Description du projet
Sécurité et degré de confiance des systèmes cyber-physiques autonomes
L’utilisation croissante de systèmes cyber-physiques autonomes (ACPS), tels que les voitures autonomes, qui s’appuient sur l’IA, est très prometteuse pour les futurs développements de l’IA. Le manque d’assurance concernant la qualité et la sécurité de fonctionnement des solutions ACPS actuelles suscite toutefois des inquiétudes. Soutenu par le programme Actions Marie Skłodowska-Curie, le projet InnoGuard se propose de développer de nouvelles méthodes d’assurance de la qualité de l’ACPS en créant des programmes de formation sur mesure pour des chercheurs en début de carrière. Le projet entend automatiser l’évaluation de la qualité de l’ACPS et à faire évoluer le comportement du système. Il renforcera également la durabilité environnementale, l’efficacité énergétique et la fiabilité des méthodes d’IA pour l’ACPS, tout en améliorant la fiabilité globale et la conformité légale.
Objectif
InnoGuard addresses novel challenges imposed by quality assurance of Autonomous Cyber-Physical Systems (ACPS), which have integrated Artificial Intelligence (AI) components. Their use in our daily lives is increasing, such as in public transport systems, highlighting the need for novel, robust development methodologies to ensure their dependability, whose compromise could have severe consequences, as exemplified by various incidents (e.g. accidents caused by Tesla's autopilot system). To ensure ACP dependability, InnoGuard targets developing novel methods for ACPS quality assurance by creating a tailored training program for Early-Stage Researchers (ESRs), with scientific objectives including methods to automate ACPS quality assessment and behavior evolution using AI techniques and enhancing ACPS dependability through real-time security, privacy, and uncertainty handling solutions. Additionally, InnoGuard seeks to improve the trustworthiness of AI methods, enhance environmental sustainability by increasing the energy efficiency of ACPSs with the usage of AI methods, including Large Language Models (LLMs), and validate such techniques in open-source contexts such as Robot Operating Systems (ROS)-based systems. Ultimately, InnoGuard will deliver novel techniques and methodological principles for ACPS quality assurance, ensuring high trustworthiness, reliability, and legal compliance. The project's holistic approach encompasses technical advancements and training initiatives, with a broad plan to disseminate results to industry stakeholders, associations, and the wider public. Through these efforts, InnoGuard strives to elevate the status of ESRs as future experts in ACPS engineering.
Programme(s)
- HORIZON.1.2 - Marie Skłodowska-Curie Actions (MSCA) Main Programme
Régime de financement
HORIZON-TMA-MSCA-DN - HORIZON TMA MSCA Doctoral NetworksCoordinateur
20500 Arrasate
Espagne