Detecting the onset of bifurcations and their precursors from noisy data

We study the problem of the detection of noise-induced precursors of periodic motion instabilities in stochastic dynamical systems. In particular, we concentrate on the period-doubling bifurcation. We have developed a statistical method to detect the onset of bifurcations and their precursors based on the previously established topological recurrence technique.