Personal tools
You are here: Home Education Theses finished Theses
 
Multimodal Interactive Systems - Old Diploma- and Semestertheses

Alte Themen für Diplom- und Semesterarbeiten


Find Location using PlaceLab (SA, DA, HiWi) (vergeben)

Emerging applications in the field of mobile, wearable and ubiquitous computing try to include more of the user's context, in order to better adapt the user interface to the user. Ideally the computer would proactively act on the user's behalf.

While many different kinds of context are possible and useful, the user's location allows to tell many things about the user's overall situation. Intel Research Seattle [1] has developed a location system, PlaceLab [2], that is based solely on the readings of Wireless LAN cards. It provides high precision at low cost, since it does not require additional hardware.

In this Semester- or Diploma-Thesis we would like to setup a PlaceLab Environment here at the TU Darmstadt. The thesis will involve setting up the environment, recording WLAN Access Point Data for the TU Campus and comparing the inside/outside precision against GPS measurements.

If you are interested in this thesis or have any questions please feel free to send me an e-mail or pass by my office any time! (ich spreche natürlich auch Deutsch:)

Further Information

Environment

Begin: as soon as possible
Development Environment: Java, WLAN
Prerequisites: Experience in Java will be of help
Supervision: Nicky Kern
< vorname.nachname AT informatik DOT tu-darmstadt DOT de >
S02/02 B111, Tel. 16-6182

Published 2.12.2004, Nicky Kern


Calling Behaviour with Uncertainty Behaviour (SA, DA)

In today's mobile phones the caller has no means of determining if it is convient for the callee to receive a call in a given situation. We would like to investigate how people would interact with a system that displays them the interruptability of the callee before actually making the call.

We have developed a model of human interruptability that distinguished between interrupting the user and his environment. Alongside we have developed an algorithm that allows to estimate the user's interruptability from wearable sensors (see this page [1] for a more complete description).

In this thesis we would like to investigate how the behaviour of mobile phone users changes when they are made aware of the callees interruptability. In a first iteration this will be done by interrupting the callee anyway - later on it would be desirable that the system is integrated with the automatic recognition component.

If you are interested in this thesis or have any questions please feel free to send me an e-mail or pass by my office any time! (ich spreche natürlich auch Deutsch:)

Further Information

Environment

Begin: as soon as possible
Development Environment: Java, Mobile Phone
Prerequisites: Experience in Java will be of help
Supervision: Nicky Kern
< vorname.nachname AT informatik DOT tu-darmstadt DOT de >
S02/02 B111, Tel. 16-6182

Published 2.12.2004, Nicky Kern


Activity Recognition using Resistive, Knitted Stretch Sensors (SA, DA)

Emerging applications in the field of mobile, wearable and ubiquitous computing try to include more of the user's context, in order to better adapt the user interface to the user. Therefore it is interesting to know the user's current physical activity.

In previous work we have investigated the use of acceleration sensors to capture the user's physical activity, such as sitting, walking, standing, etc. A first version of our approach can found in this paper [1]. While these sensors are small and use little power they are still micro-chip-sized and sensitive to humidity, heat, etc. In short: it is difficult to wash an acceleration sensor.

In this thesis a new type of sensor - knitted, resistive, stretch-sensors - shall be investigated for activity recognition (see for example this paper [2]). These could for example easily be sown into clothing and pose less problems with water and humidity - truly washable computing. For this thesis such sensors shall be connected to a computer using a smart-it [3], data recorded and initial recognition experiments performed.

Further Information

Environment

Begin: as soon as possible
Development Environment: Hardware, Linux, Machine Learning
Prerequisites: Experience in soldering and machine learning desireable, but not strictly necessary
Supervision: Nicky Kern
< vorname.nachname AT informatik DOT tu-darmstadt DOT de >
S02/02 B111, Tel. 16-6182

Published 2.12.2004, Nicky Kern


Evaluation Strategies of Wearable and Mobile Applications (SA, DA)

The Design and Systematic Evaluation of Wearable and Mobile Application is a little understood problem. In this thesis a simple mobile application shall be developed and (multiple) evaluation methods evaluated regarding their suitability for mobile applications.

The evaluation of desktop user interfaces is a well understood problem. Every year conference such as CHI [1] publish yet another evluation of such interfaces. The evaluation of applications in a real, uncontrolled, mobile setting has been little tried.

In this thesis a simple, but novel, mobile application shall be implemented and evaluated. The application could for example be a organizer-controlled ring-tone setting for a mobile phone, but other are certainly possible. This application shall be evaluated in a lab and/or field user study.

If you are interested in this thesis or have any questions please feel free to send me an e-mail or pass by my office any time! (ich spreche natürlich auch Deutsch:)

Further Information

Environment

Begin: as soon as possible
Development Environment: Java, Mobile Phone
Prerequisites: Experience in user studies and/or statistics will be of hlep
Supervision: Nicky Kern
< vorname.nachname AT informatik DOT tu-darmstadt DOT de >
S02/02 B111, Tel. 16-6182

Published 2.12.2004, Nicky Kern


Location Estimation using Wearable Sensors (SA, DA, HiWi) (vergeben)

Emerging applications in the field of mobile, wearable and ubiquitous computing try to include more of the user's context, in order to better adapt the user interface to the user. Ideally the computer would proactively act on the user's behalf.

While many different kinds of context are possible and useful, the user's location allows to tell many things about the user's overall situation. GPS is the most popular system for location estimation outdoors, but it fails inside buildings, where people spend most of their time.

In this Semester- or Diploma-Thesis we would like to explore estimation of a user's location using wearable sensors. While these sensors only provide relative location information, they might help improve the accuracy of other approaches that yield absolute data, such as the one used by  the PlaceLab system. We will use the XSens [1] sensor kit, which includes acceleration and magnetic field sensors that can be used to determine orientation and relative location. The thesis will involve setting up the sensor kit, design, implementation of algorithms for location estimation, and possibly integration with an approach such as PlaceLab.

If you are interested in this thesis or have any questions please feel free to send me an e-mail or pass by my office any time! (ich spreche natürlich auch Deutsch:)

Further Information

Environment

Start: as soon as possible
Development Environment: C/C++, Windows
Prerequisites: Experience in C/C++ will be of help
Supervision: Nicky Kern, Tam Huynh
< vorname.nachname AT informatik DOT tu-darmstadt DOT de >
S02/02 B111, Tel. 16-6182

Published 2.21.2004, Tam Huynh


A Journaling Tool for Wearable Computers (SA, DA, Hiwi) (vergeben)

In Wearable Computing it is the goal that the computer perceives the world from a user's perspective -- "seeing what he sees, hearing what he hears and feeling what he feels". Based on sensor data it can infer the user's context, such as the user's current activity ("walking", "riding bike", "shopping"), estimating his current location or even predicting what he will do next.

One way to obtain annotated data for classifier training is to equip users with sensors, record data and annotate it afterwards. These annotations, together with the sensor data, can later be used as training data for machine learning models.

In this Semester- or Diploma-Thesis, we would like to improve on an existing journaling tool for offline browsing and annotation of data collected from a wearable computer. There is a basic implementation available (developed at the MIT Media Lab) that we would like to extend:

  • handling of video, audio, acceleration and other types of sensors
  • a file format for annotations that is based on an open standard such as XML
  • efficient handling of large volumes of sensor data
  • an interface to Matlab for postprocessing of journal data

Further Information

Environment

Begin: as soon as possible
Development Environment: Java (and maybe Matlab)
Prerequisites: Experience in C++, Java and/or Matlab will be of help
Supervision: Nicky Kern, Tam Huynh
< vorname.nachname AT informatik DOT tu-darmstadt DOT de >
S02/02 B111, Tel. 16-6182

Published 2.12.2004, Nicky Kern


Design and implementation of a system for processing of flight data on tablet PC (DA)

(vergeben)

Besides controlling air traffic, Deutsche Flugsicherung GmbH also brings the airlines trafficking German airspace to account. All information from national and international airports and air traffic control centers concerning take-offs, landings and flight routes is here checked for correctness and completeness before the invoices are sent out. Every day, about 8000-10.000 flights have to be verified. The flight information is printed out on paper lists, and each person responsible for the flight data processing goes through three thick binders of lists every day. Briefly described, the lists are leafed through, any actions needed are noted on the paper and later entered into the computerized system.

Today, the flight data processing is performed through working with a combination of the paper-based list, annotation tools like coloured pens and post-its, and the computer for entering changes, which is a highly cumbersome procedure. In order to improve the working conditions for the affected people, DFS now wants to develop a tablet PC-based system, where all validation, annotation and alteration tasks can be performed efficiently and ergonomically.

The main goal of this diploma thesis is to develop a prototype of such system, using an iterative design/development process where a number of different input modalities are evaluated, all in close cooperation with the responsible users, i.e. the end user of the system.

If you are interested in this thesis or have any questions please feel free to send me an e-mail (ich spreche auch Deutsch:)

This diploma thesis is external, and the diploma candidate will ideally sit mainly at the DFS head offices in Langen.

Environment

Begin:

As soon as possible

Prerequisites:

  • Completed intermediate diploma (Vordiplom)

  • Good-very good knowledge in Java programming (Swing)

  • Basic knowledge about tablet PC programming would be ideal

  • Interest in and ideally also some knowledge about human computer interaction design methods

Supervision:

Victoria Carlsson

< nachname AT mis DOT tu-darmstadt DOT de >

S02/02 B111, Tel.16-6227

Published 22.09.2005, Victoria Carlsson

by webmfritz last modified 2006-05-29 14:41