Roy L. Streit, Ph.D.
RESEARCH PAPERS

1. Tracking with Distributed Sensors Using Geometric Invariants (with T.A. Wettergren and J.R. Short), IEEE Transactions on Aerospace and Electronic Systems, 40 (2004), 1366-1374. [.pdf]

2. A Numerical Optimization Approach to Acoustic Hull Array Design (with T.A. Wettergren and J.P. Casey), Journal of the Acoustical Society of America, 112 (2002), 2735-2741.[.pdf]

3. Hidden Gauss-Markov Models for Signal Classification (with P.L. Ainsleigh and N. Kehtarnavaz), IEEE Transactions on Signal Processing, SP-50 (2002), 1355-1367. [.pdf]

The following two papers extend the Probabilistic Multi-Hypothesis Tracking (PMHT) algorithm to include Bayesian classification, and to utilize sensor data without peak picking.

4. Tracking, Association, and Classification: A Combined PMHT Approach (with S. Davey and D. Gray), Digital Signal Processing, 12 (2002), 372-382. [.pdf]

5. Multitarget Tracking of Distributed Targets Using Histogram-PMHT (with M.L. Graham and M.J. Walsh), Digital Signal Processing, 12 (2002), 394-404. [.pdf]

6. PMHT: Problems and Some Solutions (with P.K. Willett and Y. Ruan), IEEE Transactions on Aerospace and Electronic Systems, AES-38 (2002), 738-754. [.pdf]

The following paper incorporates transmission loss models into target localization problems, and treats uncertainty in the model by statistical inflation.

8. Bearings-Only Target Motion Analysis With Acoustic Propagation Models of Uncertain Fidelity (with M.J. Walsh), IEEE Transactions on Aerospace and Electronic Systems, AES-38 (2002), 1122-1137. [.pdf]

8. Detection of Long-Duration Narrowband Processes (with Z. Wang and P.K. Willett), IEEE Transactions on Aerospace and Electronic Systems, AES-38 (2002), 211-227.[.pdf]

The next paper derives the recognition differential for passive sonar using a discrete-continuous Poisson process to model the number and spatial distribution of interferers.

9. Detection of Gaussian Signals in Poisson Modulated Interference, Journal of the Acoustical Society of America, 108 (2000), 1752-1762.[.pdf]

The following paper develops a new method for detecting transients and spread spectrum signals using a missing data model and the expectation-maximization method.

10. Detection of Random Transient Signals via Hyperparameter Estimation (with P.K. Willett), IEEE Transactions on Signal Processing, SP-47 (1999), 1823-1834. [.pdf]

The following papers estimate Gaussian mixtures from sample data, applications to neural networks, and the use of cross-class pooling to achieve robustness with small data sets.

11. Estimation of Gaussian Mixtures With Rotationally Invariant Covariance Matrices (with T.E. Luginbuhl), Communications in Statistics: Theory and Methods, 26 (1997), 2927-2944. [.pdf]

12. Maximum Likelihood Training of Probabilistic Neural Networks (with T.E. Luginbuhl), IEEE Transactions on Neural Networks, NN-5 (1994), 764-783. [.pdf]

The following paper presents the first application of hidden Markov models to frequency line tracking, and to automatic classification of transient and other non-stationary signals.

13. Frequency Line Tracking Using Hidden Markov Models (with R.F. Barrett), IEEE Transactions on Acoustics, Speech, and Signal Processing, ASSP-38 (1990), 586-598. [.pdf]

14. The Moments of Matched and Mismatched Hidden Markov Models, IEEE Transactions on Acoustics, Speech, and Signal Processing, ASSP-38 (1990), 610–622. [.pdf]

The following papers deal with practical problems of frequent occurrence and give useful tools for rectifying them.

15. The Effect of Inter-Channel Crosstalk on Array Performance, Journal of the Acoustical Society of America, 86 (1989), 1827-1834. [.pdf]

16. In Situ Optimal Reshading of Arrays With Failed Elements (with M.S. Sherrill), Special Issue on Underwater Acoustic Signal Processing, IEEE Journal on Oceanic Engineering, OE-12 (1987), 155-162. [.pdf]

The following three papers present the first algorithm for solving an approximation problem that occurs in diverse applications, and develop the associated mathematical theory.

17. Solution of Systems of Complex Linear Equations in the L Norm With Constraints on the Unknowns, SIAM Journal on Scientific and Statistical Computing, 7 (1986), 132-149. [.pdf]

18. Algorithm 635. An Algorithm for the Solution of Systems of Complex Linear Equations in the L Norm with Constraints on the Unknowns, ACM Transactions on Mathematical Software, 11(1985), 242-249. [.pdf]

19. Saddle Points and Over-Determined Complex Equations, Linear Algebra and Its Applications, 64 (1985), 57-76. [.pdf]

.The next papers derive a family of filters for independently controlling passband width and sidelobe taper, treating both discrete (digital) and continuous filter designs.

20. A Two-Parameter Family of Weights for Nonrecursive Digital Filters and Antennas, IEEE Transactions on Acoustics, Speech, and Signal Processing, ASSP-32 (1984), 108-118. [.pdf]

21. Limits of Chebyshev Polynomials When the Argument Is A Ratio of Cosines, Journal of Approximation Theory, 40 (1984), 393-395. [.pdf]

The following two papers use linear programming methods to optimize conventional beamformers for arbitrarily configured sonar arrays.

22. A Note on the Semi-Infinite Programming Approach to Complex Approximation (with A.H. Nuttall), Mathematics of Computation, 40 (1983), 599-605. [.pdf]

23. A General Chebyshev Complex Function Approximation Procedure and an Application to Beamforming (with A.H. Nuttall), Journal of the Acoustical Society of America, 72 (1982), 181-190. [.pdf]

24. Concertina-like Movement in the Absence of a Chebyshev System (with J.T. Lewis), Journal of Approximation Theory, 36 (1982), 364-367. [.pdf]

25. Real Excitation Coefficients Suffice For Sidelobe Control in a Linear Antenna Array (with J. T. Lewis), IEEE Transactions on Antennas and Propagation, AP-30 (1982), 1262-1263. [.pdf]

The next paper presents a method for shading large conformal sonar arrays to achieve wide beamwidths and low sidelobes for reliable communication.

26. Optimization of Discrete Arrays of Arbitrary Geometry, Journal of the Acoustical Society of America, 69 (1981), 199-212. [.pdf]

My earliest papers extend the Dolph-Chebyshev design technique for equispaced linear arrays to a class of symmetric linear arrays and to a class of concentric ring arrays.

27. Optimized Symmetric Discrete Line Arrays, IEEE Transactions on Antennas and Propagation, AP-23 (1975), 860-862.

28. Sufficient Conditions for the Existence of Optimum Beam Patterns for Unequally Spaced Linear Arrays With an Example, IEEE Transactions on Antennas and Propagation, AP-23 (1975), 112-115. [.pdf]

RESEARCH PAPERS APPEARING AS BOOK CHAPTERS


The following paper represents target location as a superposition of Gaussian densities and proposes an inflation/deflation technique for robustness to model prediction mismatch.

1. Bearings-Only Target Motion Analysis With Propagation Loss Models of Uncertain Fidelity, in Theoretical and Computational Acoustics ‘97, Y. Teng, E. Shang, Y. Pao, M. Schultz, A. Pierce, Editors, World Scientific, Singapore, 1999, 571-584.

The following three papers explore dynamical signal models and their suitability for high-performance super-computing architectures.

2. A Hidden Markov Model Approach to Frequency-Line Tracking of Non-Stationary Random Signals (with R.F. Barrett), Chapter 11 of Time-Frequency Signal Analysis, B. Boashash, Editor, Longman-Cheshire-Wiley, Melbourne, Australia, 1992, 250-276.

3. Connection Machine Implementation of Hidden Markov Models for Frequency Line Tracking (with J. L. Munoz), Chapter 14 in Very Large Scale Computation in the 21st Century, J.P. Mesirov, Editor, Society of Industrial and Applied Mathematics, Philadelphia, 1991, 204-217.

4. Frequency Line Tracking Algorithms (with A.K. Steele and R.F. Barrett), Proceedings of the NATO Advanced Study Institute on Underwater Acoustic Data Processing, Y.T. Chan, Editor, Kluwer Academic Publishers, Dordrecht, The Netherlands, 1989, 497-501.

The next paper applies matrix preconditioning and conjugate gradients to solving non-self-adjoint partial differential equations.

5. Polynomial Iteration for Nonsymmetric Indefinite Linear Systems (with H.C. Elman), in Lecture Notes in Mathematics #1230, J.P. Hennart, Editor, Springer-Verlag, New York, 1986, 103-117.

The following paper is my very first paper — it investigates the detailed behavior of certain approximations in Haar basis systems as the domain is systematically perturbed.

6. Extremals and Zeros in Haar Systems Are Monotone Functions of One Endpoint, in Theory of Approximation with Applications, A.G. Law and B.N. Sahney, Editors, Academic Press, New York, 1976, 387-401.

OTHER REFEREED RESEARCH PAPERS


The following papers develop and extend statistical pattern recognition ideas.

1. Maximum Likelihood Adaptive Neural System Application to Transient Signal Classification (U) (with L.I. Perlovsky, R.P. Coons, T.E. Luginbuhl, and S.G. Greineder), U.S. Navy Journal of Underwater Acoustics, Special Issue on Full Spectrum Processing, 43 (1993). (Confidential)

2. Procrustes: A Feature Set Reduction Method (with S.G. Greineder and T.E. Luginbuhl), U.S. Navy Journal of Underwater Acoustics, Special Issue on Full Spectrum Processing, 43 (1993). (Confidential)

The following paper evaluates the performance of an optimum signal processor against sensor noise generated by turbulent fluid flow.

3. Adaptive Array Processing in Non-homogeneous Flow Induced Self-Noise Fields, U.S. Navy Journal of Underwater Acoustics, 7 (1987), 105-124. (Confidential).

The following paper revealed a serious design flaw in state-of-the-art software for scientific computation (International Mathematical and Statistical Libraries); the modifications it proposed were incorporated in subsequent library upgrades.

4. Solution of Large Hermitian Eigenproblems on Virtual and Cache Memory Computers, Association for Computing Machinery SIGNUM Newsletter, 16 (1981), 6-7.