elling and Estimation for Doppler-shifted Gaussian Random Processes
Christophe Couvreur, Yoram Breslert · 1996
We address the problems oj modeling Doppler-shifted wide-band Gaussian random processes and of estimating the Doppler parameter from a jinite series of discrete-time samples. Relations between the continuous-time process, the Doppler shift parametel; and the discrete-time process obtuined by sampling are estublished. Appsoxinznte rutionul models are proposed. Various estimators are proposed for Doppler parameter when the second-order statistics of the original continuous-time random process are known. The Cramdr-Rao hound is derived. The estimators are compared expesinientully on synthetic Doppler-shifted data. We also hint at some extensions ofthe method to non-stationay processes and time-varying Doppler shifts.