Atlanta, USA – 19th-23rd May 2025
ICRA 2025 (IEEE International Conference on Robotics and Automation) is the flagship annual event of the IEEE Robotics and Automation Society and took place in Atlanta (USA) from 19th-23rd May this year. In its 41st edition, this event once again assembled globally leading researchers, industry experts, and innovators, showcasing most recent research, technological, and industry developments.
The program featured:
Plenary and keynote presentations
Peer‑reviewed contributed papers
Workshops, tutorials, and forums
A technology expo with corporate partners
Student competitions, arts in robotics, and vibrant social events
This year’s event marked a record number of 7,006 participants, and 2,107 technical presentations delivered in 326 technical sessions and 23-24 parallel tracks with newest research findings highlighted at the ICRA Expo and technological innovations showcased by 141 exhibitors and 14 sponsors.
From MAGICIAN’s side, the event was attended by a delegation of University of Trento’s IDRA research group who participated in several workshops accompanied by paper and oral presentations.
The first workshop “Learning Meets Model-Based Methods for Contact-Rich Manipulation” focused on the challenge of endowing robots with the capability to autonomously execute complex, contact-intense interactions, and thus triggered discussions on the value of of structured frameworks typical of model-based approaches while also integrating the adaptability of learning-based models working with large datasets.
The second workshop “The Future of Intelligent Manufacturing: From innovation to implementation” explored the gap between innvation and industrial implemenation by triggering vibrant discussions among participants from academia and industry. Shopfloor challenges and limitations were addressed among others for industrial robotics & automation and human-machine interaction.
In these workshops, Matteo dalle Vedove (Phd student @University of Trento) presented his research work on motion policy learning. The presentation highlighted a novel DMP-based method for learning motion policies on surfaces lowering the requirement of the manifold to be smooth, but only need a polyhedral mesh. His presentation and related paper on MeshDMP: Motion Planning on Discrete Manifolds using Dynamic Movement Primitives, earned him the “Best Paper Award”,
a major achievement and event highlight.
The paper combines both research work from the EU co-funded projects MAGICIAN and INVERSE, and exemplifies the impressive research work done by Phd students in both projects.
Matteo dalle Vedove is co-supervised by MAGICIAN’s coordinator Daniele Fontanelli, INVERSE’s coordinator Matteo Saveriano and Luigi Palopoli.
