Iterative Robust Capon Beamformer

Many adaptive beamforming algorithms deteriorate drastically in practical applications because they are highly sensitive to mismatches between exploited assumptions and the actual characteristics of the sensor array, environment and sources. Recently, the Robust Capon Beamformer (RCB) is derived which utilizes the uncertainty set of the desired array steering vector and is capable of handling arbitrary array steering vector mismatches. However, when large steering vector mismatches occur, the uncertainty set has to be expanded to accommodate for the increased uncertainty and this degrades the output signal-to-interference-plus-noise ratio (SINR) of the RCB. In this paper, an iterative RCB is proposed which makes use of a small uncertainty sphere to search for the desired array steering vector. This proposed method provides higher SINR than the conventional RCB by avoiding the use of a large uncertainty search sphere. Unlike the conventional RCB, this iterative RCB does not require knowledge of the uncertainty level in the desired array steering vector. Theoretical analysis and computer simulations illustrate the effectiveness of the proposed beamformer.