| Pattern division multiple access (PDMA), which is a novel non-orthogonal multiple access scheme proposed by CATT, targets to satisfy the requirements of massive connectivity, high spectral efficiency and low latency for the fifth generation (5G) mobile communication system. PDMA is based on the joint optimization design of transmitter and receiver, with non-orthogonal pattern of unequal diversity at the transmitter and general SIC detection algorithm at the receiver. PDMApattern defines the sparse mapping of data to a resource group on multiple signal domains (including time, frequency, and spatial domain) to differentiate different users. At present, PDMA has been actively discussed in 3rd Generation Partnership Project (3GPP), and accepted in international telecommunication union (ITU) report which is named Future technology trends of terrestrial IMT systems, and hence, it is a candidate key technology for the 5G systems.When bringing PDMA technology from theory to practice, many unexpected challenges are found. In the aspect of theory research, the theoretical performance analysis of the PDMA matrix pattern is still incomplete, and there is lack of near-optimal design method of PDMA matrix pattern based on the joint design of transmitter and receiver. In the practical application of 5 G system, there are lack of high performance low complexity detection algorithm and link-level performance estimation algorithm, and no complete radio transmission scheme for massive connections in 5G. In this dissertation, we focus on the research of key technologies of PDMA for 5G, and try to solve above-mentioned problems.First of all, we study the PDMA pattern matrix design at the transmitter side. Firstly, for uplink (UL) mMTC, the design criteria of PDMA pattern matrix are to minimize sum squared correlation (SSC)with a given correlation coefficient threshold. Secondly, for UL enhanced mobile broadband (eMBB), the design criterion is to maximize constellation constrained-capacity (CC-Capacity) with a given maximum row weight. Finally, the performance of a given PDMA pattern matrix is analyzed. Extensive simulation results show that the proposed PDMA pattern matrices achieve good performance with manageable complexity.Then, low-complexity high-performance detection algorithms for PDMA are researched. Firstly, a low-complexity minimum mean-square error -interference rejection combining (MMSE-IRC) signal detection algorithm is proposed for uplink degraded PDMA and massive MIMO system, which exploits a dimension-reduction technique based on the eigenvalue decomposition to reduce the complexity compared with the conventional MMSE-IRC algorithm. Secondly, we proposed a belief propagation-iterative detection and decoding (BP-IDD) receiver for generally coded PDMA uplink system, which can fully exploit coding potentials and diversity gains. The simulation results show that the proposed algorithms can balance the trade-off of performance and complexityIn addition, we propose a simple yet accurate link performance estimation technique for PDMA uplink system, which is named the genie-aided interference cancellation (GIC)-only (GO) method. By utilizing the GIC receiver with a single optimal parameter ?o to estimate the effective SINR, accurate link performance prediction has been achieved. Based on numerical results, we have confirmed that the proposed GO method achieves greater link performance estimation accuracy over the conventional method and is robust to different configuration parameters.Finally, we investigate radio transmission scheme design of PDMA for 5G uplink and downlink (DL). Firstly, for UL mMTC scenario, we propose a general solution for PDMA UL grant-free transmission, and design a new pilot scheme for grant-free transmission. Secondly, for DL eMBB scenario, a general solution of PDMA wideband and subband scheduling is given. The simulation results show that the proposed PDMA scheme has better performance over OMA scheme in terms of conneted users and spectral efficiency. |
PDF Full Text Request