| Item type |
SIG Technical Reports(1) |
| 公開日 |
2024-05-02 |
| タイトル |
|
|
タイトル |
Exploring Passive Activity Recognition using Multi-shot UWB CIRs |
| タイトル |
|
|
言語 |
en |
|
タイトル |
Exploring Passive Activity Recognition using Multi-shot UWB CIRs |
| 言語 |
|
|
言語 |
eng |
| キーワード |
|
|
主題Scheme |
Other |
|
主題 |
行動認識 |
| 資源タイプ |
|
|
資源タイプ識別子 |
http://purl.org/coar/resource_type/c_18gh |
|
資源タイプ |
technical report |
| 著者所属 |
|
|
|
Department of Computer Science, School of Engineering, Tokyo Institute of Technology |
| 著者所属 |
|
|
|
Department of Computer Science, School of Engineering, Tokyo Institute of Technology |
| 著者所属 |
|
|
|
LY Corporation |
| 著者所属 |
|
|
|
Ritsumeikan University |
| 著者所属 |
|
|
|
Department of Computer Science, School of Engineering, Tokyo Institute of Technology |
| 著者所属(英) |
|
|
|
en |
|
|
Department of Computer Science, School of Engineering, Tokyo Institute of Technology |
| 著者所属(英) |
|
|
|
en |
|
|
Department of Computer Science, School of Engineering, Tokyo Institute of Technology |
| 著者所属(英) |
|
|
|
en |
|
|
LY Corporation |
| 著者所属(英) |
|
|
|
en |
|
|
Ritsumeikan University |
| 著者所属(英) |
|
|
|
en |
|
|
Department of Computer Science, School of Engineering, Tokyo Institute of Technology |
| 著者名 |
Han, Lin
Atsushi, Nomura
Kota, Tsubouchi
Nobuhiko, Nishio
Masamichi, Shimosaka
|
| 著者名(英) |
Han, Lin
Atsushi, Nomura
Kota, Tsubouchi
Nobuhiko, Nishio
Masamichi, Shimosaka
|
| 論文抄録 |
|
|
内容記述タイプ |
Other |
|
内容記述 |
As a promising device-free solution, wireless sensing techniques have been applied to various applications while addressing privacy concerns associated with traditional computer vision-based methods. There has been limited research focusing on the use of UWB CIR for activity recognition and showing improved performance over Wi-Fi CSI. However, previous studies using one-shot CIR cannot correctly capture the variations inherent in dynamic activities. This research proposes a device-free activity recognition approach by utilizing multi-shot CIRs, consisting of several CIRs arranged in a time series, in order to cover the transition of activities. And three variants of wavelet denoising across various dimensions are introduced to remove noise on signal. Experiments were conducted with horizontal and vertical device settings to test the sensing performance in different scenarios. The results were also compared with Wi-Fi CSI at the same frequency to benchmark its effectiveness. |
| 論文抄録(英) |
|
|
内容記述タイプ |
Other |
|
内容記述 |
As a promising device-free solution, wireless sensing techniques have been applied to various applications while addressing privacy concerns associated with traditional computer vision-based methods. There has been limited research focusing on the use of UWB CIR for activity recognition and showing improved performance over Wi-Fi CSI. However, previous studies using one-shot CIR cannot correctly capture the variations inherent in dynamic activities. This research proposes a device-free activity recognition approach by utilizing multi-shot CIRs, consisting of several CIRs arranged in a time series, in order to cover the transition of activities. And three variants of wavelet denoising across various dimensions are introduced to remove noise on signal. Experiments were conducted with horizontal and vertical device settings to test the sensing performance in different scenarios. The results were also compared with Wi-Fi CSI at the same frequency to benchmark its effectiveness. |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AA11838947 |
| 書誌情報 |
研究報告ユビキタスコンピューティングシステム（UBI）
巻 2024-UBI-82,
号 9,
p. 1-7,
発行日 2024-05-02
|
| ISSN |
|
|
収録物識別子タイプ |
ISSN |
|
収録物識別子 |
2188-8698 |
| Notice |
|
|
|
SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc. |
| 出版者 |
|
|
言語 |
ja |
|
出版者 |
情報処理学会 |
IPSJ-UBI24082009.pdf (20.5 MB)
