Summary
IEICE Information and Communication Technology Forum
2018
Session Number:SESSION01
Session:
Number:SESSION01_2
RFID-Based Deep Shopping Data Acquisition Scheme with Multiple Feature Extraction
Shinichiro Aita, Hiromu Asahina, Kentaroh Toyoda, Iwao Sasase,
pp.-
Publication Date:2018/8/31
Online ISSN:2188-5079
DOI:10.34385/proc.32.SESSION01_2
Summary:
In order to accurately collect the deep shopping data, such as which item a consumer picks in a shop, radio frequency identification (RFID) tags based deep shopping data acquisition scheme have drawn attention. The existing RFID-based deep shopping data acquisition scheme that uses received signal strength (RSS) is not accurate enough to use in practice. In this paper, we propose an RFID-based deep shopping data acquisition scheme with multiple feature extraction. To realize more accurate deep shopping data acquisition, we use not only RSS but also phase and RFID tag read count. Supervised machine learning technique together with such multiple features improve the accuracy. By introducing read count and the phase, it is possible to detect when the fluctuation amount of the RSS cannot be well observed. The proposed scheme was tested in the lab environment and it improves detection accuracy when the conventional scheme does not work well.