A novel method for signal component incision in the time-frequency plane

Abstract

A time-frequency signal analysis technique for detection and extraction of signal components is developed. The proposed technique is observed to be successful even under high noise levels, for signals whose components have compact support in the time-frequency plane. It is an iterative algorithm which detects and extracts one component at a time. The robust estimates of instantaneous frequency which are attained by principle curve projections, have increased the performance. The performance of the algorithm is analyzed under different noise levels using synthetic signals.