An adaptive detection of spread targets in locally Gaussian clutter using a long integration time

Abstract

This paper deals with the problem of detecting a collision target in ground clutter, using a long integration time. A single reception channel being available, classical space time adaptive processing (STAP) cannot be used. After range processing, ground clutter can be modeled as a known interference subspace in the Doppler domain depending on its radial and orthoradial speeds. We exploit this a priori knowledge to perform an adpative detection of a collision target supposed to lie in a known and different subspace. A GLRT detector is first derived for known clutter covariance matrix. Then, the unknown covariance matrix is adaptively estimated from the projection of the data onto the modeled clutter subspace, and is plugged in the GLRT to form a suboptimal detector. The proposed scheme can be viewed as a synthetic STAP, for which the space domain is replaced by a clutter orthoradial information and longer integration time.