Ph.D XXXVIII

Supervisor: Fabio Remondino

Room: FBK

Mail: razzino.armandovittorio@spes.uniud.it

Research Project

Multi-modal learning-based Simultaneous Localization and Mapping (SLAM)

Simultaneous Localization and Mapping (SLAM) is concerned with creating and updating the map of an environment while a moving robot is exploring it. State of the art SLAM models and frameworks generally focus on a single technique, e.g. LIDAR (Laser Imaging Detection and Ranging), Inertial, Visual or learning-based, or merge multiple techniques at the data analysis stage, severely reducing adaptability to environments and datasets not especially designed for a specific technology or to data acquisition using multiple kinds of onboard sensors. The project aims to achieve methods fusion during data acquisition or environment exploration and exploit Machine and Deep Learning to increase accuracy of SLAM, e.g. in trajectory estimation and environment reconstruction, especially with self-made robots and datasets.