Seminari di Automatica e Robotica - From constraints to opportunities: efficient visual task learning for humanoid robotics

Dettagli

Seminari di Automatica e Robotica: E. Maiettini- From constraints to opportunities: efficient visual task learning for humanoid robotics - 29.5.2022

Il prossimo incontro del ciclo di "Seminari di Automatica e Robotica", organizzato dal gruppo di Automatica del Dipartimento di Ingegneria, Università degli Studi di Perugia, laboratorio ISARLab (https://isar.unipg.it/) si terrà venerdì 29 aprile, ore 11.30, nell'aula 3 del Dipartimento di Ingegneria.

Relatore: Elisa Maiettini, Humanoid Sensing and Perception (HSP) research line at the Istituto Italiano di Tecnologia
Titolo: From constraints to opportunities: efficient visual task learning for humanoid robotics.

Il seminario potrà essere seguito anche online, a questo indirizzo: https://bit.ly/3K8xwyb.

Abstract
Reliable visual perception and efficient adaptation to novel conditions are priority skills for humanoids that function in ever changing environments The vast advancements in latest computer vision research, brought by deep learning methods, are appealing in a robotic setting However, their adoption in applied domains is not straightforward since adapting them to new tasks is strongly demanding in terms of annotated data, optimization time and computational resources These requirements do not generally meet current robotics constraints In past and current work, we proposed techniques to overcome some of these issues by exploiting opportunities coming from humanoid robotic platforms (like e g their embodiment in the environment and their sensors) During this presentation, I will provide an overview of our latest work on this topic

LOCANDINA