A diagonal transmission matrix is adopted by utilizing the singular value decomposition (SVD) on the circulant matrix. After that, we present a Multiuser Multidimensional STC. This code expand the previous diagonal transmission matrix to multiple dimensions. Hence it produces a diversity due to the linear combination of all input elements. Two modified constellations are presented aiming at reducing the constellation expansion and the implementation burden. Secondly, a brief survey on state-of-the-art NOMA systems is provided, where their performance in terms of bit error rate (BER), overload capability and spectrum efficiency are compared. Thirdly, we study a NOMA system systematically. Firstly, an uplink NOMA system based on the Steiner Triple System (STS) is presented. A multiuser detection algorithm by using a reduced size vector correlator which will decrease the receiver complexity with an acceptable performance is proposed. Performance-oriented algorithm and overload-oriented algorithm aiming at handling different channel scenarios are studied. After that, we present a novel constellation forming methodology of downlink NOMA system aiming at solving the surjective-only mapping problem in NOMA systems. Compared with other methods by using constellation shaping, phase shifting and power scaling [1-4], our paradigm is capable of providing a new way to solve the surjective-only mapping with low design complexity. As far as we know, this method has not been used before. Finally, the impact of channel estimation error for NOMA system is simulated. The idea behind channel estimation is transmitting the pilot sequence and estimating the channel coefficient based on the received pilot sequence. As a future work, we will work on ARQ, cryptography and optimal designs.