2015 Fiscal Year Annual Research Report
LatticeNET: Practical Lattice Codes for Cooperative Wireless Networks
| Project/Area Number |
26289119
|
|---|
| Research Institution | Japan Advanced Institute of Science and Technology |
|---|
Principal Investigator |
KURKOSKI Brian 北陸先端科学技術大学院大学, 情報科学研究科, 准教授 (80444123)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
和田山 正 名古屋工業大学, 工学(系)研究科(研究院), 教授 (20275374)
松本 正 北陸先端科学技術大学院大学, 情報科学研究科, 教授 (40452114)
八木 秀樹 電気通信大学, 情報理工学(系)研究科, 准教授 (60409737)
|
|---|
| Project Period (FY) |
2014-04-01 – 2017-03-31
|
| Keywords | 格子理論 / 符号理論 / 情報理論 / ネットワーク符号 / 無線通信 |
|---|
| Outline of Annual Research Achievements |
The LatticeNet Project is divided into five work packages (WP). We are progressing according to the original plan, and in 2015 achieved the following results.
WP1 is coding lattice designs which are efficient to encode and decode. We developed a new decoding algorithm for low-density lattice codes (LDLC), which has the best performance/complexity tradeoff at dimension 100 and 1000. This three/two Gaussian parametric LDLC lattice decoding algorithm and its analysis was presented at IEEE Information Theory Workshop (Oct 2015, Jeju Island, Korea).
WP2 We have developed a method to allow the shaping lattice and the coding lattice to be distinct, exploiting the advantages of each. The consequence are practical lattice codes that can be used for power-constrained physical layer network coding.
WP3 is the design codes for the relay channel. This research clarified that a write-once memory code designed for data storage can be applied to coding for the relay channel. The main results of his research are (1) WOM codes can achieve the capacity of the AMAC channel, and these codes have a special property that allows them to be efficiently decoded.
WP4. New algorithms for wireless networks using PLNC presented at ICITCC (Mar 2016, Cambridge, United Kingdom)
WP5. LatticeNet supported the visit to Japan of two prominent researchers, from USA and Australia. They gave presentations at "Seminar on Wireless Communications and Lattices," held at the University of Electro-Communications (Jun 22, 2015, Tokyo), attended by about 40 people. Intensive research discussions were held.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
As summarized in the Research Achievements section, the research is smoothly progressing according to the original plan for the five work packages (WP1-WP5). In particular, of the four technical work packages (WP1-WP4), WP1 and WP2 are nearly complete with journal papers under review or near submission. WP3 is making progress, with a journal paper under review. WP4 is now in the research stages. WP5 may be regarded as complete, since two mini-workshops were held in 2014 and 2015, according to the plan.
A LatticeNet project meeting was held on June 23, 2015 at the University of Electro-Communications. It was attended by two Advisory Committee (AC) members from overseas. The AC members gave favorable feedback about the progress, and helpful suggestions about future directions.
|
| Strategy for Future Research Activity |
Research in 2016 will proceed according to the original plan. In the third year, emphasis is on journal papers for WP1-2, and progressing with WP3-4.
WP1. A journal paper on efficient LDLC decoding is has been submitted, and acceptance is expected. We recently obtained an unexpectedly appealing insight into the design of LDLC lattices, and in this year we will submit a journal paper on this result.
WP2. The core ideas of power-constrained lattice construction are established, and now a journal version of this work is nearing submission. Particularly important is the the co-design of the LDLC lattice and shaping lattice of codes for physical layer network coding.
WP3-4. A journal paper based on WOM codes for relay channels is currently under review. Since writing the original proposal, the importance of lattices for physical-layer network coding has become clear. Thus, rather than the single-user relay channel, lattices are highly suited for the two-user relay channel. We plan to show that lattices can significantly improve throughput when they are used in a multi-access relay channel. This combines the key characteristics of the relay channel (WP3) and more general multi-terminal Gaussian networks (WP4).
According to the original plan, mini-workshops were held in 2014 and 2015, and the requirements of the plan (WP5) have been satisfied. A further mini-workshop may be held in 2016, either with or without invited international collaborators.
|
| Causes of Carryover |
According to the original plan, the budget is used for employing graduate students as researchers. They conduct research essential to the project, give oral presentations, and write conference and journal papers. Because student researchers received scholarships in H27, amounts from H27 were carried forward to H28. Travel to domestic and international conferences is important for disseminating research results to the research community.
|
| Expenditure Plan for Carryover Budget |
Including the amount carried forward from H27, the total amount allocated in H28 is approximately 6,000,000 yen. Roughly speaking, this will be allocated as 4,000,000 yen for Personnel Expenditure and Renumeration, and 2,000,000 yen for Travel. This allocation efficiently supports the goals of the LatticeNet project, and is consistent with the original plan.
|
