We study algorithms that allow robots to perceive and explore the environment, by manipulating objects and interacting with humans. Motivated by research on human perception, our strategy seeks to explore active perception and multi-modal integration, whereby the robot actively explores the environment to improve perception and learning, leveraging on various sensory modalities. We perform a mix of basic and applied research in domains spanning rehabilitation, human-robot collaboration, service and industrial robotics. We also seek to extend robot autonomy with the development of software tools and methodologies for modelling and deploying robot behaviors.
Humanoid Sensing and Perception
IIT People List
Marta Caracalli Administrative Assistant Humanoid Sensing and Perception iCub Tech Event-Driven Perception for Robotics Artificial and Mechanical Intelligence
IIT Publications List
Ceola F., Rosasco L., Natale L.
RESPRECT: Speeding-up Multi-fingered Grasping with Residual Reinforcement Learning
IEEE Robotics and Automation Letters
Gabriele M.C., Maracani A., Alfano P.D., Piga N.A., Rosasco L., Natale L.
Sim2Real Bilevel Adaptation for Object Surface Classification using Vision-Based Tactile Sensors
IEEE-RAS International Conference on Robotics and Automation