2019 Fiscal Year Annual Research Report
Multidimensional compressive sensing based technologies for next-generation MIMO radar with SL3: Super-resolution, Low-complexity, Low-cost and Low-consumption
| Project/Area Number |
15K06072
|
|---|
| Research Institution | Akita Prefectural University |
|---|
Principal Investigator |
徐 粒 秋田県立大学, システム科学技術学部, 教授 (40252324)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
桂 冠 秋田県立大学, システム科学技術学部, 特任助教 (80734904) [Withdrawn]
松下 慎也 秋田県立大学, システム科学技術学部, 准教授 (20435449)
|
|---|
| Project Period (FY) |
2015-04-01 – 2020-03-31
|
| Keywords | compressive sensing / sparse representation / sparse signal processing / adaptive algorithm / MIMO radar system / DOA estimation |
|---|
| Outline of Annual Research Achievements |
This research has developed a class of new CS-based methods for under-sampled signal processing of MIMO radar systems, with higher resolution,less computational complexity and stronger robustness against non-Gaussian impulsive noise than the existing CS-based methods.Two main points are briefly stated below, while the other details can be referred to our published papers.
2D CS algorithms for 2D signal reconstruction and the corresponding 2D-CS-based methods for the range and velocity joint detection problem have been established, which no longer require to vectorize the original 2D CS problem to the 1D case and thus can reduce the computational complexity largely. Moreover, robust 2D-CS-based methods have also been proposed, which significantly improve the performance of CS-based methods for non-Gaussian impulsive noise environment.
A new algorithm with a pre-estimation to reduce the dimensionality of the measurement matrix has been proposed for the direction-of-arrival (DOA) estimation with a small number of noisy snapshots.In the first stage, the range of interest is divided into a relatively low-resolution grid, and the conventional beam former is used to quickly identify the candidate or potential areas where true sources may exist. In the second stage, the obtained candidate areas are divided into a denser sampling grid,and the l_{2,1}-norm penalty is used to solve the corresponding multiple measurement vectors (MMV) problem. Simulation results demonstrate that the proposed methods have much lower complexity and higher accuracy than the known CS-based methods.
|
| Remarks |
研究業績
http://www.akita-pu.ac.jp/system/elect/sce/xuli/publist.html
|
