Abstract
Non-orthogonal multiple access (NOMA) is an emerging technology for future cellular systems in order to accommodate more users via non-orthogonal resource allocation, especially when the number of users/devices exceeds the available degrees of freedom, resulting in an overloaded condition. To date, low density signature (LDS) and sparse code multiple access (SCMA) are two promising NOMA techniques. However, research in this area is still in its infancy and there are still several open issues in the LDS/SCMA transceiver design. This thesis aims to address some of these challenges. The contributions are summarized as follows. 1:-LDS-OFDM and low density parity check (LDPC) codes both can be represented by a bipartite graph. Inspired by their similar structure, we construct a joint sparse graph combining the single graphs of LDS-OFDM and LDPC codes, namely joint sparse graph for OFDM (JSG-OFDM). A joint detection and decoding scenario is proposed using message passing algorithm (MPA). Design guidelines for the joint sparse graph are derived through extrinsic information transfer (EXIT) chart analysis. Simulation results show that the JSG-OFDM outperforms existing techniques such as GO-MC-CDMA, LDS-OFDM and turbo structured LDS-OFDM. 2:-Due to the higher power and spectral efficiency, the filter-bank multi-carrier (FBMC) technique is a promising alternative to OFDM. We use a low density graph to model the weight matrix of intrinsic interference in the isotropic orthogonal transform algorithm (IOTA) filtered FBMC system. In addition, such graph is combined with LDS and LDPC codes to form a joint sparse graph for FBMC-IOTA (JSG-IOTA). Based on MPA, a joint detection and decoding scheme is designed for JSG-IOTA, and the joint sparse graph is optimized by EXIT chart analysis. Numerical results show the superiority of JSG-IOTA to conventional techniques. 3:-In SCMA, the processes of bit to symbol mapping and LDS spreading are combined together. We investigate multi-dimensional SCMA codebooks, and the design rules are derived to maximize the constellation shaping gain. Moreover, we propose to construct SCMA codebooks by copy-and-permute operation on protographs and three-dimensional (3D) constellation shaping. Simulation results show that SCMA outperforms LDS with high-order constellations, and the proposed optimization methods can further improve the SCMA performance.