A minimum norm approach to target classification using phased array radar

An algorithm is proposed that uses a phased array radar to profile any number of arbitrary targets as a function of bearing. The algorithm uses singular value decomposition to construct a pseudoinverse pertinent to the derivation of a minimum norm or best least squares estimator of the associated spectrum. Some results of application to the target identification problem for a broadside phased array radar are given in the case of some salient profiles, and the results are discussed with regard to the inclusion of prior knowledge, particularly with respect to noise statistics.<>