DESCRIPTION :
Offer description : The development of cyber-physical intelligent systems requires aggregating numerous data from heterogeneous devices. Nevertheless, when these systems must ensure critical missions, guaranteeing security properties in an adaptive way for the data exchanged such
as integrity or confidentiality becomes a necessity. These properties are often partly guaranteed by the use of dedicated resources (hardware and software), however with the systems heterogeneity and constraints, the usual resources are not always available. Modeling autonomous systems using the multiagent paradigm allows representing the different roles or actors of the system. The sensors and actuators of the physical system, used to accomplish several objectives simultaneously, can be represented by agents contributing to the decision-making.
Since decentralization is an essential point of multi-agent systems (MAS), information processing can be planned locally at the agent level without requiring centralized control or decision-making; however without excluding feedback for decision-making at a higher level of the system. Agents act autonomously and cooperatively to achieve their local and global objectives within the system.
Project summary: For the H4 project, we propose to develop methods to ensure an adaptive security level for the exchange of data reported by sensors
and used in decision-making of autonomous vehicle supervision systems. A main a question is to study how to secure and make reliable data from constrained devices that do not have specific security elements allowing the implementation of conventional mechanisms, requiring a particular cryptographic effort and advanced key management. The previous work on LCIS allowed the development of a public key infrastructure for multi-agent
systems (MAKI - Multi-Agent Public Key Infrastructure) designed to secure embedded multi-agent systems whose classic characteristics are openness and heterogeneity [ICCAART23, Baudet23]. Moreover, a study on the complementarity of approaches based on digital twins (DT) and multi-agent systems made it possible to set up an architecture for modeling and monitoring communications to secure them [SHIFT24, SECURWARE24].
Each physical component of a CPS (sensor) is represented by a physical agent and its digital twin (virtual agent). Digital twins are used to model and monitor the CPS behavior and take advantage of the feedback loop of the digital twins to simulate and analyze in real time the physical and software states, while SMA are useful ensure communication and distributed decision making. The alliance of SMA and DT aims to improve attack detection
and CPS resilience.
Main goals: A main objective of the thesis is to define methods allowing to collect useful and necessary data for monitoring security and safety at the system communications level. Modeling via SMA allows for the autonomy and cooperation of agents to implement adaptive security based on criteria related to transport or communications status. Decentralization aims to reduce single points of failure and improve scalability, and
the integration of digital twin will be used for real-time monitoring and better understanding of vulnerabilities. A second objective is to build an architecture to take advantage of these two complementary approaches. On one hand, the multi-agent paradigm promotes decentralization and coordination and on the other hand, an architecture with digital twins allows to model surveillance, to strengthen the security of the CPS.
The thesis will thus address the following issue: "How to ensure the security and integrity of exchanges between heterogeneous and potentially constrained equipment in a critical cyberphysical system" while (1) respecting performance constraints; (2) being implantable on all hardware resources without specific security resources and (3) being compatible with the main standardized protocols such as DDS or MQTT and implementing security standards [ISO/IEC 27005]. For the proof of concept of the research results, case studies and scenarios will be developed for the validation of the approaches and the proposed architecture. They will be based on the use of existing communication standards and can exploit open datasets or the implementation of synthetic data.
Niveau de formation : Bac+5
Temps partiel / Temps plein : Plein temps
Type de contrat : Contrat à durée déterminée (CDD)
Compétences : Sécurité Informatique, Technologie Digitale, Gestion des Clés, Message Queuing Telemetry Transport (MQTT), Infrastructure à Clé Publique, Traitement des Données, Multi-Agent Systems, Technologies Informatiques, Anglais, Prise de Décision, Sens de la Communication, Persévérance, Compétences Interpersonnelles, Esprit d'Équipe, Architecture, Système Autonome, Respect de la Confidentialité, Systèmes Embarqués, Scalabilité, Systèmes Intelligents, Réalisation de Prototypes, Gestion du Niveau de Sécurité, Simulations, Vulnérabilité, Compétences de Modélisation, Études de Cas, ISO / CEI, Capteurs, Actionneurs, Applications des Règles et Consignes de Sécurité, Preuve du Concept
Courriel :
job-ref-nyeb2jvxr8@emploi.beetween.com
nyeb2jvxr8@emploi.beetween.com
Téléphone :
0476574845
Type d'annonceur : Employeur direct