Detection, Estimation and Beamforming for Adaptive Sensor Arrays: Algorithms and Performance

Abstract

ABSTRACT

A class of adaptive detection and estimation algorithms has emerged over the past thirty years that exploit the spatial and temporal diversity available from sensor array systems in order to provide robust signal detection and parameter estimation under rather adverse/non-ideal conditions. These arrays are often deployed in high multipath environments plagued by limiting interference with unknown statistics. A uniformly most powerful test does not exist for this class of problems. Consequently, optimal detection and estimation rely heavily upon maximum-likelihood (ML) estimates of unknown parameters including use of data sample covariance matrix. Analyses embracing practicalities such as finite sample support, array response uncertainty/ mismatch, non-stationarity and nonlinear parameter estimation are quintessential for the design of systems requiring precision and robustness, e.g., adaptive radar/sonar systems. In this talk, we present an overview analysis of this class of adaptive algorithms addressing the aforementioned issues of practical interest via the use of random matrix theory. Specifically, the receive operation characteristics are considered for the detector class that includes the adaptive matched filter (AMF), Kelly/Khatri’s generalized likelihood ratio test (GLRT), Conte/Scharf’s adaptive coherence estimator (ACE), and the 2-D adaptive sidelobe blanker (ASB). The mean squared error (MSE) performance of the signal parameter estimation class that includes the nonlinear ML estimator and the Capon-minimum variance distortionless response (MVDR) beamformer/estimator often used for frequency and/or angle estimation is considered. The MSE performance is considered below threshold where local error performance bounds, like the Cramér-Rao bound, are not useful. Lastly, some discussion of robust sample covariance-based adaptive beamforming is provided and new results/insights on the statistical relationships between conventional and adaptive processing is presented.

BIOGRAPHY

Christ D. Richmond received his a B.S. in Electrical Engineering from the University of Maryland in College Park, and a B.S. in Mathematics from Bowie State University. He received his S.M., E.E., and Ph.D. degrees in electrical engineering from the Massachusetts Institute of Technology (MIT), Cambridge.

He is currently a member of the technical research staff at MIT Lincoln Laboratory, Lexington. His research interests include detection and parameter estimation theory, sensor array and multichannel signal processing, statistical signal processing, random matrix theory, radar/sonar signal processing, multivariate statistical analysis, information theory, multiuser detection, multi- input multi-output (MIMO) systems, and wireless communications.

Dr. Richmond is the recipient of the Office of Naval Research Graduate Fellowship Award 1990–1994, the Alan Berman Research Publications Award March 1994 (Naval Research Laboratory), and the IEEE Signal Processing Society 1999 Young Author Best Paper Award in area of Sensor Array and Multi-channel (SAM) Signal Processing.

He served as the Technical Chairman of the Adaptive Sensor Array Processing (ASAP) Workshop, MIT Lincoln Laboratory, 2007, 2006, and 1998, and served as a member the IEEE Technical Committee on SAM Signal Processing, and as an Associate Editor for the IEEE Transactions on Signal Processing 2002--2005. He was an invited reviewer for the book Bayesian Bounds for Parameter Estimation and Nonlinear Filtering/Tracking, by Prof. Harry L. Van Trees and Prof. Kristine Bell of George Mason University, Editors, IEEE Press 2007.