Summary
International Conference on Emerging Technologies for Communications
2020
Session Number:D3
Session:
Number:D3-3
Fast Approximation of the Time Delay Estimation using Persistent Homology
Koya Oyama, Riki Watabe, Hiroyuki Kamata,
pp.-
Publication Date:2020/12/2
Online ISSN:2188-5079
PDF download
Summary:
In nonlinear analysis, time delay and embedding dimension are estimated, and a phase space trajectory called an attractor is constructed. The estimation accuracy of these parameters is linked to the analysis accuracy. The problem with using the conventional method for estimating the time delay is that the geometrical shape of the attractor, which is important in the construction of the attractor, is not considered. Therefore, we have proposed a method that considers geometrical properties of the attractor by using persistent homology for estimating the time delay. However, the calculation cost with this method is high. Consequently, we propose a method that can approximate the result when persistent homology is used by using a simpler algorithm. The results show that the calculation cost is greatly reduced, and the time delay can be estimated in consideration of the geometric shape.