Block Least Mean Squares processing of noise radar waveforms

Noise radars usually employ correlation processing of waveforms. However, vulnerability to clutter is a serious disadvantage of this approach. This paper considers using Least Squares (LS) based methods. In particular, highly efficient Block Least Mean Squares (Block LMS) algorithm is studied in details. The formula for integration gain of Block LMS is derived. Compared to analogue quantity for correlation processing, it shows significant advantage of the proposed solution in terms of robustness to clutter. The Doppler response of the algorithm is analyzed, which - under proper choice of algorithm parameters - is identical to that of correlation approach. Simulation experiments confirm that when heavy clutter is present, the proposed method outperforms correlation processing significantly.