Time-frequency representation for time-varying signals using a Kalman filter

A new method which uses a Kalman filter to obtain a time-frequency representation for time-varying signals is introduced. In this method, a time-varying signal is modeled as a time-varying AR process whose parameters determine the instantaneous power spectral density (IPSD). Then, a Kalman filter is used to estimate the time-varying parameters which are used to compute the estimated IPSD. From simulation results, it is concluded that a good estimate of the IPSD is obtained with a 2nd order variation model of the time-varying parameters.