| Presentation | 2021-03-05 A Study on Trainable Gaussian Belief Propagation Using Deep Unfolding for Large NOMA Detection Ryuichi Kume, Shinsuke Ibi, Takumi Takahashi, Hisato Iwai, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | This paper compares transmission characteristics of sparse code multiple access (SCMA) and sparse superposition code (SSC)-aided non-orthogonal multiple access (NOMA). Furthermore, the effect of deep unfolding is validated by comparing Gaussian belief propagation (GaBP) and trainable GaBP (T-GaBP), where the internal parameters embedded in GaBP are trained by data-drive tuning, in terms of detection capability. SCMA has a sparse structure in the codewords, and this makes it possible to mitigate severe interference between users even in high-loading scenarios. SSC is one of the powerful error-correcting codes that achieve channel capacity in Gaussian channels, and its sparse structure is similar to SCMA codeword when applied to multiple access. The trainable parameters in T-GaBP works to mitigate the ill-convergence behavior of iterative detection due to the negative effects of outliers caused by poor accuracy of the large-system approximation. For each detector, computer simulations compare bit error rate (BER) performances of SSC and SCMA. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Sparse superposition code / non-orthogonal multiple access / Gaussian belief propagation / sparse code multiple access / deep unfolding |
| Paper # | RCS2020-255 |
| Date of Issue | 2021-02-24 (RCS) |
| Conference Information | |
| Committee | RCS / SR / SRW |
|---|---|
| Conference Date | 2021/3/3(3days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Online |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Mobile Communication Workshop |
| Chair | Eiji Okamoto(Nagoya Inst. of Tech.) / Masayuki Ariyoshi(NEC) / Satoshi Denno(Okayama Univ.) |
| Vice Chair | Fumiaki Maehara(Waseda Univ.) / Toshihiko Nishimura(Hokkaido Univ.) / Tomoya Tandai(Toshiba) / Suguru Kameda(Tohoku Univ.) / Osamu Takyu(Shinshu Univ.) / Kentaro Ishidu(NICT) / Keiichi Mizutani(Kyoto Univ.) / Kentaro Saito(Tokyo Inst. of Tech.) / Hanako Noda(Anritsu) |
| Secretary | Fumiaki Maehara(Kyushu Univ.) / Toshihiko Nishimura(NEC) / Tomoya Tandai(ATR) / Suguru Kameda(Univ. of Electro-Comm.) / Osamu Takyu(Mie Univ.) / Kentaro Ishidu(NTT) / Keiichi Mizutani(NIigata Univ.) / Kentaro Saito / Hanako Noda |
| Assistant | Koichi Adachi(Univ. of Electro-Comm.) / Osamu Nakamura(Sharp) / Manabu Sakai(Mitsubishi Electric) / Masashi Iwabuchi(NTT) / Tatsuki Okuyama(NTT DOCOMO) / Mai Ohta(Fukuoka Univ.) / Teppei Oyama(Fujitsu Lab.) / Kentaro Kobayashi(Nagoya Univ.) / Masaaki Fuse(Anritsu) / Akihito Noda(Nanzan Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Radio Communication Systems / Technical Committee on Smart Radio / Technical Committee on Short Range Wireless Communications |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | A Study on Trainable Gaussian Belief Propagation Using Deep Unfolding for Large NOMA Detection |
| Sub Title (in English) | |
| Keyword(1) | Sparse superposition code |
| Keyword(2) | non-orthogonal multiple access |
| Keyword(3) | Gaussian belief propagation |
| Keyword(4) | sparse code multiple access |
| Keyword(5) | deep unfolding |
| 1st Author's Name | Ryuichi Kume |
| 1st Author's Affiliation | Doshisha University(Doshisha Univ.) |
| 2nd Author's Name | Shinsuke Ibi |
| 2nd Author's Affiliation | Doshisha University(Doshisha Univ.) |
| 3rd Author's Name | Takumi Takahashi |
| 3rd Author's Affiliation | Osaka University(Osaka Univ.) |
| 4th Author's Name | Hisato Iwai |
| 4th Author's Affiliation | Doshisha University(Doshisha Univ.) |
| Date | 2021-03-05 |
| Paper # | RCS2020-255 |
| Volume (vol) | vol.120 |
| Number (no) | RCS-404 |
| Page | pp.pp.240-245(RCS), |
| #Pages | 6 |
| Date of Issue | 2021-02-24 (RCS) |