Estimation of chirp radar signals in compound-Gaussian clutter: a cyclostationary approach

Signal detection of known (within a complex scaling) rank one waveforms in non-Gaussian distributed clutter has recently received considerable attention. We expand on published solutions to consider the case of rank one waveforms that have some unknown parameters, i.e., signal amplitude, initial phase, Doppler shift, and Doppler rate of change. The contribution of this paper is the derivation and performance analysis of two joint estimators of Doppler shift and Doppler rate—the chirp parameters—for a signal with fluctuating random amplitude embedded in correlated compound-Gaussian clutter. One solution is based on the maximum likelihood (ML) principle and the other one on target signal second-order cyclostationarity. The hybrid Cramér–Rao lower bounds (HCRLB’s) and a large sample closed-form expression for the mean square estimation error (only for the Doppler shift) are also derived. Numerical examples are provided to show the behavior of the proposed estimator under different non-Gaussian clutter scenarios.