Linearly Constrained Minimum-Variance Algorithm for Radar Jammer and Clutter Suppression

We propose an augmented linearly constrained minimum-variance algorithm for radar jammer and clutter suppression. The basic LCMV algorithm is augmented to extend the simple target detection capability to target classification. In general, received radarreturns are comprised of a linear superposition of the target return signal and mainbeam and sidelobe jammer signals and clutter returns, which corrupt or obscure the desired signal. One of the objectives of signal processing algorithms is to extract andamplify the target returns and reject or suppress the jammer signals and clutter. We propose to develop and apply a real-time algorithm for isolating target signals from jammer and clutter interference, while minimizing the output variance of the receiverresponse to the latter. The signal-to-noise ratio (SNR)is expressed as a Raleigh quotient and its optimization leads to a generalized eigenvalue problem. The solution yields a linear transformation which, when applied to the receiver response, maximizesthe SNR. The result is the strongest (in the SNR sense) signal as far removed from jammer signals and clutter as possible, for maximum detectability. The results are expanded to enable target classification. The algorithm we propose has potentialapplications in remote sensing tasks in government and private industry, planetary surveying and mapping, and in private and commercial civilian aviation, as well as in military scenarios.