Multicomponent signal analysis using the polynomial-phase transform

We consider the problem of estimating signals consisting of multiple components of the form b(n)exp{Jφ(n)} where b(n) is a constant or a slowly varying amplitude, and the φ(n) is a polynomial function of time. We present an estimation algorithm based on the discrete polynomial transform and illustrate its performance by several examples. The proposed algorithm is computationally much simpler than the maximum likelihood estimator for this problem.