FREEWAY – Automating asynchronous federated learning and edge computing for efficient vehicle operation analytics

The FREEWAY project aims to deliver next-generation digital services for electromobility by enabling asynchronous federated learning (AFL) to address scalability challenges, especially for large heterogenous fleets, integrating advanced edge processing and MLOps workflows to enhance vehicle operational efficiency, safety, and uptime.

FREEWAY supports the development of advanced digital services for electromobility, covering use cases such as energy consumption forecasting, vehicle operation profiling, activity recognition, and anomaly detection.

Central to FREEWAY’s mission is the enhancement, scalability, and expansion of software capabilities on edge devices and back-end systems. The project will incorporate an MLOps approach adapted to asynchronous FL scenarios, establishing robust workflows for data retrieval, model version control, and automated testing, validation, and deployment processes. By demonstrating the feasibility and efficiency of AFL across multiple real-world use cases, FREEWAY will showcase how federated edge intelligence can further substantially benefit both electromobility stakeholders and the broader mobility landscape.