From template analysis to generating partitions I: Periodic orbits, knots and symbolic encodings

TL;DR

This paper introduces a robust algorithm for constructing symbolic encodings of chaotic attractors in three-dimensional flows, leveraging topological analysis of periodic orbits and template theory, suitable for noisy experimental data.

Contribution

It develops a new method combining topological analysis and triangulations to generate partitions from periodic orbits, improving robustness and applicability to experimental data.

Findings

Successfully applied to over 1500 periodic orbits from simulations

Achieved partition border localization with 0.01% precision

Method is robust to noise and suitable for experimental time series

Abstract

We present a detailed algorithm to construct symbolic encodings for chaotic attractors of three-dimensional flows. It is based on a topological analysis of unstable periodic orbits embedded in the attractor and follows the approach proposed by Lefranc et al. [Phys. Rev. Lett. 73, 1364 (1994)]. For each orbit, the symbolic names that are consistent with its knot-theoretic invariants and with the topological structure of the attractor are first obtained using template analysis. This information, and the locations of the periodic orbits in the section plane, are then used to construct a generating partition by means of triangulations. We provide numerical evidence of the validity of this method by applying it successfully to sets of more than 1500 periodic orbits extracted from numerical simulations, and obtain partitions whose border is localized with a precision of 0.01%. A distinctive advantage of this approach is that the solution is progressively refined using higher-period orbits, which makes it robust to noise, and suitable for analyzing experimental time series. Furthermore, the resulting encodings are by construction consistent in the corresponding limits with those rigorously known for both one-dimensional and hyperbolic maps.