Research Themes
Space/Terrestrial Comms and Security
Sample Project Titles for this theme **
Anomaly Detection in Distributed NTN Agents Using MPC
Secure Boot and Firmware Integrity for AI-Driven NTNs
Trustworthy of AI-enabled Management and Orchestration for Multi-Domain Operations (MDO)
Directional Modulation for Secure Anti-Jam Satellite Communications
**Click for more titles and the project descriptions
Autonomous operation of NoN with distributed cloudnative architecture relies on an Orchestrator and Service Management system and sub-systems, typically across multiple autonomous operators, either trusted or untrusted.
In space communication infrastructures (satellite/UAV/High-altitude platform systems) there can be a wide variety of open integration options with terrestrial network platforms, introducing distinct research challenges in performance, energy efficiency and security and privacy in different scenarios.
Network resources provided by global satellite communication systems can be simultaneously shared by multiple regional terrestrial networks, and hence there are key research questions on how we can engineer novel secure/resilient and high performing solutions against potential attacks whilst optimising the system performance.
Trustworthy AI for Secure Future Open Networks
Sample Project Titles for this theme**
Performance and Cost Challenges of Trustworthy AI for Securing Terabit Datacenter Networks
Calibrated Learning for Reliable and Trustworthy AI in Future Open Wireless Networks
Run-time Information Integration for Establishing Machine Learning Trustworthiness
AI Agents for Secure Resilient Future Networks
**Click for more titles and the project descriptions
In recent years, AI has been proposed as a powerful tool for automating network management and orchestration operations.
In the context of future open NoN-environments, interconnected or integrated network components contributed from homogeneous or heterogeneous operators (terrestrial, space, cloud/edge hyperscale providers) are expected to be managed through distributed AI-based network intelligence with their own local operation policies and resolve possible AI-conflicts.
In addition to the technical challenges of end-to-end performance optimality and consistency through harmonised AI decision-making, there are also numerous research questions about trustworthy AI for network security; for instance, how can AI accurately detect and prevent cyber-attacks in future open networks and how to ensure the AI component has not itself been attacked?
Secure and Trustworthy Hardware
Sample Project Titles for this theme**
Trustworthy RISC-V based accelerated AI Computing Architecture
Hardware Security in Compute-in-memory for AI applications-
**Click for more titles and the project descriptions
We need to ensure that the hardware that will underpin future comms infrastructure is secure and trustworthy. Cyber vulnerabilities in hardware cannot be easily patched (unlike software).
An open research challenge is how to deliver trusted and resilient hardware for future fully virtualised communications systems.
Research into postquantum cryptography and linking advanced cryptographic approaches (e.g., identity-based encryption) and hardware security to achieve hardware-rooted security could provide verifiability, anti-tampering, anti-counterfeiting properties to future communication infrastructures.
Such technologies could also be used to mitigate against the possible leakage of sensitive information in integrated communications and sensing systems of future
AI-assisted Physical Layer Security
Sample Project Titles for this theme**
Semantic Driven Covert Communications
Post-Quantum-Ready AI for Privacy-Preserving Physical Layer Security with Federated Learning
**Click for more titles and the project descriptions
Traditional approaches in research and development of the physical layer of communication systems and their security rely on rigorous mathematical modelling of the system and environment (e.g. channel modelling and heuristic systems and algorithms synthesis).
Such approaches inevitably rely on various simplifications that undermine the fidelity of the models and the efficiency of the designs in highly dynamic wireless environments. As the complexity of the environment, e.g. the propagation and channel, and the demand for more efficient and adaptable/reconfigurable physical layers, increases in the context of 5G and 6G, traditional approaches become less effective.
Emerging AI/ML technology offers an entirely new framework that is data-driven for use in the ML-based analysis and design of secure physical layers for future networks. In addition, the nature of physical layer security solutions is characterized by increased ambiguity owing to the presence of attackers whose behaviour cannot be readily modelled but can be learned from the available data. Hence, AI assisted approaches will be investigated to address the research challenges in this theme.