Presentation Information
[B-1A-16]A Study on Delayed Path Estimation Using Machine Learning
— Characteristics of Path Existence Probability and Received Power —
◎△Yuta Kiriyama1, Tetsuro Imai1, Koshiro Kitao2, Satoshi Suyama2 (1. Tokyo Denki University, 2. NTT DOCOMO, INC.)
Keywords:
Machine Learning,Wireless Communication,Delayed Path Estimation,Path Existence Probability
In the B5G/6G systems currently under investigation, further improvements in power efficiency are required along with spectral efficiency, which calls for highly accurate propagation estimation models. To this end, the authors have proposed a spatio-temporal path estimation model based on machine learning. However, sufficient estimation accuracy has not yet been achieved. In this paper, we extend the model so that it can also output the path existence probability, and we report an evaluation of the path estimation accuracy using delay measurement data obtained in an indoor environment.