Logo image
A spatial-coupled sparse code multiple access scheme
Thesis

A spatial-coupled sparse code multiple access scheme

Lin, Chien Chun
Masters, 國立清華大學, 通訊工程研究所
2016

Abstract

稀疏碼多重存取 低密度奇偶檢查碼 外部資訊傳遞演算法 sparse code multiple access low-density parity-check code extrinsic information transfer
In recent years, the demand for high data transmission throughput grows significantly with the progress of the technologies for mobile communications. Most of these technologies are based on the idea of orthogonal multiple access (OMA) in current wireless systems. However, it is concerned that spectrum resources may not be fully exploited when considering conventional OMA schemes. Consequently, the non-orthogonal multiple access (NOMA), which is one of the promising radio access technique for 5G communication systems, is proposed. Comparing to OMA-based systems, each spectrum resource can be allocated to more than one users, and hence yields a significant improvement in throughput performances. In this thesis, we propose a novel scheme for the sparse code multiple access (SCMA), which is one of the NOMA-based techniques. In the proposed scheme, the users are allowed to transmit their data not only through the resources assigned for themselves in the current transmission block, but also through other resources, which are originally assigned for other users, in the next transmission block. As a result, each of the transmission blocks is coupled with the previous and the afterward transmission block. Consequently, the \textit{a priori} information for each of the transmission blocks can be obtained before its own detection process, from the detection process of the previous transmission block. In addition, the low-density parity-check (LDPC) codes, which are considered as the outer channel codes, can be optimized by using the extrinsic information transfer (EXIT) technique. It is shown through simulations that the bit-error rate (BER) the performance of the proposed scheme is superior to that of the original SCMA scheme.

Metrics

1 Record Views

Details

Logo image