摘要

A generalized adaptive subspace detector for range-Doppler spread target (RDST-GASD) in the non-Gaussian clutter is derived in this paper. The subspace model of multi-pulse wideband radar target returns is established in the frequency-slow time domain. The clutters are modeled as nonhomogeneous spherically invariant random vectors (SIRVs); that is, the power of the clutter is different from one range cell to another. The clutter covariance matrix is estimated with the secondary data. The constant false alarm rate (CFAR) property of RDST-GASD with respect to both the power and the covariance matrix of the clutter is demonstrated theoretically. Considering that there is target range walking across range cells during a coherent processing interval (CPI) for wideband radar, the RDST-GASD does range alignment to the multiple returns of the target in a CPI. As a result, the coherent integration is implemented and the detection performance is improved.