Summary
International Conference on Emerging Technologies for Communications
2020
Session Number:I1
Session:
Number:I1-3
On Radio Propagation Estimation based on Deep Learning with Convolution and Self-attention
Riku Hashimoto, Katsuya Suto,
pp.-
Publication Date:2020/12/2
Online ISSN:2188-5079
DOI:10.34385/proc.63.I1-3
PDF download
Summary:
In this paper, we address the challenge of radio propagation estimation in unmeasured areas. Our envisioned model is designed based on state-of-the-art deep learning, i.e., convolution and self-attention networks, to estimate the shadowing effect in radio propagation from the three-dimensional building map. We demonstrate that self-attention successfully estimates the shadowing effect and improves estimation accuracy.