Juliette Grosset Thesis defense

The challenges of Industry 4.0 include optimizing data flows and decentralizing decision-making, where centralised systems often become inadequate. Autonomous Industrial Vehicles (AIVs) need to become smarter and more cooperative by exchanging relevant traffic data among themselves and with the infrastructure. This thesis aims to enhance the autonomy of AIVs through collective intelligence strategies, improving their adaptability, decision-making, and efficiency by facilitating communication and information sharing. Our methodology, based on modeling, simulation, and scenario testing, seeks to propose specific collective strategies to strengthen various key functionalities of AIVs. First, we improved an obstacle avoidance algorithm and developed a global strategy based on shared perception. We adapted and proposed standardized messages for the Industry 4.0 context and introduced a dynamic task (re)allocation system in decentralized environments. Based on the strong assumption of cooperative perception, we then proposed an architecture for generating V2X data. Fi-
nally, we developed collective energy management strategies for each AIV using a fuzzy decision model, allowing them to autonomously determine the optimal recharge times and thus reduce their downtime within the fleet.