About the project:
Cyber Security is one of the key challenges and a well known issue in the IoT ecosystem. Due to heterogeneity and resource constraints, optimisation, testing and validation of the end-to-end systems is a hard problem which leads to unsecure systems. In recent years, many machine learning and artificial intelligence (ML/AI) solutions have been investigated to overcome some of these challenges, however, these solutions are quite application specific and don't scale. ML/AI based solutions for resource-constrained devices are still in the early stage, and need to be properly tailored to IoT systems.
The proposed work will be based on a concrete IoT platform and a concrete cyber attack, to be surveyed and chosen by the student and advisors during the first year of study (for instance, denial of service attacks on LoRa networks, or timing attacks on networks based on IEEE 802.15.4 standard). The work will be supported by an experimental setup based on simulation of large-scale networks, as well as hardware prototyping of specific scenarios, developed over the first half of the PhD period, aiming to evaluate the potential use of ML/AI techniques to detect an attack and improve network resilience.
The core research will include the formalisation and definition of attack scenarios, the investigation of monitoring techniques and metrics of interest to support attack detection and resilience evaluation, the comparative analysis of ML/AI techniques (to be chosen by the student) and the strategies to deploy such solutions over network nodes, edge devices and cloud infrastructure. The following research questions will then be addressed: What kind of data obtained from network simulation and prototypes can be used for training ML/AI techniques? What are the biases, and how can they be avoided? How accurate are those techniques when detecting attacks, and how effective are they in increasing the network resilience to those attacks? Which deployment strategy provides the best trade-off between detection accuracy, resilience, and network overheads?
This plan provides a well-delimited area within the research landscape in IoT, but at the same time provides enough freedom for the student to choose an IoT cybersecurity problem that is industry-relevant and amenable to the application of ML/AI solutions.
Enquiries: [Email Address Removed]
How to apply:
Candidates must submit an application for a PhD in Computer Science at the University of York and quote this research project on the application. Find out more here at: https://www.york.ac.uk/study/postgraduate-research/apply/
For further information, please contact: Dr Poonam Yadav (https://www.cs.york.ac.uk/people/yadav), email: [Email Address Removed]
Continue with Facebook
