EECS 225A Spring 2006

schedule (project presentations), announcements, resources

Catalog description

225A. Digital Signal Processing. (3) Three hours of lecture per week. Prerequisites: 123 and 126 or solid background in stochastic processes. Advanced techniques in signal processing. Stochastic signal processing, parametric statistical signal models, and adaptive filtering. Application to spectral estimation, speech and audio coding, adaptive equalization, noise cancellation, echo cancellation, and linear prediction.

Textbook

Monson H. Hayes, Statistical Digital Signal Processing and Modeling, Wiley, 1996 (ISBN 0471594318) [homepage]

In addition, various resources (Web and handouts) will be used to supplement the text.

Course Topics

With signal processing becoming ubiquitous in today's computer literate world, a large number of application areas are growing in importance, both in industry and in the research community, such as seismic signal processing, speech data processing, medical image processing, radar signal processing, and sensor array processing. These problems have many different aspects, and a corresponding number of different solutions have been explored.

Signal Models (7 lectures)

• Deterministic signal models (least-squares fits)
• Stochastic signal models (essentially 2nd order: WSS signals & LTI systems)
• Graphical models approach
• Data-driven approaches
• Spectral estimation

Signal Representation and Approximation (8 lectures)

• Fourier, wavelets, splines
• Karhunen-Loeve Transform; principal components analysis
• Sampling (Hilbert space formulation)
• Quantization (Rate-distortion theory, Dithering)

Signals, Systems, Noise (14 lectures)

• Estimation Theory (Wiener, Kalman, Adaptive, Neural networks...)
• Detection Theory (Neyman-Pearson, etc)
• System Identification

Instructor

Michael Gastpar [homepage]

265 Cory Hall

Office hours: Wed 3-4, Thu 11-12

Email: gastpar

Student responsibilities

Please read the course announcements often. If it is posted there, you are presumed to have been informed about it. See the announcements for reading and homework assignments. The schedule includes important deadlines, such as exams and project due dates.

Students are also reminded of the Departmental Policy on Academic Dishonesty and are also urged to also read and abide by the professional ethics represented in the IEEE Code of Ethics. Especially relevant in the latter are the two guidelines:

• To be honest and realistic in stating claims or estimates based on available data.
• To seek, accept, and offer honest criticism of technical work, to acknowledge and correct errors, and to credit properly the contributions of others.

Grades

The components of the course will be weighted as follows in the final grade. The final grades will be set by matching a curve to the final course averages.

Component	Weight	Comments
Homework	10%
Midterm	30%	80 minute open-book exam.
Final Exam	30%	80 minute open-book exam.
Literature project	30%	In small groups (1-3 students), you will select a subarea of the class and explore the related literature. You will select around 5 papers and discuss and extend their contributions in a short report.

Important Dates

Exam	Date and time	Location
Midterm	March 21 @ 2-3:30	299 Cory Hall
Final Exam	May 12 @ 12:30-3:30	TBD
Project Proposals	March 8 @ midnight
Project Final Report	May 3 @ midnight

Acknowledgement

This web page is mostly drawn from the web page of Prof. D. G. Messerschmitt (Spring 2005).