EEC264 – Estimation And Detection Of Signals In Noise
4 units – Fall Quarter; alternate years
Lecture: 3 hours
Discussion: 1 hour
Prerequisite: EEC 260
Grading: Letter; problem sets (10%), MATLAB assignments (10%), midterm exam (35%), final exam (45%).
Introduction to parameter estimation and detection of signals in noise. Bayes and Neyman-Pearson likelihood-ratio tests for signal detection. Maximum-likelihood parameter estimation. Detection of known and Gaussian signals in white or colored noise. Applications to communications, radar, signal processing.
Expanded Course Description:
- Hypothesis Testing (2 weeks)
- Bayesian likelihood ratio tests for binary decisions
- Receiver operating characteristic
- Non-Bayesian minimax and Neyman-Pearson tests
- M-ary hypothesis testing
- Parameter Estimation (2 weeks)
- Bayesian, maximum a posteriori, and maximum-likelihood estimation of parameter vectors
- Cramer-Rao lower bound, bias, efficient estimates
- Linear least-squares estimation and its geometric interpretation
- Orthogonal Expansion of Gaussian Processes (1 week)
- Orthogonal expansion of deterministic signals
- Karhunen-Loeve expansion of discrete and continuous-time Gaussian processes
- Detection of Known Signals (2-1/2 weeks)
- Detection of known signals in white Gaussian noise (WGN)
- Sufficient statistics
- Correlator and matched filter receiver implementations
- Performance evaluation
- M-ary detection in WGN
- Detection of known signals in colored noise: resolvent and whitening filter approaches.
- Detection of Signals with Unknown/Random Parameters (2 weeks)
- Detection of signals with unwanted parameters. Composite hypothesis testing
- Estimation of waveform parameters in noise
- Application to the estimation of pulse amplitude and delay and sinewave amplitude, phase and frequency.
- Joint estimation and detection, generalized likelihood ratio test (GLRT)
- Detection of signals with random parameters. Detection of signals with incoherent phases and/or random amplitudes. Envelope detectors
- Detection of Gaussian Signals in WGN (1/2 week)
- Generalized correlator receiver structure for detecting Gaussian signals in WGN
- R. N. McDonough and A. D. Whalen, Detection of Signals in Noise 2nd Edition, Academic Press, 1995.
Instructors: Ding, Levy
THIS COURSE DOES NOT DUPLICATE ANY EXISTING COURSE.
Last revised: January 2001