Blind Doppler estimation for satellite with interference EMI

Abstract

Doppler frequency estimation in satellite communications plays an important role in the geolocations of electromagnetic interference (EMI) emitters. Due to the high speed of satellites and various trajectories of the orbits, one cannot assume that the Doppler frequency is constant. The Doppler estimations are usually performed repeatedly for each given time duration, within which the Doppler frequency can be modeled as a chirp. Doppler estimation algorithms based on the phase requires the transmitted signals removed from the received signal. This is hard to do in practical systems, since the transmitted signals from EMI source are unknown. In this paper, blind doppler frequency estimation is considered for unknown EMI signals using frequency analysis. After the received signal at satellite is down converted, the estimation is performed for each given time instants at different signal-to-noise ratios. To reduce the estimation errors at very low signalto-noise ratio (SNR), the extremely bad estimates are detected as outliers and the estimates are replaced by those at previous time instants.