A Sliding Mode-based Filter for Non-stationary Signals

This paper presents a sliding mode-based filter for estimating the oscillation amplitude of non-stationary signals. The proposed algorithm makes use of the structure of sliding mode observers to ensure the estimation of the orthogonal component of a non-stationary signal, when the frequency of the trigonometric function varies over time. A stability analysis is provided to show the filter’s ability to remove the DC components, and numerical simulations illustrate its performance under noisy conditions.