All projects
AI for IoTActive
Federated Edge AI for Resilient IoT
Federated learning and fault-tolerant architectures for running AI directly on constrained, safety-critical IoT devices.
Sensor fusion diagram over a fault-tolerant edge computing architecture
Overview
Running AI at the edge means working within tight compute and connectivity budgets — and, in safety-critical settings like gas-leak detection, tight tolerance for failure. This project line combines multi-modal sensor fusion with federated learning to train TinyML models without centralising raw sensor data, and designs fault-tolerant edge architectures for deployments that can’t afford downtime.
Related publications
- Multi-modal Sensor Fusion and Federated Learning for TinyML on Resource-Constrained IoT Devices — International Journal of Parallel, Emergent and Distributed Systems, 2025
- Resilient Edge Computing: An Elixir-BEAM Architecture for IoT Gas Leakage Detection — Proceedings of the 2025 10th International Conference on Cloud Computing and Internet of Things
- AI-Driven Anomaly Detection in IoT Systems: Techniques and Applications — book chapter, Innovations and Challenges in Computing, Games, and Data Science, 2025
Status
Active, spanning both the learning side (federated TinyML) and the systems side (fault-tolerant edge architecture) of resilient IoT deployment.