Separation of chaotic signals by reservoir computing

Abstract

We demonstrate the utility of machine learning in the separation of superimposed chaotic signals using a technique called reservoir computing. We assume no knowledge of the dynamical equations that produce the signals and require only training data consisting of finite-time samples of the component signals. We test our method on signals that are formed as linear combinations of signals from two Lorenz systems with different parameters. Comparing our nonlinear method with the optimal linear solution to the separation problem, the Wiener filter, we find that our method significantly outperforms the Wiener filter in all the scenarios we study. Furthermore, this difference is particularly striking when the component signals have similar frequency spectra. Indeed, our method works well when the component frequency spectra are indistinguishable—a case where a Wiener filter performs essentially no separation.

Document Details

Document Type
Pub Defense Publication
Publication Date
Feb 01, 2020
Source ID
10.1063/1.5132766

Entities

People

  • Brian R. Hunt
  • Edward Ott
  • Michelle Girvan
  • Sanjukta Krishnagopal

Organizations

  • National Science Foundation
  • Santa Fe Institute
  • United States Department of Defense
  • University of Maryland

Tags

Fields of Study

  • Engineering

Readers

  • Image Processing and Computer Vision.
  • Neural Network Machine Learning.
  • Wave Propagation and Nonlinear Chaotic Dynamics.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms