An Efficient Real-Time Method for Time-Varying Filter Region of Support Estimation

This paper outlines a time-frequency (TF) analysis based method for time-varying (TV) filter´s region of support (FRS) estimation. The proposed method, applied to a cross-terms-free TF representation that additionally provides high quality instantaneous frequency (IF) estimation, enables multiple detection of the local FRS in the observed increment of time. In this way, the method provides a very efficient real-time filtering of nonstationary mono- and multicomponent frequency-modulated (FM) signals. Moreover, real-time implementation of the TV filter, based on the proposed method, optimizes hardware requirements and produces comparable computational cost regarding the existing TV filters´ designs.