Chirp transform in the nonlinear tracking performance analysis of the LMS adaptive predictors

A chirp transform is defined for the /spl Delta/-step LMS adaptive predictors for linearly chirped signals embedded in additive white Gaussian noise. By converting the chirped signals to stationary baseband signals, this transform provides a different approach in analyzing the tracking performance of the LMS adaptive predictors. This transform also provides an approach for analyzing the nonlinear effects of the LMS adaptive predictor for nonstationary input signals. It is also shown that the chirp transform can be applied to the 1-step RLS predictor with chirped input signals.