Signal Processing and Speech Communication Laboratory
home › phd theses

PhD Theses

Anneliese Kelterer: The prosody of interactional and discursive strategies in Austrian conversational speech

Prosody has many functions in speech; e.g., cueing information structure (“Max bought a HOUSE.” vs. “MAX bought a house.”), sentence type (“Max bought a house?”), or communicative functions such as turn management (do I want to continue telling you about Max’s new house or am I done talking). This thesis investigates the prosody of yet another kind of communicative function, the expression of attitude (also called stance-taking, evaluation).

Wolfgang Roth: Bayesian Methods in Machine Learning

Many machine learning algorithms seek for a mode in a posterior distribution and then make predictions solely based on that mode. These approaches waste lots of information available in the posterior and are prone to overfitting when a bad local mode is found. The full Bayesian approach computes its prediction by an expectation over the posterior. Computing this expectation analytically is in general intractable and subject of current research. The task of the thesis is to develop techniques based on the full Bayesian approach:

Stefan Grebien: Cognitive MIMO Radar for RFID Localization

It is hypothesized that the use of cognitive system concepts as well as multiple input, multiple output techniques can enable accurate and robust indoor localization using a passive radio frequency identification (RFID) system.

Julian Linke: Contribution of prosodic information in ASR systems for read and spontaneous speech by integrating long-term and short-term acoustic features

State-of-the-art ASR systems perform well on read and conversational speech (see modern virtual assistants like Alexa or Siri) but yet the recognition of spontaneous speech is associated with many difficulties which potentially could benefit from new ideas for the speech recognition task. This thesis presents speech recognition experiments which incorporate prosodic information to improve ASR systems for read and - spontaneous - conversational speech. Specifically, this approach is suitable for languages with lower available resources. One of the main reasons for the arising difficulties are the many pronunciation variants which need to be understood and learned when developing modern ASR systems. On this account the main focus lies on the improvement of the acoustic model (which represents one main part of a modern ASR system) by integrating, for example, different long-term and short-term acoustic features. Consequently, the trade-off between knowledge-based and data-driven approaches is levered out by illustrating - and contrasting - the advantages of prosodic information included in the modeling process of ASR systems.

Martin Ratajczak: Deep Learning and Structured Prediction

Linear-chain conditional random fields (LC-CRFs) have been successfully applied in many structured prediction tasks. LC-CRFs can be extended by different types of deep models.

Johanna Rock: Deep Learning for Resource-Constrained Radar Systems

Radar systems provide information about object distances, velocities and positions but they can also be used for object recognition, such as pedestrian classification. The constant sensing of the various sensors in a car produces a huge amount of raw data, which requires an intelligent extraction of relevant data as well as distributed processing on the sensing units themselves. In this project we investigate the use of machine learning methods such as Deep Neural Networks to address these issues. Special attention is placed on resource efficiency of the used models in order to deploy them directly to the hardware.

Eric Kurz: Flexible Measurement Technology for spatial and directional Sound Field Analysis

Statistical room acoustics is mainly based on the analysis of room acoustic impulse responses, which are determined by means of omnidirectional measurement methods. Almost all classical room acoustic quality measures are calculated from these impulse responses. All these measures of quality, which are used to describe the acoustic behaviour of a room are based on the assumption of a diffuse sound field that produces a uniform and isotropic sound energy distribution in the entire room. However, for many room acoustic problems it is not sufficient to perform analyses based on this assumption.

Mate Andras Toth: Interference Mitigation for Automotive Radar

Christoph Obermair: Machine Learning Models for Failure Prediction in RF Cavities for Accelerators

Radio Frequency (RF) breakdowns are one of the most prevalent limits in RF cavities for particle accelerators. During a breakdown, field enhancement associated with small deformations on the cavity surface results in electrical arcs. Such arcs lead to beam aborts and if they occur frequently, they can cause irreparable damages on the RF cavity surface.

Truc Nguyen: Machine Learning for Acoustic Scene and Respiratory Sound Classification

Acoustic scene classification is an important task of a machine hearing system used for classification of general sounds in everyday environments. It aims to characterize and label the acoustic environment of an audio recording. While respiratory sounds classification is a key task on computer-aided lung sound research to support medical diagnosis. It focuses on automatic recognition of respiratory adventitious sounds or healthy and several categories of pathological lung sounds. The thesis develops machine learning methods for these application tasks.

Lukas Wielandner: Narrowband positioning exploiting massive cooperation and mapping

Narrowband positioning of sensors in an indoor environment is highly inaccurate. This problem can be tackeld by using large sensor networks and exploiting massive cooperation and information about the environment.

Saskia Wepner: Contribution of prosodic features to language models in automatic speech recognition systems and human perception

State-of-the-art automatic speech recognition (ASR) systems achieve accuracies close to human capabilities for read speech. When it comes to conversational speech (CS), systems perform much worse. Reasons for that are, on the one hand, incomplete sentences, wrong grammar, slang vocabulary and a broad pronunciation variation resulting from both linguistic phenomena such as reduced pronunciation in a familiar context and dialect speaking styles. On the other hand, there is usually not enough data available to train existing systems sufficiently. Making use of linguistic knowledge about CS, such as prosodic features, shall generate a better understanding of how the conversational character of a dialogue affects pronunciation and how sentences are grammatically deformed in CS. With the focus on language models (LMs), this understanding is expected to improve current ASR systems for CS without the need of large data bases. The aim of this research is to find prosodic features that contribute to the performance of both ASR systems and human speech perception. Therefore, humans shall challenge the adapted LM(s) in perception experiments yielding further knowledge on to date human superiority in speech recognition of CS which can then again be exploited in ASR and vice versa.

Alexander Venus: Reliable and Robust Localization and Positioning

This thesis will be focused on the reliable and secure localization of electronic keys using wireless technologies, aiming at reaching some defined level of reliability and trusworthiness. That is, essentially, it seeks to investigate trustworthy models and algorithms for distance bounding and localization with respect to defined geometric zones by studying the physical layer signal processing within the wireless transceiver chips and system level aspects building upon it.

Thomas Wilding: Robust Positioning in Ultra-Wideband Off-Body Channels

The positioning accuracy that can be achieved using communication over a wireless link between transceivers is envisioned to be improved by the introduction of multiple antenna elements at the RF-devices.

Michael Rath: Signal Processing for Localization and Environment Mapping

In the upcoming years we face the reality of everyday objects being connected to a large network, the ‘Internet of Things’ (IoT). A key aspect of the IoT is dependable communication and localization, where the participants act as ‘Smart Things’, communicating with each other and being aware of their environment.

Jakob Möderl: Using UWB Radar to Detect Life Presence Inside a Vehicle

Approximatley 40 tragic deaths of small children locked in vehicles occur in the US each year due to the extreme heat or cold inside the parked vehicle. If the vehicle is able to detect the presence of a child (or any other life such e.g. a pet) it can either alert the owner or adjust the climate control in order to avoid these tragic accidents.

Finished Theses