Statistical Signal Processing

This post graduate course is designed to cover techniques for statistical signal processing, detection and parameter estimation. It will briefly review the preliminaries on linear algebra and statistics. The rest of the course is broadly divided into three parts. The first part will deal with the design, implementation and performance evaluation of detectors; this would cover composite and M-ary hypothesis testing. The second part of the course deals with estimation techniques like Maximum Likelihood, MAP and MMSE estimation. The third part introduces adaptive filtering approaches; this will cover stochastic and data-driven approach with emphasis on least squares based techniques. Homework will be a mix of theory and programming assignments.

1. Students are able to apply hypothesis testing to signal and event detection problems.

2. Students are able to evaluate detector performance using ROC curves, sensitivity/specificity etc.

3. Students can choose among MLE, MAP and MMSE estimators given a parameter estimation tasks.

4. Students are able to apply and design least squares based adaptive filters for stochastic signals.