Low‐complexity methods for joint delay and Doppler estimation of unknown wideband signals

Abstract

A joint time delay and Doppler shift estimation of unknown wideband signals is studied in the sense that the variation of time delay over time and the migration of Doppler shift over signal frequency cannot be neglected. By partitioning the received signals into short time segments, a novel wideband signal model is proposed. Based on this model, the maximum likelihood estimate (MLE) of the time delay at the starting point of observation and the Doppler shift at the signal carrier frequency is derived. With well compensation of the migrations of these parameters, the MLE derived could achieve higher estimation accuracy. To lower the computational complexity, a fast estimation algorithm based on the Keystone transform is also presented. After correcting the migrations through the Keystone transform, successive one-dimensional Fourier transform, instead of two-dimensional grids search, is used to achieve coherent processing of the received signals. The Cramer-Rao lower bound analysis for wideband cases is presented and the computational complexity of the proposed algorithms is analysed. Simulation results demonstrate the superiority of the proposed methods for wideband cases.