• Diese Seite als PDF öffnen.

Signal Characterization by Machine Learning

Description

A major challenge in the field of signal processing is the extraction of characteristic signal features form given complex baseband data corrupted by noise. On the one hand, there exist classical signal processing methods to cope with such tasks. On the other hand, machine learning based techniques to extract particular signal characteristics under certain constraints are presented in literature. Appropriate detection methods based on machine learning and/or classical estimation theory should be examined and compared in this master thesis.

Supervisor:

Student

• Diese Seite als PDF öffnen.

Message Passing Decoding of BMERA Codes

Description

In this research internship, the student will implement a belief propagation (BP) decoder for message passing decoding of BMERA codes, also known as convolutional polar codes. For this, the student will first start with a conventional  BP decoder for polar codes and then take the necessary steps to adapt the algorithm for BMERA codes. Eventually, a numerical performance evaluation by means of frame error rate simulations should be made.

Prerequisites

Supervisor:

Student