ELIAS Open Call Winners

The European Lighthouse of AI for Sustainability (ELIAS) unites top AI researchers and innovators to tackle pressing challenges such as climate change, societal shifts, and the energy crisis—while advancing responsible and ethical AI.

We are thrilled to introduce the winners of our 1st Open Call, turning innovative AI methods, software, and benchmarks into real-world applications that drive sustainable and socially conscious impact.

LUCIA

PINNSSHS

TENSORIA

AI-REFRAME

Project Overview

LUCIA

Learning Urban Cognitive Intelligence for Autonomous buildings

LUCIA is a lightweight AI platform that helps buildings operate efficiently and comfortably. By connecting to existing sensors, devices, and cameras, it detects inefficiencies such as unoccupied areas with lights left on or open windows affecting HVAC (Heating, Ventilation, and Air Conditioning) systems, and proposes or applies adjustments to reduce waste while maintaining comfort.

Designed for low-power edge deployment, LUCIA minimises reliance on cloud services and processes sensitive data locally to protect privacy. Its decisions are explainable and auditable, and human operators can always review or override automated actions. The platform is modular and vendor-neutral, enabling easy integration with existing building-automation systems and future extension to additional subsystems. By prioritising privacy-by-design, explainability, and local intelligence, LUCIA aims to support greener building operations and more efficient facilities management, aligned with European goals for sustainable and trustworthy AI.

Challenges

Why Sustainability Matters:

Buildings are major consumers of energy, and inefficient operations lead to unnecessary waste and emissions. By enabling AI-driven, autonomous optimisation of energy use while preserving comfort and privacy, LUCIA directly reduces resource consumption, supports greener operations, and contributes to a more sustainable, resilient built environment.

Energy Efficiency
Cybersecurity & Infrastructure

Technical Details

AI Approach

LUCIA combines edge-optimised deep learning with a multi-agent AI architecture to enable smarter, energy-efficient building management. Vision models, based on LLaVA multimodal frameworks, detect room occupancy and open windows from local camera frames, while on-device LLM-based decision agents (Ollama) interpret sensor data, reason about context, and propose energy-saving actions. Rule-augmented prompts and context-aware reasoning ensure that all decisions remain explainable, safe, and auditable, while lightweight, reinforcement-inspired adaptive logic allows agents to refine their actions over time based on past outcomes and user feedback.

By analysing real-time telemetry and visual cues, LUCIA identifies inefficiencies—such as empty rooms with lights on or HVAC running while windows are open—and automatically or suggestively adjusts systems to reduce unnecessary energy use. This demand-driven, adaptive control not only minimises consumption of the most energy-intensive building systems but also preserves comfort and human oversight.

With its privacy-preserving design, explainable AI, and modular, vendor-neutral deployment, LUCIA enhances operational efficiency, supports inclusion by simplifying building management, reduces emissions, extends equipment lifespan, and provides a scalable solution for greener, more resilient buildings across public and private infrastructures.

The mission of LUCIA is to optimise energy efficiency and resource management in smart buildings through AI-driven, autonomous control, while ensuring privacy, scalability, and ease of use for diverse stakeholders.

Smarter energy management and waste reduction
Explainable AI for human-in-the-loop control
Privacy-preserving, edge-based processing
Potential reduction in HVAC and lighting energy usage

LUCIA aims to reduce energy waste, lower carbon emissions, improve system resilience, and empower building managers with privacy-preserving AI tools for smarter, inclusive building operations.

{

We are dedicated to harnessing the power of AI to build smarter, more sustainable solutions that optimise energy use, enhance efficiency, and contribute to a greener future.

Raquel Barroso Reyes

R&D&I Director

César Fernández González

CTO

EWALA IT Services

Ewala IT Services SL is an Asturian-based technology company specialising in industrial (OT) and IT cybersecurity, IoT-powered automation, and AI-driven monitoring. The company develops custom threat-intelligence tools, advanced asset monitoring, and real-time anomaly detection solutions, supporting enterprises and public organisations in secure, efficient, and resilient operations.

Related Initiatives

LUCIA

PINNSSHS

TENSORIA

AI-REFRAME