A Bayesian approach to robust adaptive beamforming

Bell, Kristine L.; Ephraim, Yariv; Trees, Harry L. Van

doi:10.1109/78.823966

Cited by 196 publications

(147 citation statements)

References 56 publications

(61 reference statements)

Supporting

Mentioning

147

Contrasting

Order By: Relevance

“…Unlike that of the exactly known signal steering vector, the performance of the DL-SMI beamformer degrades severely especially at high SNR because of DOA mismatch. Because the steering vector uncertainties in this example is caused by the DOA mismatch only, the Bayesian beamformer [23] performs as well as the proposed robust beamformer. …”

Section: Example 2: Signal Look Direction Mismatchmentioning

confidence: 87%

“…The proposed robust beamformer performs almost as well as the former two scenarios, whereas the others are not. Because the equivalent DOAs of the array sensors are not uniform, namely, the steering vector uncertainties are not caused by the DOA uncertainty only, the Bayesian beamformer [23] loses it's robustness, especially at high SNR. Such phenomena also occur in the worst-case based robust beamformer [14] and the DL-SMI beamformer (10).…”

Section: Example 3: Random Perturbation Of the Elements In A Steeringmentioning

confidence: 99%

“…(20), the SMI beamformer (8), the DL-SMI beamformer (10), the eigenspace-based beamformer [13], the Bayesian beamformer [23], the worst-case based robust beamformer [14], and the Bayesian beamformer with order recursive implementation [18]. In addition, the optimal SINR is also shown in all figures.…”

Section: Simulationsmentioning

confidence: 99%

“…In the proposed robust beamformer, the initialized DOA of each sensor is set to θ s , and the Θ is composed of L = 11 evenly spaced points at {−5.739, −4.589, −3.440, −2.292, −1.146, 0, 1.146, 2.292, 3.440, 4.589, 5.739} + θ s , where θ s is the presumed DOA. The Θ is also adopted in the Bayesian beamformer [23]. The a priori pdf is uniform with p(θ) = 1 11 .…”

Section: Simulationsmentioning

confidence: 99%

“…The a priori pdf is uniform with p(θ) = 1 11 . The parameter γ is set to 0.3, as [23] denotes. In all of the beamformers, K = 50 snapshots are used.…”

Section: Simulationsmentioning

confidence: 99%

See 4 more Smart Citations

Robust Adaptive Beamforming for Steering Vector Uncertainties Based on Equivalent Doas Method

Gu¹,

Shi²,

Chen³

et al. 2008

PIER

View full text Add to dashboard Cite

Abstract-The adaptive beamformers often suffer severe performance degradation when there exist uncertainties in the steering vector of interest. In this paper, we develop a new approach to robust adaptive beamforming in the presence of an unknown signal steering vector. Based on the observed data, we try to estimate an equivalent directionof-arrival (DOA) for each sensor, in which all factors causing the steering vector uncertainties are ascribed to the DOA uncertainty only. The equivalent DOA of each sensor can be estimated one by one with the assumption that the elements of the steering vector are uncorrelated with each other. Using a Bayesian approach, the equivalent DOA estimator of each sensor is a weighted sum of a set of candidate DOA's, which are combined according to the value of the a posteriori probability for each pointing direction. In this way, the signal steering vector and the diagonal loading sample matrix inversion (DL-SMI) version adaptive beamformer can be obtained. Numerical simulations illustrate the robustness of the proposed beamforming algorithm. 278Gu et al.

show abstract

Section: Example 2: Signal Look Direction Mismatchmentioning

confidence: 87%

Section: Example 3: Random Perturbation Of the Elements In A Steeringmentioning

confidence: 99%

Section: Simulationsmentioning

confidence: 99%

Section: Simulationsmentioning

confidence: 99%

“…The a priori pdf is uniform with p(θ) = 1 11 . The parameter γ is set to 0.3, as [23] denotes. In all of the beamformers, K = 50 snapshots are used.…”

Section: Simulationsmentioning

confidence: 99%

See 3 more Smart Citations