Enhanced Minimum Variance Beamforming

A Competitive Mean-Squared Error Approach to Beamforming

“…The problem stems from the fact that the bias, and therefore the MSE of (9) is a function of . One approach to eliminating this dependency is to force the bias to 0 and then minimize the MSE, which leads to the problem (8). However, this does not guarantee a small MSE, so that on average the resulting estimate may be far from .…”

“…This beamformer can alternatively be obtained as the solution to subject to (8) which, as we show below, has the interpretation of minimizing the MSE subject to the constraint that the beamformer is unbiased. Although this approach has several optimality properties, it does not necessarily result in a good signal estimate.…”

“…Some beamforming techniques are designed to mitigate this effect [5]- [8], whereas others are developed to also overcome uncertainty in the steering vector [9]- [13], [14]. Here, we assume that the steering vector is known exactly and our aim is to design a beamformer for signal estimation.…”

Mean‐Squared Error Beamforming for Signal Estimation: A Competitive Approach

“…is a zero-mean, Gaussian, complex random vector, temporally and spatially white, [33,34], loading Capon beamformer (L-CAPON) [9,10], eigenspace-based beamformer (EIG) [11,12], and the robust beamformers of (1.10), (1.11) and (1.12) which we refer to, respectively, as ROB1, ROB2, and ROB3 [14,15,16]. In Example 2 we compare our methods against the principal eigenvector beamformer [35], which we refer to as PEIG, with R given by (1.8) and L-CAPON The parameters of each of the compared methods were chosen as suggested in the literature.…”

Prewhitening of colored noise fields for detection of threshold sources

“…"e I 2 and Minimum Norm (Min-Norm) 3 compute angular response as a function of cosine of bearing angle. These methods associate peaks in the response function with source locations, providing direction of arrival estimates.…”