This page lists some currently planned thesis projects. There will be additional projects as time and resources permit. For examples, of past projects associated with prof. Maguire (and some others) please see the annocements of past thesis presentations.
This thesis project concerns the design, implementation, and evaluation of a context aware SIP user agent which can provide secure multimedia VoIP calls and presence information. For example, it should enable users to make use of multimedia (and perhaps other) devices around them. For example, the user should be able to send the video stream to a network attached data projector or large screen display; while sending the audio to a network attached speaker; while providing outgoing video images of an artifact in from of them via a network attached camera -- with all of the streams being managed by the user's device.
This project builds upon several previous thesis projects in this area. Based upon this earlier work, one of the important protcols is the Service Location Protocol (SLP). It is important to understand just how much of this protocol is needed on constrained devices to support dynamic service discovery and to identify an efficient mechanism to establish trust relationships between the user's device and devices located near the user.
A goal is to minimize the effort required by the user to make used of these nearby devices while protecting the user and their device from "rogue" device. Thus to move the video from your handset (the user's personal device) to a large display and the audio from your headphones to speaker in the room and the reverse is not simply redirection of the media stream, but knowing when and when not to do so. Using context aware device aggregation the user should be able to make use of the devices around them as they wish they could -- while minimizing the effort of the user (i.e., the user should not have to spend their life configuring their device(s)).
Note: There is one student already involved in this project. There is room in this project for additional students.
(all thesis projects below are available from http://web.it.kth.se/~maguire/DEGREE-PROJECT-REPORTS/index.html)
One of the projects that I would like to see done is a port of minisip to the Symbian platform (which has a SIP and RTP implementation), but has not implemented SRTP and MIKEY. I think that this would be a very interesting system from both a theoretical point of view (in how to integrated this with the existing stack - for which I have some ideas) and from a practical point of view as there have not been SRTP and MIKEY implementations in handsets. (Contrast with the multiple device below in "Secure personal telephony gateway #2".)
The is also an interest in exploiting set top boxes with storage to provide peer-to-peer video content distribution, see for example: Dario Vodopivec, On-demand Television combined with non-real-time Peer-to-Peer Content Delivery for Television Content Providers, Masters thesis, Royal Institute of Technology (KTH), School of Information and Communications Technology, TRITA-ICT-EX-2010:211, August 2010. This can be seen as a way to exploit traffic shaping on a longer time scale.
A growing area of research is haptic (tactile) interaction. It has been shown that use of haptic feedback can improve the performance of users when doing tasks such as filling out spreadsheets and defining regions of interest for radiation therapy treatment planning. Additionally, haptic interaction is often a part of tele-surgery. Traditionally these systems used low delay, high reliability links. A research question is to character the impact of delay and delay variance on haptic applications (perhaps in a simlar way to the R. G. Cole and J. H. Rosenbluth modeled the effect of delay on voice over IP performance).
Note that this thesis project would probably interact with the following thesis project.
For children who are receiving radiation therapy for cancer, it is very important that the organ contours used for the radiation treatment plan be as accurate as possible. In children, the organs are reduced in size, making the task of defining such organ contours more difficult than for an adult. The goal of this project is to increase the accuracy of this contour definition process (the task ) while not increasing the time required for the task. Reducing the required time required to perform this task would be an added benefit.
The haptic device is capable of producing several different kinds of force: spring loading, viscous, direct, frictional, and hard surface. The 3D visualization software can compute functions over the pixels or voxels, to, for example, define edges using several different filters. At present, the software implements a gradient force. The haptic force produced from this gradient can be used as is or can be modified by an exponential. Additionally, either of these can be combined with a viscous force based on the velocity and direction of the haptic device movement. This existing version has been shown (in an earlier exjobb) to reduce the time required to draw a contour, although the number of volunteers was not sufficiently large to indicate if there was a statistically significant difference in the accuracy, there was a significant reduction in the time required.
This exjobb will (1) design an experiment which will use different types of haptic forces, (2) conduct this experiment using a number of volunteers, and (3) analyze the results to determine which gives the best improvement with regard to accuracy and time in comparison with the null hypothesis - i.e., without haptic force.
In Ymir Vigfusson, Hussam Abu-Libdeh, Mahesh Balakrishnan, Ken Birman, Robert Burgess, Gregory Chockler, Haoyunan Li, and Yoav Tock, "Dr. Multicast: Rx for Data Center Communication Scalability", Eurosys 2010 - the authors claim tha there is a tragedy of the commons due to the limited availablity of IP multcast state space on network interface and network switches (see their table 1) and that when multiple applications are using multicasting that "filtering becomes ineffective when a large number of groups are used, and this can burden end-host kernels with high rates of unwanted traffic, overwhelming receivers who in turn begin dropping packets."
While their paper uses sources from 2004 (their reference [15]) and 2008 (their reference [22]) for background. Their own experiments show that with an increasing number of multicast groups that the packet loss rate increases - despite the fact that the transmitted packet rate was constant.
The topic of this thesis project would be to perform the packet loss experiments that they describe in their section 2.4 and plot in their figrue 1 - in both a local setting (i.e., attached to a single switch) and on a wide area basis (i.e., across multiple switches). The later experiments might examine the impact on ISP's access networks that are supporting multicast.
There have been a number of attempts to speed up the performance of DNS - ranging from converting to a database to specfic DNS providers. Today two such DNS providers are: Google and OpenDNS.
Why is DNS performance interesting? See for example: http://code.google.com/speed/public-dns/docs/performance.html.
This batchelors thesis project would measure the performance that a user might experience using DNS caching, using their local ISP, using one or more of the special DNS providers. Another goal is to understand what users might do to improve their effect DNS performance.
Add support for a personal telephony gateway to minisip to create a secure personal telephony gateway. Test this software using hardware based upon the Tiger 560B telephony chip.
This is a project well suited to a pair of students, one with knowledge of C++ (as minisip is implemented in C++) and the other with knowledge of device drivers (as the hardware is attached via a USB interface).
Using an existing ATA device create a personal telephony gateway to enable SRTP and improved security for customers using an ATA that does not provide security.
Consider for example a Tilgin AB Vood 322 device + PC which could support MIKEY and SRTP.
IPv6 is seeing increasing use and is essential if "everything" is going to be attached to the network. To further the goal of making appliances, sensor, etc. as source, sinks, and sources+sinks - it is essential that various devices become "network enabled". This series of batchelors thesis projects concerns how to network enabled appliances. There should be opportunities to work with students in the IT Product Realization program, SOC, etc.
A first step in this area is the thesis: Thor Hådén, IPv6 Home Automation, Batchelors thesis, Royal Institute of Technology (KTH), School of Information and Communications Technology, TRITA-ICT-EX-2009:28, June 2009
Today many networks are optimized for downloading (i.e., a transfer through the networking infrastructure to a host). Examples, include web browsing, audio and video streaming from traditional media providers, etc. One of the coming trends is increasing generation of content (even if it is only a sensor saying "the temperature in the basement has fallen below 8 degrees C"). This will result in increasing amounts of traffic in the uplink direction. This series of batchelors thesis projects concerns how to accelerate the growth in the amount of uplink traffic. (Note that the links that we are concerned with can be both wired and wireless.)
The last blocks of IPv4 addresses will likely be assigned this year. This means that there will not be new addresses, this implies either a partitioned network (where some parts of the network are unreachable from other parts of the network) or a transition to IPv6.
The aim of this thesis project is to investigate and help users to transition to IPv6. This thesis will need to address technical, political, and economic issues to facilitate increasing IPv6 adoption in Sweden. For example, this project might generate a set of short articles - suitable for a newspaper - on how users can and should transition to IPv6.
Given the user by SL of RFID cards there will be an increasing number of users with RFID cards. An interesting question is: Can these be used to provide an inexpensive source of context information? One method to start to explore this is incorporating simple RFID readers into mobile platforms. This project would involve adding a RFID reader to a PDA to use the RFID card serial numbers as a source of context information. Note that this project does not concern reading the contents of the RFID card, but only reading card serial nunmbers.
This thesis can build on work of an earlier MS thesis: Antonio Aguilar, A Patient Identification System using RFID and IEEE 802.11b Wireless Networks, Masters thesis, Royal Institute of Technology (KTH), School of Information and Communications Technology, COS/CCS 2007-13, March 2007
For information contact maguire@it.kth.se
Last update:
5 October 2011