15/06/2026
Presented "Multimodal Federated Learning in Air Quality Monitoring Under Heterogeneous Resource Constraints" at MAC.
This work investigates how Federated Learning can enable privacy-preserving and efficient AI across heterogeneous edge devices such as Raspberry Pi and NVIDIA Jetson platforms. By combining multimodal data sources, including environmental images and tabular sensor measurements, the proposed framework addresses key challenges in real-world IoT systems, including system heterogeneity, communication overhead, non-IID data distribution, and privacy concerns.
The study demonstrates the potential of deploying intelligent, distributed, and privacy-aware AI solutions for air quality monitoring and other real-world applications such as healthcare and smart environments.