University Events

Description: Topic: Source Enumeration, Localization and Spectral Estimation Using Co-Prime and Other Sparse Sensor Arrays

Location: Lester W. Cory Conference Room, Science & Engineering Building (SENG) Room 213A

Abstract:
Sparse arrays often refer to a class of non-uniformly spaced line arrays, where the inter-element spacings are integer multiples of the half spatial wavelength of the impinging signals. Popular sparse arrays such as minimum redundancy arrays, coprime arrays and nested arrays have broad applications in radio astronomy, radar, sonar, and wireless communication systems, including source enumeration, detection, direction-of-arrival (DOA) estimation and spatial power spectral density (PSD) estimation. Through array augmentation, a sparse array with O(N) elements is capable of localizing O(N2) sources by exploiting the second-order statistics of the propagating field. When used for beamforming, sparse arrays potentially achieve the resolution of a fully populated uniform linear array of comparable aperture using many fewer sensors at the expense of much higher sidelobes. To achieve these performances, sparse array processing techniques often require large numbers of snapshots to ensure more accurate estimates of the signal spatial correlations or the spatial PSDs. The large numbers of snapshots required may be available in electromagnetic scenarios, but are often unrealistically optimistic for many acoustic environments, largely due to the speed of field propagation, use of large array aperture and the field being non-stationary.

This dissertation explores several research directions and proposes new algorithms improving the source enumeration, detection and estimation performances for passive sparse sensor array systems. This dissertation makes three major contributions. The first focuses on the coprime sensor arrays (CSA) and proposes a new processor, termed the min processor, as an alternative to the more popular product processor. Compared with the product processor, the min processor has many attractive features such as lower sidelobes and positive semi-definite spectra while maintaining the same array resolution. These features improve a CSA's capability in Gaussian source detection and spatial PSD estimation. Secondly, the CSA product and min processors are extended for correlation processing both temporally narrowband and wideband sources by exploiting the Fourier relationship between the PSD and the correlation function. The statistical properties of the PSD estimates using the min processor propagate to the correlation estimates and benefit high-resolution DOA estimation. Finally, this dissertation proposes new coherent wideband subspace processing algorithms for enumerating and estimating the DOAs of more sources than sensors using any sparse arrays. The new algorithms extend the concepts of periodogram averaging and spatial resampling originally proposed for ULA applications to sparse array applications. By reinforcing the sources' spectral information through frequency averaging, the proposed algorithms achieve great performance in wideband source enumeration and DOA estimation, especially in low SNR and snapshot limited scenarios.

Note: All ECE Graduate Students are ENCOURAGED to attend.
All interested parties are invited to attend. Open to the public.

Committee Members: Dr. David A. Brown and Paul J. Gendron, Department of Electrical & Computer Engineering; Dr. Kathleen E. Wage, Department of Electrical and Computer Engineering, George Mason University.

*For further information, please contact Dr. John Buck at 508.999.9237, or via email at jbuck@umassd.edu.