Universal Adaptive Beamforming
Abstract
This project proposes exploiting universal algorithms, order statistics and Random Matrix Theory (RMT) to address the challenges posed for adaptive beamformers by large arrays in dynamic environments. The technical approaches proposed fall within two themes. First, we will extend our recent research applying universal algorithms from information theory to important array processing challenges. Second, we propose developing approximate implementations of optimal adaptive beamformers (ABFs) that sacrifice a very small amount of performance to achieve a substantial reduction in computational requirements. These tradeoffs between performance and computation requirements will become critical in future passive sonar systems containing very large numbers of sensors. We propose five main technical challenges for investigation. First, we propose exploiting context tree weighting universal algorithms for generalized sidelobe canceling ABFs to develop computationally efficient ABFs that rival the performance of the optimal Capon beamformer with far less computation for very large arrays. Second, we propose developing a universal ABF across subaperture size for large aperture arrays. This UABF will vary the coherent processing aperture in response to changes in wavefront coherence across the array. Third, we will expand our recent covariance matrix taper universal ABF to develop ABFs that create wider notches for moving interferersbut are more economical with degrees of freedom, resulting in better gain against white noise. Fourth, we will adapt the Algorithmic Noise Tolerance framework from VLSI design to beamformers, resulting in beamformers that intelligently "fall back" from complicated adaptive beamformers to simpler beamformers when challenging environmental conditions or distorted array shape result in degradedperformance. Fifth, we will generalize Bartlett-Welch periodogram averaging by exploiting linear order-statistic filters to designspectral estimationalgorithms which are more robust to outliers in the data while maintaining nearly optimal performance when thereis no mismatch present. We plan to test our new algorithms on real data from the UMass Dartmouth Signal Processing Group s currentmicrophone array and also existing ocean acoustic data sets from prior ONR-funded ocean acoustic experiments where possible.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 12, 2023
- Source ID
- N000142312133
Entities
People
- John R. Buck
Organizations
- Office of Naval Research
- United States Navy
- University of Massachusetts