Presentation | 2016-01-18 Effectiveness of L1 Regularization for Sparse Impulse Response Estimation Using colored Noise Keito Kito, Ryo Tanaka, Takahiro Murakami, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | In this paper, we show an effectiveness of L1 regularization for a sparse impulse response estimation using a colored noise. When an impulse response is estimated, a model to be estimated is modeled using a white noise as an input signal. However, in practice, the noise is not an exact white noise due to an influence of frequency characteristics of measurement instruments. Therefore, a technique that estimates an accurate impulse response using a colored noise is required. When the colored noise is used as the input signal, the performance of the impulse response estimation using the second order statistics such as the MSE (Mean Square Error) algorithm is degraded. Then, we apply the negentropy and the L1-norm to the impulse response estimation. By maximizing the negentropy, the performance of estimation is improved in terms of the probability distribution. By minimising the L1-norm, the performance of estimation is improved in terms of sparsity of the impulse response. In the proposed method, we use the FOBOS (Forward Backward Splitting) algorithm to optimize the negentropy and the L1-norm. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | impulse response / independent component analysis / negentropy / sparsity / L1 regularization / FOBOS |
Paper # | IT2015-58,SIP2015-72,RCS2015-290 |
Date of Issue | 2016-01-11 (IT, SIP, RCS) |
Conference Information | |
Committee | RCS / IT / SIP |
---|---|
Conference Date | 2016/1/18(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Kwansei Gakuin Univ. Osaka Umeda Campus |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Signal Processing for Wireless Communications, Learning, Mathematical Science, Communication Theory, etc. |
Chair | Makoto Taromaru(Fukuoka Univ.) / Yasutada Oohama(Univ. of Electro-Comm.) / Osamu Houshuyama(NEC) |
Vice Chair | Hidekazu Murata(Kyoto Univ.) / Satoshi Denno(Okayama Univ.) / Yukitoshi Sanada(Keio Univ.) / Tadashi Wadayama(Nagoya Inst. of Tech.) / Makoto Nakashizuka(Chiba Inst. of Tech.) / Masahiro Okuda(Univ. of Kitakyushu) |
Secretary | Hidekazu Murata(Mitsubishi Electric) / Satoshi Denno(NTT DoCoMo) / Yukitoshi Sanada(Univ. of Electro-Comm.) / Tadashi Wadayama(Wakayama Univ.) / Makoto Nakashizuka(NEC) / Masahiro Okuda(Ritsumeikan Univ.) |
Assistant | Jun Mashino(NTT) / Tetsuya Yamamoto(Panasonic) / Takamichi Inoue(NEC) / Tomoya Tandai(Toshiba) / Toshihiko Nishimura(Hokkaido Univ.) / Takuya Kusaka(Okayama Univ.) / Takamichi Miyata(Chiba Inst. of Tech.) |
Paper Information | |
Registration To | Technical Committee on Radio Communication Systems / Technical Committee on Information Theory / Technical Committee on Signal Processing |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Effectiveness of L1 Regularization for Sparse Impulse Response Estimation Using colored Noise |
Sub Title (in English) | |
Keyword(1) | impulse response |
Keyword(2) | independent component analysis |
Keyword(3) | negentropy |
Keyword(4) | sparsity |
Keyword(5) | L1 regularization |
Keyword(6) | FOBOS |
1st Author's Name | Keito Kito |
1st Author's Affiliation | Meiji University(Meiji Univ.) |
2nd Author's Name | Ryo Tanaka |
2nd Author's Affiliation | Meiji University(Meiji Univ.) |
3rd Author's Name | Takahiro Murakami |
3rd Author's Affiliation | Meiji University(Meiji Univ.) |
Date | 2016-01-18 |
Paper # | IT2015-58,SIP2015-72,RCS2015-290 |
Volume (vol) | vol.115 |
Number (no) | IT-394,SIP-395,RCS-396 |
Page | pp.pp.61-66(IT), pp.61-66(SIP), pp.61-66(RCS), |
#Pages | 6 |
Date of Issue | 2016-01-11 (IT, SIP, RCS) |