[Skip navigation]

[LU]

The statistical signal processing group

Thesis projects

The Statistical Signal Processing Group (SSPG) continuously offer M.Sc. thesis projects within its scope of research. These projects are commonly either linked to currently ongoing research projects within the group, or are part of industrial collaboration projects within the group's areas of interest. Below are listed some of the current and availabe thesis projects of the group.

Current M.Sc. project

Blood veclocity estimation using ultrasound
Student: Pamela Buratti, Supervisor: Andreas Jakobsson
Classification and detection of liquid explosives
Student: Niclas Ek, Supervisor: Andreas Jakobsson

Offered projects

High-resolution data-adaptive time-frequency representation of signals
This M.Sc. thesis project deals with finding an appropriate time-frequency representation of non-stationary signals. Typically, most real-world signals are non-stationary, and can only be considered stationary over limited time intervals. However, in a wide variety of applications, there is a significant interested in determining how the spectral content of the signal is varying with time. The proposed project aims at developing a novel high-resolution data-adaptive time-frequency representation of non-stationary signals. Comparative studies using simulated data sets will examine how the new technique compares to traditional approaches, such as the spectrogram and the Wigner-Ville and Choi-Williams time-frequency representations. The method will also be evaluated on real measurements of interference signals from an NQR sensor used to detect of hidden explosives. Supervisor: Andreas Jakobsson
Detection and interference rejection of non-circularly symmetric sources
Until recently, most array filtering problems have only considered second-order circular sources. As many forms of communication signals are non-circular, there is currently a significant interest in examining how to best detect desired non-circular sources, as well as how to best reject interference from undesired such sources. The suggested project will examine these issues in further detail. Supervisor: Andreas Jakobsson
Audio Signal Processing
When sending out audio on a loudspeaker the audio will be affected by the characteristics of the loudspeaker and quite often the loudspeaker will introduce non-linear artifacts in the audio. This thesis is about finding and experimenting with methods that can limit non-linear artifacts or the effects of it in audio. This project will be performed at ST-Ericsson in Lund. Supervisor: Andreas Jakobsson