ET4386 Estimation and Detection
IntroductionThis course covers the basics of detection and estimation theory, as used in statistical signal processing, adaptive beamforming, speech enhancement, radar, telecommunication, loclization, system identification, and elsewhere.
Part I: Optimal estimation covers minimum variance unbiased (MVU) estimators, the Cramer-Rao bound (CRB), best linear unbiased estimators (BLUE), maximum likelihood estimation (MLE), recursive least squares (RLE), Bayesian estimation techniques, and the Wiener filter.
Part II: Detection theory covers simple and multiple hypothesis testing, the Neyman-Pearson Theorem, Bayes Risk, and testing with unknown signal and noise parameters.
The course complements EE4c03 Statistical digital signal processing and modeling, and gives a solid background for EE4715 Array Processing and EE4685 Machine learning, a Bayesian perspective.
Preliminary knowledgeTo follow the course with profit, you will need the background knowledge provided by an elementary course in Random Signals.
In principle, the exam in the study year 2021/2022 will be a written exam. However, depending on the developments of the Covid-19 Pandemic, the course is subject to change. In case the Covid-19 pandemic makes it difficult to have a written exam, this will be changed into an oral exam.
The exam is closed book, but, students are allowed to bring a double sided self handwritten A4 formula sheet.
As part of the course, there is a compulsary mini project, which helps you to get experienced with the theory and to apply this to a practical problem. The available mini projects will be announced via the course website, after which students can sign in via Brightspace. The mini projects can be done in pairs. To sign up for a project, go to brightspace, enroll for the course, and then go to the tab "collaboration", and select "groups".
The final report on the project needs to be uploaded in pdf before January 10th 2022 via brightspace.
The project report will be graded and counts for 20 % of the final grade. Access to the final exam will only be granted if the project is handed in.
- Project 1: "Multi-Microphone Speech Enhancement". Description - Data
- Project 2: "Signal Detection". Description - Data
- Project 3: "Cognitive radio". Description - Data
- Project 4: "Clock Synchronization for wireless sensor networks". Description - Data
- Project 5: "Accelerometer Calibration". Description - Data
To sign up for the mini-projects, go to the course page on brightspace. Then go to the tab "collaboration", and then select groups. Signing up can be done until December 1st 2021. After that the enrolment for the projects will close. To upload your final assignment (before January 10th 2022), do this via "assignments" in brightspace.
For questions on the projects, please email at et4386-EWI@TuDelft.nl
- Fundamentals of Statistical Signal Processing, Volume I: Estimation Theory; S.M. Kay, Prentice Hall 1993; ISBN-13: 978-0133457117.
- Fundamentals of Statistical Signal Processing, Volume II: Detection Theory; S.M. Kay, Prentice 1993; ISBN-13: 978-0135041352.
InstructorsThe lectures will be given by Dr. ir. Richard C. Hendriks (RCH) and Dr. Raj Thilak Rajan (RTR).
ContactPlease use the following email address for any inquiries regarding the course and the mini projects:
Course materialIndividual files in PDF format are available below. As the course develops additional files with e.g., solutions to the exercises, will be posted.
ScheduleThe schedule for 2021/2022 is as follows:
|Introduction. Estimation theory - MVU, CRB||Vol.1 Chapters 1 and 2||MVU||Introduction/MVU|
|Estimation theory: Cramer Rao Lower Bound (CRB)||Vol.1 Chapter 3 - 3.5 , Chapter 5||CRB|
|3.||Nov 15||RTR||Estimation theory: Best Linear Unbiased Estimators (BLUE), Maximum likelihood estimation (MLE)||
Vol.1: Ch. 3.7, Ch. 6.1 - 6.5 and Ch. 7.1 - 7.6
|4.||Nov 18||RTR||Estimation theory - Least squares (LS)||Vol.1: Ch. 8.1 - 8.4 and 8.8-8.9||Least Squares||LS|
|5.||Nov 22||RTR||Estimation theory - Bayesian philosophy||Vol.1 Ch. 10-10.6||Bayesian||Bayesian philosophy|
Estimation theory - Bayesian estimators
Vol.1 Ch. 11-11.5
vol. 1 Ch. 12-12.5
|MAP - LMMSE||MAP/LMMSE|
|7.||Nov 29||RTR||Wiener filters||
Vol.1 Ch. 12.7
|8.||Dec 2||RCH||Detection theory - Introduction, Neyman Pearson theorem||Vol.2 Chapters Ch. 3-3.7||Introduction detection||Intro Det. Theory|
|9.||Dec 6||RCH||Detection theory - Deterministic signals||Vol.2 Chs. 4-4.4||Deterministic Signals|
|10.||Dec 9||RCH||Detection theory - Random Signals||Vol.2 Chapters 5-5.6||Detection - Random signals||Random Signals/GLRT|
|11.||Dec 13||RCH||Detection theory - GLRT||Vol.2 Chapters 6-6.4||GLRT|
|13.||Dec 20||RCH||Estimation of Sensor Array Signal Model Parameters Using Factor Analysis: Biomedical and audio array processing|
|14.||Dec 23||RTR||Estimation & Detection: Sensor calibration|
The book contains many examples and exercises. A (incomplete) list with recommanded exercises from the book can be downloaded here. In addition, some extra examples and exercises are given in the list below:
- Example MVU & CRLB
- Exercise 10.3
- Example MVU, CRLB & Bayesian Estimation.
- Bayesian Estimation (Answers)
- Detection Theory (Answers)
- Estimation Theory (Answers)