研究実績の概要 |
We are progressing according to the original plan in 2016, and achieved the following results.
We focused on the design of spatially coupled concatenated codes. We proposed two kinds of spatially coupled concatenated codes. One is the RA-extended SC-LDPC code, whose uncoupled version consists of an LDPC outer code for spatial coupling and a repeat-accumulate (RA) inner code to adjust code rate. We calculated the potential threshold of the base code ensemble to predict the iterative decoding performance of the proposed code. Numerical results show that in the low rate region, the proposed code outperforms the conventional codes. The proposed code is promising in low rate coding workplace, e.g., for the multiple-access channel.
The other proposed code is a spatially coupled repeater-combiner-convolutional (SC-RCC) code, whose uncoupled version consists of a repeater and combiner as outer code for spatial coupling and a convolutional code as inner code. For point-to-point channel and two-user multiple-access channel, numerical results show the proposed code is universal, in the sense that it approaches the achievable power regions of both channels. For multi-user multiple-access channel, the sum rate of the proposed code, by concatenated a spreading, i.e., repetition code, are also close to the channel capacity.
|
現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
As summarized in the Research Achievements section, the research is smoothly progressing according to the original plan in 2016. The proposed RA-extended SC-LDPC code is rate-compatible, and the proposed SC-RCC code concatenated with a repetition code provides a rate-variable code. Both codes are available for multiple-access channel. We show that the proposed SC-RCC code is universal in the sense that it approaches the achievable power regions of single- and two-user multiple-access channels. It seems hard to prove that the code is the universal in the multi-user multiple-access channel.
|
今後の研究の推進方策 |
Research will proceed according to the original plan. In the second year, emphasis is on the decoding performance of the codes with finite length. Although the proposed RA-extended SC-LDPC code and SC-RCC code work well with finite length, we still prefer a code with variable rate that can approach the capacity of multiple-access channel, even with a moderate code length. A way to this purpose is to spatially couple irregular LDPC codes. We have some attempt in this direction in 2016, and will thoroughly investigate the spatially couple irregular LDPC code in 2017.
|