ECEN 5652 - Detection and Extraction of Signals from Noise
3 credit hours
Prerequisite ECEN 5612, Noise and Random Processes
Textbook: Statistical Signal Processing, L. Scharf, 1991.
Reference: An Introduction to Detection and Estimation, V.H. Poor, Springer-Verlag, 1989. Detection, Estimation and Modulation Theory, H.L. VanTrees, Wiley, 1968. Principles of Communication Engineering, Wozencraft and Jacobs, Wiley, 1965.
Goals: Understanding of the fundamentals of hypothesis testing and estimation and their engineering applications in various signal detection and extraction problems.
Topics:
- Review of applied probability and random processes, the Karhunen-Loeve expansion. Statistical modeling and an introduction to detection and estimation problems.
- Fundamentals of linear algebra: vector spaces, linear independence, QR factorizations, linear subspaces, singular value decompositions, projections, rotations, psuedoinverses.
- Detection theory: simple hypothesis testing under the Bayes' criterion and the Neymann-Pearson criterion; sufficient statistics. Composite hypothesis testing, the Neymann-Pearson criterion and the notion of invariance. The generalized likelihood ratio test and its optimality. Minimax detection and sequential detection. Applications include detection problems in communications and radar/sonar signal processing including the linear statistical model and the multivariate Gaussian model.
- Estimation theory: maximum likelihood estimation and sufficiency, Cramer-Rao inequality, Bayesian and minimax parameter estimation including minimum mean-squared error and maximum a posteriori estimation, linear minimum mean-squared estimation; applications including the multivariate normal model, linear statistical model, Kalman and Wiener filtering.
- Moderate.
- None.
