Discrete Message Passing Algorithms for Receiver Design

Betreuer: Tobias Monsees
Art der Arbeit: Projekt (MSc), Masterarbeit (MSc), Bachelorarbeit (BSc.), Vertiefungsprojekt (BSc.)
Ausgabe: -
Bearbeiter: -
Status: zu vergeben

The sum-product algorithm is an efficient message-passing algorithm to calculate (exact or approximate) marginal distributions by exploiting the factorization structure the underlying probability distribtuion [1]. Various algorithms in digital communications and signal processing (such as the Viterbi algorithm, the iterative "turbo" decoding algorithm, the Kalman filter and many others) can be derived as specific instances of the sum-product algorithm [2]. Nevertheless, this algorithm requires the processing of continuous messages and functions which can only be approximated by digital signal processing with a finite number of bits per message.
The goal of this work is to investigate discrete variants of the sum-product algorithm with a suitable low-bit resultion for each message. 

- Literature seach regarding message passing algorithms for reveicer design
- Investigation of appropriate discrete message passing algorithms
- Comparison between discrete and "continuous" variants 

Requirements: Communication Technologies, Basic MATLAB and/or C programming skills.

[1] M. C. Bishop, "Pattern Recognition and Machine Learning", Springer-Verlag, 2006.
[2] F. R. Kschischang, B. J. Frey, H. -A. Loeliger, "Factor Graphs and the Sum-product Algorithm", IEEE Transactions on Information Theory, vol. 47, no. 2, pp. 498-519, 2001.

Zuletzt aktualisiert am 01.04.2019 von T. Monsees
AIT ieee tzi ith Fachbereich 1
© Arbeitsbereich Nachrichtentechnik - Universität BremenImpressum / Kontakt