Master Theses

In their master thesis, students demonstrate their ability to independently solve problems in the area of their specialization using state-of-the-art scientific methods.

The following list of master theses reflects a variety of different topics and projects to work on. Students are encouraged to propose further topics suiting their personal professional interest. In any case, please contact the advisor in due course.

Already finished master thesis can be found here

Transportation Informatics

Ongoing Master Thesis works

Name:	Title:
RAMANNAGARI SHYAM PRASAD GOUD	Evaluation and improvement of a Requirement Specification to support formalisation of requirements
KOTTE SANDEEP	REAL TIME CAR-DRIVER EMOTION DETECTION SYSTEM USING SPEECH SIGNALS
SCHAUTZ MARIO	Systementwicklung und prototypische Implementierung des frameworks eines "extended floating car data" Konzepts zur Unterstützung des Eco-Driving
SPIEL MARC THOMAS	Verkehrstelematik im städtischen Gebiet anhand der Stadt Klagenfurt - eine Kosten-Nutzen-Analyse bestehender Verkehrsanlagensysteme und die Entwicklung eines Expertensystems für die Planung von Verkehrsanlagen
PALURU SANTOSH KUMAR	OPTIMISATION OF PROCESS TOOL BASED MANAGEMENT IN FLEXIBLE IT
SEEBACHER SIEGFRIED	Qualitätsmanagementkonzept für Sensorsysteme auf der Autobahn
GUDISEVA KIRAN KUMAR	Design, implementation and test of an automatic measurement setup for high efficiency Photovoltaic Converters (Solar-Cell)
DHANEKULA SIVA NAGA KRISHNA	INTRODUCTION OF STATISTICAL METHODS FOR PRODUCTION CONTROL
VEERANKI SAI PAVAN KUMAR	Map Matching Algorithm for Floating Car Data
BOJNIK FRANZ	MPEG-7 based Driver Assistance
GENSER FRITZ	Das Laris Service Portfolio - Innovatives Service Engineering durch Nutzung von Routendaten
NARIPELLA RAHUL PRAKASH	Implementation of an Ant-Colony-System to Solve the Dynamic Vehicle Routing Problem
SCHELANDER MARTINA	Entwicklung eines mobilen, ortsabhängigen Verkehrsinformationsdienstes auf Basis eines elektronischen Mautsystems

Open Master Thesis topics

PID Controller and PWM Generator for Motor Driver Based on PIC or ATMEL uController

Description

Goal of this project is design and implementation of a micro controller-based PID controller for controlling a motor’s robot. Due to the landscape and slope we should control the motor torque. In this case we must get a feedback from motor, according to the input pulse and environment situation, and system must generates proper PWM pulse to ensure that the motor is working according to the input reference.

PDF-Download

Design and Implementation of Image Processing Framework on GPU - Phase - 1

Description

One of the main important parts of machine vision is Image processing unit, there are lot of critical technical operation in image processing unit that make bottle neck in the system. Almost all of image processing operators and filters are pixel based. Each pixel is surrounded by the pixel neighbors witch access to this data is necessary for the image processing calculations. In the classical method and system architecture, due to the limitation of hardware, we have to process pixels sequentially, which is time consuming and slow, by some software trip in advanced CPU architecture we can speed up the system such as multi thread programming (Time Sharing) and pipelining, but still system is slow. Parallel processing is one the best way to reach 100 or more speed up in processing time.

GPU (Graphical Processing Unit) is architecture for graphic acceleration. NVIDIA released a SDK (CUDA), for high performance computing based on GPU. In this SDK, there are some essential functions for transaction between CPU host and GPU device. Such as, memory allocation, memory copy, etc.

Each GPU is dedicated for calculation of floating point operation, and programmer is able to create several threads dynamically for increasing the speed.

In this research, we are going to design and implement a fully object oriented framework based on C++ and CUDA SDK. At the end, we have a high level library which is possible to write a behavioral code for machine vision.

GPUCV is an Open source project similar to this project (it has started Oct 2007), and student can start his/her project on this framework. This project is not completed jet.

PDF-Download