On July 9, 2024, we held a presentation and meeting with our partners from the University of Nicosia Research Foundation (UNRF) – Artificial Intelligence Laboratory. This presentation focused on the AdaptoFlow project, which aims to revolutionize adaptive sensing and machine learning inference in 5G/6G deployments.

About AdaptoFlow:
AdaptoFlow is a cascade project of the TrialsNet EU project. It seeks to enhance TrialsNet’s use cases—namely, smart crowd monitoring (UC1) and smart traffic management in smart cities (UC4) — through the integration of advanced AI/ML techniques, 5G networking and edge computing infrastructure. The project's goal is to shift the monitoring and AI/ML inference paradigm from a traditional, centralized approach to a dynamic, decentralized one, enabling in situ data analysis at the network edge.

Key Mechanisms:
🔹 Adaptive Data Stream Processing: AdaptoFlow dynamically adjusts data collection intensity based on real-time data stream behavior. This ensures efficient data processing and reduces the load on analytics modules during low-volatility periods.
🔹 Energy-Aware Model Swapping: AdaptoFlow supports on-the-fly AI/ML model swapping to manage limited resource capacity effectively. This approach conserves energy while maintaining user experience and application performance.

By implementing these intelligent mechanisms, AdaptoFlow aims to minimize resource wastage, enhance service responsiveness, and maintain acceptable quality of service (QoS) levels.

For more information about AdaptoFlow, visit: https://linc.ucy.ac.cy/index.php?id=315

We look forward to continuing our collaboration with UNRF and leveraging AdaptoFlow to push the boundaries of smart city technologies.