Student project ideas
Doing a Master’s thesis on a research topic allows you to try new
ideas, challenge your creativity and to get a taste of what research
is all about. If things go well, you might also get the change to
present your work at some international conference or workshop.
Below is a list of ideas for thesis work. If you are interested in
one of these or have some ideas of your own, don't hesitate to write
me an email or drop by my office. You can also look at the list of
completed theses where I have been
supervisor or examiner.
Smart vehicle coordination using Bitcoin-inspired algorithms
The virtual currency Bitcoin is having a disruptive effect on the financial sector. While initially considered as a curiosity, the technology behind bitcoin, so called blockchains now receive investments from companies like Goldman Sachs and IBM. The potential of blockchains go beyond that of currencies, allowing decentralised coordination and smart contracts, and IBM is advocating it as a basis for coordination in the Internet of Things. With the help of Turing-complete script languages such as being used by the Ethereum platform, the possibilities are endless. In this project the goal is to design, implement and evaluate a blockchain-based coordination mechanism for vehicular networks. The project involves both theoretical analysis and practial implementation. The scope is suitable for at least on 30hp project, possibly also a combination of two 16hp projects.
Attacks and mitigations for the connected car
As cars become more advanced and connected, the number of incidents where cars are being hacked also increase. An promising technology in future intelligent transportation systems is the ability to cooperate in groups of vehicles. Such cooperation require group membership protocols. This project will investigate vulnerabilities and attacks against such group membership protocols, and suggest mitigations and counter measures that can protect the systems under certain circumstances. This project is suitable for students with an interest in information security.
Theoretical analysis of distributed coordination
Autonomous cars are already on the horizon. The next generation of
self-driving cars will interact with other vehicles using wireless
communication to coordinate their actions and and enable more
efficient traffic solutions than human drivers can
accomplish. However, distributed coordination is still an open
problem, especially in situations where the local actions taken by
vehicles propagate to other parts of the system causing congestion and
unpredictable delays.
This Master’s thesis project will consider an abstraction of this
problem in a theoretical setting to investigate the asymptotic
behaviour of distributed coordination under varying
circumstances. Consider a system composed of N mobile agents
that move around in a directed graph G=(V,E). At any point in
time a node v ∈ V can contain either zero or one mobile
agents. Obviously |V| ≥ N.
Each agent can communicate locally with agents located in
neighbouring nodes, and nodes up to h hops away (h
being a parameter under study). Moreover, for each agent a there is a
set of nodes Va such that the agent wishes to visit all
nodes in Va as many times as possible. The performance of an
agent is determined by the number of visits to the node
in Va with the least number of visits.
The task of this thesis project will be to investigate the bounds
on achievable performance under varying conditions (starting with a
restricted setting, and using increasingly complex models). A suitable
student should be comfortable with mathematical reasoning, and
interested in theoretical analysis of distributed algorithms.
Vehicular group membership resilient to malicious attacks
Next generation of heavy-duty vehicles will use wireless communication
and advanced sensor technology to enable platoon driving where
multiple trucks coordinate their driving in order to save fuel. Such
advanced applications require vehicles to be able to accurately
determine which other vehicles that are participating in the
cooperation and to detect and adapt to changes as vehicles come and
go. A key challenge is how such a membership protocol can withstand
attacks from malicious actors in the environment (other vehicles or
even from vehicles within the platoon).
This project is focused on implementation of a group membership
protocol for vehicular platoons. The core idea of the protocol is to
combine the use of communication, on-board sensors, and rearrangement
of the platoon to detect malicious actors. The student will implement
the protocol in a simulation environment and evaluate the ability to
detect different attack types. The work is performed in cooperation
with Scania.
|