Media Informatics and Multimedia Systems

Offene Abschlussarbeiten

Im Folgenden finden Sie offene Themen für Abschlussarbeiten. Die Themen sind jeweils grob formuliert und bieten Spielraum für verschiedene Ausrichtungen innerhalb dieser Themenbereiche. Wenn Ihnen ein Thema gefällt, kontaktieren Sie bitte den entsprechenden Ansprechpartner. Bei allgemeinen Anfragen wenden Sie sich bitte an Marion Koelle (marion.koelle(at)uol.de).

Attention Guidance to Information Displays through Adaptive Salience

Summary: In this Bachelor or Master Thesis you will design, develop and evaluate a system to guide the attention of a user to a specific information display in a large monitoring environment, such as a control room or a ship bridge. The target display should be highlighted by increasing its salience and/or decreasing the salience of others (e.g. manipulating illumination intensity, color, orientation, motion). Depending on your experience and interest, the system can be either implemented using an augmented reality (AR) video-see-through head-mounted device (Google Cardboard), or with a monitor setup and a body-worn camera. For a Bachelor Thesis the complexity of the topic will be reduced regarding the lower experience and the shorter time frame.

The core parts of the research project are:  

  • systematically derive system design 
  • implement system 
  • evaluate system in user study

Prerequisites: Programming skills, image processing basics (OpenCV), basic statistics 

Working language: German or English 

Contact: Tim Claudius Stratmann (tim.claudius.stratmann@uol.de)

Detection of Radar Image Interpretation on Basis of Eyetracking Data

Summary: In this Bachelor or Master Thesis you will derive a predictive model to detect if an radar image is being interpreted by a human operator based on eyetracking data. This is helpful for attention guidance systems based on gaze direction, as it answers the question of whether the operator not only looked at, but also retrieved information from a radar display with given certainty. The data collection is part of this Bachelor or Master Thesis, we have contacts and will assist in finding dates with nautical students and experts for a user study in which we will collect the necessary data. For a Bachelor Thesis the complexity of the topic will be reduced regarding the lower experience and the shorter time frame.

The core parts of the research project are:  

  • identify areas of interest (AOI) in radar image 
  • setup eyetracking on radar simulator (basic programming for AOI logging)
  • collect data in user study (with nautical students and experts) 
  • create predictive binary classification model (e.g. using logistic regression)

Prerequisites: Programming skills, basic statistics

Working language: German or English 

Contact: Tim Claudius Stratmann (tim.claudius.stratmann@uol.de)

Tangible Toolkit to Design and Make own Things

Um komplexe Probleme lösen zu können eignet sich analytisches und algorithmisches Denken. Und dies sollte als sogenanntes „Computational Thinking (CT)“ gelernt werden. Es wird immer wichtiger in der heutigen Zeit und somit wichtiger Kinder diese Fähigkeit für ihre Zukunft beizubringen. Programmierung ist ein Weg CT zu lehren. Es hat sich gezeigt, dass es für Kinder motivierender ist, wenn sie etwas für sich bauen können. Zusätzlich ist Tangible Programming intuitiver und greifbarer für Kinder. Um Kinder spielerisch an die Programmierung heranzuführen sowie sie zu fordern und nicht zu überfordern, soll ein Toolkit entwickelt werden, welches den Kindern ermöglicht eigene Dinge zu entwickeln. Dabei sollen verschiedene Sensoren (Licht, Temperatur, Bewegung,...) und Aktuatoren (Sound, Licht, Vibration) verwendet werden und dessen Zusammenspiel sowie weitere Programmierkonzepte beigebracht werden. Die Entwicklung des Toolkits soll entlang des Human-Centered Design Process erfolgen um die Usability und User Experience des Interaktiven Systems für die Kinder zu erhöhen.

Kontakt: Erika Root (erika.root@offis.de)

Designing Cues for Assisting In-vehicle Multitasking

In the coming age of automation, we will observe a fundamental shift in the way in which we interact between humans and machines. Automation will not come overnight. It is agreed that in the decades to come with an increasing level of automation many situations will remain in which humans and automated systems will share control. When automation finds its limitations human will be in charge to take over control from the vehicle. An intuitive example is that of a human reading a book in an automated car on the motorway and returning to manual driving when the car exits the motorway and returns an urban road.  Initiation of this transition from automated to manual control, so-called takeover requests (TOR) will have to be designed as part of the driving assistant system that allow to return from a secondary task back to the primary task of driving. 

In this project we will design and evaluate human machine interfaces which assist drivers in take-over situations with

  • Shift their attention from the non-driving task
  • Preparing them for the take-over maneuver
  • Inform them about the driving environment at take over moment

Prerequisites: Arduino, Android/Java, interested in prototyping, basic statistics knowledge

Contact: Shadan Sadeghian (shadan.sadeghian[at]offis.de)

Designing and evaluating an HMI for assisting the Ship Bridge crew in canal passage

Summary: Mariners make use of complex ship bridge systems to gather information elements and look for cues in the environment outside of the bridge to complete a navigation task. Designers are asked to develop assistance systems that reduce the cognitive load of the crew and allow efficient task execution. Research on new systems development in the maritime domain is focused on navigation tasks in internal waters, such as canal passage, to decrease the risk for incidents and accidents. The thesis will explore a Goal Based design of an HMI for ship bridges, where a combination of an FCM modelling approach and an HMI design method will be applied.

The core parts of the research project are: 

  • Systematically derive the HMI design with the combination of the methods (FCM modelling, HMI design method ‘Konect’)
  • Implement an HMI design
  • Evaluate the HMI (inkl. cognitive effort, correctness of perception) 

Prerequisites: Java, interested in prototyping, basic statistics 

Offered: Real time experiments in ship bridge simulators 

Working language: English (preferred) or German 

Contact: Marie-Christin Ostendorp (Ostendorp [at] offis.de), Stella Parisi (parisi [at] offis.de)