Presentation 2021-10-14
An Experimental Study on Improving Accuracy of Location Estimation in Finger Print Using CNN and ResNet
Yu Sakanishi, Satoru Aikawa, Shinichiro Yamamoto, Yuta Sakai,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Recently, indoor navigation system is one of the important technologies. We are studyingan indoor location estimation technique using wireless LAN employing a Finger Print scheme. A database (DB) and user data (UD) are measured from wireless LAN radio waves to estimate the location using the Finger Print. A CNN is utilized to compare the UD and DB. However, the CNN estimation accuracy may deteriorate due to the degradation problem caused by the increase in the number of middle layers and the gradient vanishing problem. ResNet is a solution to this problem. In this study, we have evaluated estimation accuracies of CNN and ResNet experimentally.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Wireless LAN / Finger Print / CNN / ResNet
Paper # CS2021-52
Date of Issue 2021-10-07 (CS)

Conference Information
Committee CS
Conference Date 2021/10/14(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Online
Topics (in Japanese) (See Japanese page)
Topics (in English) Broadband access, Home network, Network service, Communication applications, etc.
Chair Jun Terada(NTT)
Vice Chair Daisuke Umehara(Kyoto Inst. of Tech.)
Secretary Daisuke Umehara(NICT)
Assistant Takahiro Yamaura(Toshiba) / Yuta Ida(Yamaguchi Univ.)

Paper Information
Registration To Technical Committee on Communication Systems
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) An Experimental Study on Improving Accuracy of Location Estimation in Finger Print Using CNN and ResNet
Sub Title (in English)
Keyword(1) Wireless LAN
Keyword(2) Finger Print
Keyword(3) CNN
Keyword(4) ResNet
1st Author's Name Yu Sakanishi
1st Author's Affiliation University of Hyogo(Univ of Hyogo)
2nd Author's Name Satoru Aikawa
2nd Author's Affiliation University of Hyogo(Univ of Hyogo)
3rd Author's Name Shinichiro Yamamoto
3rd Author's Affiliation University of Hyogo(Univ of Hyogo)
4th Author's Name Yuta Sakai
4th Author's Affiliation University of Hyogo(Univ of Hyogo)
Date 2021-10-14
Paper # CS2021-52
Volume (vol) vol.121
Number (no) CS-198
Page pp.pp.1-5(CS),
#Pages 5
Date of Issue 2021-10-07 (CS)