Charles Hutchins

MINDS CDT, 2020

MINDS alumnus | Knowledge Exchange and Enterprise Fellow, University of Southampton

Keeping drone networks safe from hidden attacks

Research focus: Cybersecurity for flying ad hoc networks (drone networks)

Charles Hutchins is a MINDS alumnus whose doctoral research investigated how to detect and respond to covert cyberattacks in self-organising drone networks. He is now a Knowledge Exchange and Enterprise Fellow at the University of Southampton.

Career at a glance

MINDS status	Alumnus
MINDS metadata	MINDS CDT, 2020
Doctoral research	Cybersecurity for flying ad hoc networks
Research approach	Machine learning, game theory and network simulation
Key output	Open dataset for early detection of grey hole attacks (Scientific Data, 2025)
MINDS-linked experience	Placement at The Alan Turing Institute
Current role	Knowledge Exchange and Enterprise Fellow
Current organisation	University of Southampton (Mechanical Engineering)

Hidden attacks in drone networks

When groups of drones work together, for example in a search-and-rescue operation, they often need to share information without relying on masts, base stations or other fixed ground infrastructure. They can do this through a flying ad hoc network: a self-organising communication network in which each drone can pass messages on to the others.

Information travels across these networks in small units called data packets. If one drone has been compromised, it can carry out a grey hole attack. Rather than blocking all traffic, the compromised drone quietly drops some packets while forwarding others.

This makes the attack hard to spot. Drones move in and out of range all the time, so some packet loss is a normal part of how these networks behave. A defence that reacts too strongly to ordinary disruption could harm the network’s performance, while one that misses a genuine attack could leave the group of drones exposed.

Charles investigated how drone networks could recognise grey hole attacks early and choose a proportionate response. His doctoral thesis made three main contributions:

• a time-series dataset of drone-network traffic, with an early classification model designed to identify grey hole attacks quickly and accurately;
• a game-theoretic framework for ranking defence protocols across different types of attack and network conditions;
• a trust mechanism that takes account of a drone’s current movement as well as its past behaviour.

He also questioned how drone-network security systems are evaluated, arguing that optimising for a single performance measure can make a system less robust across the wider range of threats it may face in practice.

Why the research matters

How the research works

Charles’s work combined machine learning with game theory. Machine learning was used to classify network behaviour from time-series data: sequences of measurements recorded over time. Early time-series classification aims to reach a reliable decision before the full sequence has been observed, so that an attack can be recognised sooner.

Game theory is a way of modelling decisions where different actors or systems may have competing objectives. In this setting it offers a structured way to compare defensive options, weighing the cost of reacting to a false alarm against the cost of failing to act on a genuine attack.

The work was developed and tested through simulations in the NS-3 network simulator.

Research and outputs

Charles’s principal published output is an openly available dataset for studying the early detection of grey hole attacks, described in a Data Descriptor in Scientific Data. The dataset, FAN-GHETS24, was generated by simulating drone-network traffic in a search-and-rescue scenario. It contains labelled sequences covering normal behaviour and several types of grey hole attack, giving other researchers a shared resource for developing and comparing detection methods.

• Hutchins, C., Aniello, L., Gerding, E. and Halak, B. (2025) ‘A flying ad-hoc network dataset for early time series classification of grey hole attacks’, Scientific Data, 12, 1431. Read the Data Descriptor (doi:10.1038/s41597-025-05560-1)
• FAN-GHETS24: A Flying Ad Hoc Network Dataset for Early Time Series Classification of Grey Hole Attacks (2024). Access the dataset on Zenodo
• ‘Using contextual reinforcement learning to design FANET defence protocols to combat grey hole attacks’, IEEE Network Operations and Management Symposium (NOMS 2024). Read the paper (doi:10.1109/NOMS59830.2024.10575177)
• ‘MANET-Rank: A Framework for Defence Protocols against Packet Dropping Attacks in MANETs’, IEEE Network Operations and Management Symposium (NOMS 2024). Read the paper (doi:10.1109/NOMS59830.2024.10575601)
• Doctoral thesis (2026): An investigation of defence protocols for the mitigation of grey hole attacks in flying ad hoc networks

Charles’s wider publication record is available on his Google Scholar profile and ORCID record.

MINDS doctoral project: Game Theory and Multi-Agent Systems to Detect and Respond to Cyber Intrusions in IoT Environments

Placement at The Alan Turing Institute

As part of his MINDS training, Charles completed a placement hosted at The Alan Turing Institute. The placement took his security thinking into a different setting: AI accelerator chips used in applications such as driverless vehicles, where fast and accurate image recognition is safety-critical.

The work involved building these chips virtually, developing attacks that could trigger misclassification, and designing testing programmes to detect and remove those vulnerabilities before deployment.

From MINDS research to knowledge exchange

Charles is now a Knowledge Exchange and Enterprise Fellow at the University of Southampton. In this role he has worked on a project developing products for TriboAI, which uses tribology research and machine learning to detect anomalies and failure modes in machinery through electrostatic sensor data. The work has involved close collaboration with the Offshore Renewable Energy Catapult in Newcastle, including testing materials, visiting operational sites and building the industry relationships that applied research depends on.

Charles’s MINDS experience