Processing of Noise Radar Waveforms using Block Least Mean Squares Algorithm

The problem of signal reception in a noise radar is considered. It is assumed that heavy clutter reflections are present. In such conditions, correlation processing suffers from the problem of noise floor in the ambiguity function of random signals. It is proposed to replace correlation processing with least squares (LS) based methods. In particular, a highly efficient block least mean squares (block LMS) algorithm is studied in detail. It is demonstrated that, compared with the correlation processing, block LMS shows significant advantages in terms of robustness to clutter. Simulation experiments confirm that when heavy clutter is present, the proposed method outperforms correlation processing significantly.