Non-Orthogonal Transmission for Future Wireless Communication Systems

Abstract

The 5G networks are envisioned to support large data traffic, massive connectivity, and ultra low-latency communications. This thesis focuses on the low-complexity design and performance analysis of non-orthogonal transmissions, which is one of the potential candidates to address the aforementioned challenges. We first develop a low-complexity transceiver design for massive MIMO channels. By exploring a sparse representation of the MIMO channel in the virtual angular domain, we generate a set of transmit-receive beam pairs to support the transmission of multiple data streams. These data streams can be easily separated via SIC, and the power allocation is optimized with water-filling. The near-optimal DoF and capacity achieved by the proposed approach are analysed. Next, we investigate the joint antenna selection (AS) problem for MIMO TWRNs. Two near-optimal algorithms, namely the joint relay-source AS (JRSAS) and the separated relay-source AS (SRSAS), are proposed in a greedy manner. Numerical results show that both JRSAS and SRSAS can approach the optimal AS algorithm but with much lower computational complexity. Subsequently, we extend the joint AS problem to a MIMO NOMA system. For the rate-adaptive NOMA, the throughput-oriented AS algorithms are first proposed. For the fixed-rate NOMA, the outage-oriented AS approaches are then developed. The average capacity and outage performance of the proposed algorithms are analysed and compared to the OMA counterparts. By naturally integrating the short-packet and NOMA techniques, we finally investigate the potential ability of short-packet communications using NOMA to fulfil low-latency transmission. For a given set of reliability constraints of users, the transmission latency of a fundamental two-user model is first studied, which is followed by a more practical multi-user system. Numerical results demonstrate the superior performance of NOMA in reducing physical-layer transmission latency in short-packet communications.