Research and expertise

​​​​There are currently a number of research projects taking place at C11 Cyber Security and Digital Innovation Centre, led by researchers and PhD students from the University of Gloucestershire.

You can find out more about our researchers and current research projects below.

 

Research project

Applied Artificial Intelligence

Dr William Sayers is currently conducting research on Physical Systems Modelling with Artificial Intelligence, this area of research investigates the optimisation of physical engineering systems, using multi-objective optimisation techniques. Due to the computing demands of modelling these systems, artificial intelligence based meta-models show promise in estimating the model results.

This requires modifying multi-objective optimisation techniques to include the use of a pre-trained and online-trained deep neural network based discriminator, which will classify engineering models as promising, or non-promising. The promising models can then be evaluated fully, thus ensuring that compute power is focused on models with good potential.

Effectively, the artificial intelligence models learn what aspects of a model are promising, and allow the optimisation algorithm to make “leaps of intuition” where exploration yields new promising model setups. Initial results, based around flood risk reduction and water distribution system reliability, have shown very promising results.

He is also working with Dr Hassan Chizari from UOG on his project generating strong cryptographic keys, described on this page, as well as on several KTP bids with Professor Shujun Zhang​.

Find out more about Dr. William Sayers

 

Research project

Towards greater security within IoT environments

Using Big Data Analytics together with Artificial Intelligence (A.I), this area of research looks into the different means by which the large amounts of data generated by different devices as well as IDS (Intrusion Detection Systems) can be leveraged to detect the presence of cybersecurity attacks in both IoT as well as in traditional computer networks. This involves using different machine and deep learning algorithms to automate the feature learning and detection process.

Current work, led by Dr Thomas Win​, combined the trained model with real-time big data streaming technologies (e.g., Apache Spark) to detect attack presence in the large amounts of data collected. This involves real-time collection of data from various sources within the network, performing statistical analysis of the feature characteristics to generate a feature vector. The generated feature vector is then passed into the trained deep learning model to detect cybersecurity attack presence.

A related PhD study, supervised by Dr Salah Al-Majeed and Prof Kamal Bechkoum, is about developing a security system on chip for IoT devices. The aim is to develop a solution that enhances security between Edge nodes and Preprocessing Gateways through implementing cryptographic algorithms at the perception Layer.

Prof Bechkoum is also working with Sepideh Mollajafari on “security issues in blockchain technology”.

Find out more about Dr. Thomas Win, Dr. Salah Al-Majeed & Prof. Kamal Bechkoum​

 

Research project

Treating Malicious Software on Windows Based Operating Systems

This area is led by Peter Bentley and looks at Advanced Persistent Threats (APTs). The research investigates ways in which it is possible to treat APTs before, during and after the malware has been laid down on the victim's machine. Current scope is limited to desktop and laptop computers with hard disk drives but can be extended to other devices.

 
Research project

Generating strong cryptographic key from body physiological signal in securing medica implants communication

The latest smart Implantable Medical Devices (IMDs) are providing many benefits to patients and health care providers. With real-time monitoring, the IMD is able to adjust its activity based on patient need, receive firmware updates to improve functionality and contribute a stream of real-time data for medical research. With an aging population, increases in chronic disease, and technological breakthroughs, Harvard Business Review estimate the market for implantable medical devices stands at $398 billion, anticipating growth of 3% per year until at least 2022. There are currently millions of people relying on implantable medical devices.

Whilst the benefits of smart implantable medical devices are clear, they are not without their concerns. Namely, their vulnerability to malicious interference as a result of cyber-attack.

Recently, a set of new solutions has been proposed to use body physiological signals as the source of randomness for this secure communication. In these solutions, both the gateway and IMD start to read a physiological signal together and they simultaneously generate a communication key from that signal.

Work undertaken by researchers at UOG, under the leadership of Dr Hassan Chizari, showed that, despite the current claims, those physiological signals could not be used as a strong source of randomness.

More work is required to develop a secure method of authentication between IMD and its gateway. This would eliminate a large set of vulnerabilities in IMDs that are related to providing a secure communication and authentication between the IMD and its gateway.

Find out more about Dr. Hassan Chizari​