Phase reconstruction from oscillatory data with iterated Hilbert transform embeddings -- benefits and limitations (Q2077830)

The authors apply Iterated Hilbert Transform Embeddings (IHTEs) to explore the phase reconstruction from scalar observables associated to oscillatory dynamical systems. As an illustration, they consider the stable self-oscillations of the Stuart-Landau system under quasiperiodic external forcing. The authors explore a range of frequencies of the forcing and of the limit cycle stability parameter. The main results are: 1. In the case of purely phase-modulated signals, the true phase can be nearly perfectly reconstructed; 2. In the more realistic situation when modulations of the amplitude and of the phase are present, the reconstruction is moderate if any. In general the authors verify that in the cases of strong stability of the oscillations and high frequency of the forcing, the IHTEs significantly improve the quality of the reconstructed phase. On the other hand they show that for a significant amplitude modulation the IHTEs do not provide any improvement with respect to other techniques.