EQ2400 Adaptive Signal Processing 6.0 credits

Adaptiv signalbehandling

Please note

This course has been cancelled.

This course treats adaptive signal processing algorithms for extracting relevant information from noisy signals. The emphasis is on recursive, model based estimation methods for time-varying systems. Applications in, for example, communications, control and medicine are treated.

Fundamentals for adaptive systems; mean-square estimation, Wiener filters. Introduction to adaptive structures and the least squares method. State space models. Kalman filters. Search techniques: Gradient and Newton methods. LMS (least mean squares), RLS (recursive least squares). Analysis of adaptive algorithms: Learning curve, convergence, stability, excess mean square error, mis-adjustment. Generalizations of LMS and RLS. 

  • Education cycle

    Second cycle
  • Main field of study

    Electrical Engineering
  • Grading scale

    A, B, C, D, E, FX, F

Last planned examination: autumn 20.

At present this course is not scheduled to be offered.

Intended learning outcomes

This course treats adaptive signal processing algorithms for extracting relevant information from noisy signals. The emphasis is on recursive, model based estimation methods for signals and systems whose properties change in time. Applications in, for example, communications, control and medicine are discussed.

Learning outcomes:

After the course, each student is expected to be able to:

  • Design and apply optimal minimum mean square estimators and in particular linear estimators. To understand and compute their expected performance and verify it.
  • Design, implement and apply Wiener filters (FIR, non-causal, causal) and evaluate their performance.
  • Use a combination of theory and software implementations to solve adaptive signal problems. Especially:
  • Identify applications in which it would be possible to use the different adaptive filtering approaches.
  • Design, implement and apply LMS, RLS, and Kalman filters to given applications.
  • Analyze the accuracy and determine advantages and disadvantages of each method.
  • Use the theoretical understanding to do troubleshooting, e.g., in cases the observed performance is not as expected.
  • Report the solution and results from the application of the above filtering techniques to given problems.

Course main content

Fundamentals for adaptive systems; mean-square estimation, Wiener filters. Introduction to adaptive structures and the least squares method. State space models. Kalman filters. Search techniques: Gradient and Newton methods. LMS (least mean squares), RLS (recursive least squares). Analysis of adaptive algorithms: Learning curve, convergence, stability, excess mean square error, mis-adjustment. Generalizations of LMS and RLS.

Eligibility

For single course students: 180 ECTS credits and documented proficiency in English B or equivalent

Recommended prerequisites

EQ1220 Signal Theory or  EQ1270 Stochastic Signals and Systems 
EQ2300 Digital Signalbehandling

Literature

Lecture notes: Adaptive Signal Processing, Hjalmarsson & Ottersten, KTH-EE

Examination

  • PRO1 - Project, 1.0, grading scale: P, F
  • PRO2 - Project, 1.0, grading scale: P, F
  • TENA - Examination, 4.0, grading scale: A, B, C, D, E, FX, F

Requirements for final grade

2 Project assignments (PRO1, 1 ECTS credits, grading P/F; PRO2, 1 ECTS credits, grading P/F) completed and reported in pairs of at most 2 students before given deadlines.
Written exam (TENA, 4 ECTS credits, grading A-F)

Offered by

EES/Information Science and Engineering

Contact

Magnus Jansson

Examiner

Magnus Jansson <janssonm@kth.se>

Add-on studies

EQ2430/EQ2440 Project course in signal processing and digital communications

EQ2450/2460 Seminars in Signals and Systems/Wireless Systems

Version

Course syllabus valid from: Autumn 2007.
Examination information valid from: Autumn 2007.