Near Optimal Adaptive Robust Beamforming

The performance degradation in traditional adaptive beamformers can be attributed to the imprecise knowledge of the array steering vector and inaccurate estimation of the covariance matrix. The inaccurate estimation of the covariance matrix is due to the limited data samples and presence of desired signal components in the training data, especially when the desired signal is the dominant signal in the training data. The mismatch between the actual and the presumed steering vectors can be due to the error in the position (geometry) and/or in the look direction estimate. Due to mismatch in the actual and the presumed steering vectors and the inaccurate estimation of covariance matrix, the output SINR of the adaptive beamformers cannot adapt to the increase in the input SNR. In other words the output SINR of the adaptive beamformers saturates as the input SNR increases. From the following graphs, it can be observed that the proposed algorithms in the following references were able to overcome the effect of saturation. The results of the proposed algorithms were compared with some of the state-of-the-art algorithms such as RCB [1], IRCB [2], SQP [3-5]. We present the performance graphs of different algorithms presented in our papers in terms of output SINR for different input SNR.