An application of the LMS algorithm in smoothing pseudo-Wigner distribution

It is shown that the least-mean-square (LMS) adaptive algorithm, if implemented in the frequency domain, can separate the autoterms from the cross terms in Wigner-Ville distributions. This separation is achieved over the process stationary interval, and as such, is more evident in a slowly varying environment. It is demonstrated that the LMS simulates an exponentially smoothed pseudo-Wigner distribution. The time constant of the smoothing window is inversely proportional to the step-size parameter used in the algorithm.<
>