Federated AI Orchestration for EV Fleets
ARIEL
ARIEL is a privacy-preserving AI orchestration solution that enables distributed model training across heterogeneous edge environments using Federated Learning. Powered by FedMaestro, it supports secure, scalable, and efficient collaboration without moving raw data outside the local infrastructure.
Project Snapshot
- Selected under the O-CEI Open Call
- Powered by Indigma’s FedMaestro orchestrator
- Focused on privacy-preserving distributed AI
- Validated through Pilot 3 integration
Overview
ARIEL introduces a federated learning orchestration framework designed to support distributed AI across heterogeneous environments. The solution enables multiple stakeholders to collaboratively train machine learning models while keeping data local and secure.
At its core, ARIEL leverages FedMaestro, Indigma’s orchestration engine, which coordinates training workflows, client participation, and model aggregation across edge and cloud environments.
Project Context
ARIEL is a selected project under the O-CEI Open Call, focusing on advancing AI capabilities across the cloud-edge continuum. The project is validated within Pilot 3, which targets the execution of machine learning processes at the edge under realistic conditions.
Key Features
ARIEL combines flexible orchestration, privacy-preserving AI, and edge-ready deployment to support distributed intelligence in real-world settings.
Federated Learning
Enables collaborative model training without sharing raw data.
Edge Deployment
Supports execution on distributed and resource-constrained devices.
Privacy-Preserving Design
Integrates secure aggregation and privacy-enhancing techniques.
Flexible Orchestration
Manages workflows across heterogeneous edge and cloud infrastructures.
Use Cases
ARIEL supports forecasting and optimisation scenarios aligned with Pilot 3 priorities and the broader vision of distributed, trustworthy AI.
Solar Forecasting
Distributed clients collaboratively train models to predict solar generation using local historical signals and optional contextual variables.
Charging Demand Forecasting
Local clients associated with EVs, chargers, or charger groups collaboratively predict near-future charging demand and energy delivery.
Technologies
The project combines modern AI, orchestration, and privacy technologies to enable secure and scalable deployment.
Impact
ARIEL contributes to the development of scalable, privacy-preserving AI solutions by enabling distributed intelligence without compromising data ownership.
The project aligns with European priorities on trustworthy AI and supports real-world deployment across edge environments. Beyond Pilot 3, ARIEL establishes a foundation for broader adoption of federated AI across energy, mobility, and distributed infrastructure domains.