Abstract
Autonomous underwater vehicles (AUVs) and remotely operated vehicles (ROVs) are transforming marine research, defense, and industrial applications. However, significant challenges persist in enabling robust perception, energy-efficient operation, and adaptive autonomy in dynamic underwater environments. This talk explores recent advances in vision-based navigation, multi-modal sensor fusion, and AI-driven decision-making for subsea autonomy. The first part of the talk presents our recent work on simulation-driven learning, leveraging synthetic data for robust underwater perception and localization. The second part highlights advancements in integrating optical, acoustic, and magnetoelectric sensing for stealth navigation and long-distance communication in low-visibility conditions. Two key aspects will be discussed in several applications: (i) multimodal sensing (vision and sonar), and (ii) perception-enabled ROVs for real-time environmental monitoring.
For more info, please follow this link.