Summary
International Conference on Emerging Technologies for Communications
2020
Session Number:B2
Session:
Number:B2-2
Received Power Prediction using Multiple RGB cameras in Indoor Environment for Wireless LAN Systems
Kahoko Takahashi, Riichi Kudo, Tomoaki Ogawa,
pp.-
Publication Date:2020/12/2
Online ISSN:2188-5079
PDF download
Summary:
This paper proposes a received power prediction scheme that uses deep-neural-network based camera image object detection in indoor environments. The proposed scheme uses two-step machine learning. The first block detects the connected device using multiple camera images and the second block predicts the received signal strength index (RSSI) using the detected object’s information. To evaluate its performance, an autonomous locomotion robot was developed and indoor experiments were conducted using a wireless LAN system with 5.6 GHz channel. Experiments show that the proposal matched or bettered the performance of conventional robot position-based prediction regardless of lead time.