| Project/Area Number |
26820141
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Communication/Network engineering
|
|---|
| Research Institution | Nagoya University |
|---|
Principal Investigator |
MORI Yojiro 名古屋大学, 工学(系)研究科(研究院), 助教 (10722100)
|
|---|
| Project Period (FY) |
2014-04-01 – 2016-03-31
|
| Project Status |
Completed (Fiscal Year 2015)
|
| Budget Amount *help |
¥3,900,000 (Direct Cost: ¥3,000,000、Indirect Cost: ¥900,000)
Fiscal Year 2015: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Fiscal Year 2014: ¥2,340,000 (Direct Cost: ¥1,800,000、Indirect Cost: ¥540,000)
|
|---|
| Keywords | 光ファイバ通信 / 光ネットワーク / 機械学習 / 光ファイバ / 光フィルタ / 光ノード |
|---|
| Outline of Final Research Achievements |
In optical fiber communication systems, the transmissible capacity and distance are strictly limited by transmission impairments such as fiber nonlinearity in links. To maximize network performance, signal distortion caused by such impairments must be eliminated. Machine learning can discriminate symbols in accordance with their distortion. As a result, the demodulation performance can be improved. However, cycle slips induced by laser phase noise disrupt the learning process and hence constructed decision boundaries are distorted. To resolve this problem, we propose a novel configuration in which cycle-slip detection and correction are introduced into the machine learning process. With such a scheme, we can reduce the number of bit errors in a stable manner. The proposed scheme substantially elongate the transmissible distance without reducing capacity.
|
