A viterbi algorithm-based IF estimator for overlapped multicomponent signal

Abstract

Instantaneous frequency (IF) estimation of multicomponent signals with time frequency (TF) overlapped components is a challenging task in radar, sonar and other applications. Viterbi algorithm (VA) on TF distribution can be potentially applied to estimate IFs of multicomponent signals, but it may track wrong IFs when signal components intersect each other in the TF domain. In order to suppress this switch problem in the original VA, this paper assumes the linearity of IFs in the overlapped regions remains unchanged, then, a novel penalty function in the modified VA based on linear least square fitting technique using adjacent several TF points is developed. Performance of the proposed algorithm on several artificial multicomponent FM signals indicates the proposed algorithm can obtain more accurate IFs by suppressing the switch problem.