Presentation | 2018-05-24 [Poster Presentation] Investigation of Feature Reduction for Body Motion Identification using Radio Channel Characteristics in Wireless BAN Yuki Ichikawa, Minseok Kim, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | In this article, the reduction of the features used for human motion classification using decision tree machine learning algorithm has been investigated.We developed a custom wearable BAN channel measurement system using a commercial cheap product of a sensor network, where a radio signal strength between four sensors and coordinator can be measured. After preprocessing of interpolation and denoising for the raw data, the time domain features are calculated by varying the sample window sizes where six window sizes were chosen from the period of the body motion.After that, Feature Importance was calculated for feature reduction.As a result, we clarified that the sensor can be reduced to those in the vicinity of the coordinator when the sample window size used for feature calculation is set to about half of the maximum period of the body motions, and that Mean Value Slope is not involved for classification. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | machine learning / feature / decision tree / BAN / radio channel / LQI / motion identification |
Paper # | RCC2018-4,MICT2018-4 |
Date of Issue | 2018-05-17 (RCC, MICT) |
Conference Information | |
Committee | RCC / MICT |
---|---|
Conference Date | 2018/5/24(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Tokyo Big Sight |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Shinsuke Hara(Osaka City Univ.) / Masaru Sugimachi(National Cerebral and Cardiovascular Center) |
Vice Chair | Kazunori Hayashi(Osaka City Univ.) / Ryu Miura(NICT) / Shinsuke Hara(Osaka City Univ.) / Takahiro Aoyagi(Tokyo Inst. of Tech.) |
Secretary | Kazunori Hayashi(Kagawa Univ.) / Ryu Miura(Hokkaido Univ.) / Shinsuke Hara(Niigata Univ.) / Takahiro Aoyagi(Saga Univ.) |
Assistant | Toshinori Kagawa(NICT) / Kentaro Kobayashi(Nagoya Univ.) / Takumi Kobayashi(Yokohama National Univ.) / Shintaro Izumi(Kobe Univ.) / Tomoko Tateyama(Hiroshima Inst. of Tech.) / Ami Tanaka(Ritsumeikan Univ.) / Daisuke Anzai(Nagoya Inst. of Tech.) |
Paper Information | |
Registration To | Technical Committee on Reliable Communication and Control / Technical Committee on Healthcare and Medical Information Communication Technology |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | [Poster Presentation] Investigation of Feature Reduction for Body Motion Identification using Radio Channel Characteristics in Wireless BAN |
Sub Title (in English) | |
Keyword(1) | machine learning |
Keyword(2) | feature |
Keyword(3) | decision tree |
Keyword(4) | BAN |
Keyword(5) | radio channel |
Keyword(6) | LQI |
Keyword(7) | motion identification |
1st Author's Name | Yuki Ichikawa |
1st Author's Affiliation | Niigata University(Niigata Univ.) |
2nd Author's Name | Minseok Kim |
2nd Author's Affiliation | Niigata University(Niigata Univ.) |
Date | 2018-05-24 |
Paper # | RCC2018-4,MICT2018-4 |
Volume (vol) | vol.118 |
Number (no) | RCC-53,MICT-54 |
Page | pp.pp.17-20(RCC), pp.17-20(MICT), |
#Pages | 4 |
Date of Issue | 2018-05-17 (RCC, MICT) |