H∞ filtering for autoregressive modeling based Space-Time Adaptive Processing

Space-Time Adaptive Processing (STAP) is now commonly used in radar engineering to detect the targets by using a phased array antenna system. However, the computational cost of the standard version and the memory storage are high. In addition, the detection could be more robust against interfering targets. To solve the above problems, autoregressive (AR) modelling of the disturbances, namely the sea clutter and the additive thermal noise, leads to a variant of the STAP. In that case, the key issue is the estimations of the multichannel AR process from the secondary data, i.e. the data received when analyzing the “cells” in the neighbourhood of the area under study. Off-line methods have been proposed, but they require a large number of secondary data. To reduce it, on-line method can be considered. Nevertheless, since the clutter has a K-distributed amplitude distribution, the Gaussian assumptions necessary to use Kalman filtering do not hold. To relax them, we suggest investigating the relevance of H∞ algorithm in this paper.