Supervision : Cédric HERPSON

Architecture d’agent embarquée pour l’autonomie ajustable

The objective of this thesis is - building on a preliminary work carried out within the framework of 2 research internships [1,4] - to propose, design and implement a decision-making architecture able to adapt the degree of autonomy of an artificial entity over time. Outraged the development of an architecture supporting the different modes mentioned, special attention should be given to the form that the transition of a mode takes (autonomous, semi-autonomous, remotely operated) to the other. Indeed, the level of trust that the user places in the system has a strong influence on transition conditions from one mode to another. However, as this confidence evolves over time, it is necessary to study the impact of this dynamic in the way the system represents and manages the "viability zones" of each mode of operation.