Skip to main content
Aller à la page d’accueil de la Commission européenne (s’ouvre dans une nouvelle fenêtre)
français fr
CORDIS - Résultats de la recherche de l’UE
CORDIS

CONTEXT-AWARE, ENERGY-EFFICIENT, AND TRUSTWORTHY PERCEPTION SYSTEMS FOR CCAM WITH DISTRIBUTED AI AND NEXT-GENERATION SENSORS

Objectif

Recent European trials (e.g. Hi-Drive) and early driverless robotaxis deployments have shown that robust environment perception remains a key bottleneck for CCAM: today’s perception stacks still struggle with dense VRU presence, occlusions, and degraded visibility (fog/rain/night/reflective surfaces). Yet improving robustness by adding more sensors and larger models raises compute demand, power consumption, heat, and costs, or shifts the burden to costly roadside infrastructure. A central barrier, therefore, remains: how to deliver reliable and trustworthy perception without unsustainable energy and cost growth?

CONCENTRA proposal aims to overcome this barrier through an integrated set of transformative enablers spanning novel sensing systems, emerging AI models, and novel systems engineering approaches. We will combine next-generation sensors with context-adaptive sensor selection and modular fusion to optimise accuracy-power trade-offs at runtime. Solid-state LiDAR and 4D imaging radar, both featuring embedded AI from our industry partners, will reduce latency in the sense-think-act loop while lowering power consumption and cost relative to spinning LiDAR baselines. Effective runtime integrity monitoring will detect degradations and support timely safe fallback (e.g. remote assistance requests) when alert limits are exceeded, which is crucial for higher levels of autonomy (e.g. Level 3 and 4). Distributed AI optimisation will let resource-limited vehicles benefit from powerful cloud-trained ensembles, while multi-RAT V2X connectivity will provide resilient collective perception with path redundancy. Standardised interfaces for plug-and-play reuse of selected sensors and perception modules will support circularity and cost-effective scaling, complemented by exploratory mapping of sensor-fusion modules to chiplet-based computing. Validation will culminate in physical proof-of-concept demonstrations on public roads and an advanced smart-city site.

Champ scientifique (EuroSciVoc)

Mots‑clés

Programme(s)

Thème(s)

Régime de financement

HORIZON-RIA - HORIZON Research and Innovation Actions

Voir tous les projets financés dans le cadre de ce programme de financement

Appel à propositions

(s’ouvre dans une nouvelle fenêtre) HORIZON-CL5-2026-01

Voir tous les projets financés au titre de cet appel

Coordinateur

THE QUEEN'S UNIVERSITY OF BELFAST
Contribution nette de l'UE
€ 835 350,00
Adresse
UNIVERSITY ROAD LANYON BUILDING
BT7 1NN BELFAST
Royaume-Uni
Région
Northern Ireland Northern Ireland Belfast
Type d’activité
Higher or Secondary Education Establishments
Liens
Coût total
Aucune donnée

Participants (10)