A minimum norm array processing algorithm for super resolution target profiling
Andrew J. Willis, Robert de Mello Koch · 2002
The problem of target characterisation by phased array sonar is approached as one of spectral estimation of a band limited function for a finite number of arbitrarily located samples. A unique solution is extracted from the resultant under-determined system by application of singular value decomposition techniques to derive a minimum norm, or best least squares estimator of the target profile. The algorithm produced using this method represents an advance on previous array processing techniques designed to identify a limited number of point targets through parametric search. The more rapid process of a single matrix decomposition is employed to give a minimum norm estimate applicable to an arbitrary target profile - this being derived, in principle, from a single snapshot received at an array unrestricted by number or distribution of sensors. Furthermore if a priori knowledge in target location or shape is available this can be incorporated to additionally enhance the quality of the reconstruction, providing for super resolution of certain features.>