This year, between May 23-27, the IEEE International Conference on Robotics and Automation (ICRA 2022) took place in Philadelphia, USA. The hybrid conference has received 3,344 submissions from 92 countries around the world while welcoming more than 8,000 participants. Our SoftGrip team is happy to announce that our two submissions were accepted and presented at the conference.
Our first accepted work to ICRA 2022 is a research paper entitled "Reproduction of Human Demonstrations with a Soft-Robotic Arm based on a Library of Learned Probabilistic Movement Primitives". It has been co-authored by Paris Oikonomou, Athanasios Dometios, Mehdi Khamassi and Costas Tzafestas of the Institute of Communications and Computer Systems (ICCS). This work proposes an architecture that aims to control a two-module bio-inspired soft-robotic arm, exploiting the enhanced parameterizability of Probabilistic Movement Primitives, and assuming that any trajectory might be defined as the proper composition of individual primitives obtained by a learned Movement Primitive library. The in-person oral presentation took place on the 3rd day of the conference (May 25) in "Soft Robot Applications" session, along with the corresponding poster session where the first author (Paris Oikonomou) answered questions and had fruitful discussions with the other attendees, related to this work.
Secondly, Thanasis Mastrogeorgiou and Panagiotis Mavridis from TWI Hellas and Panos Chatzakos from the University of Essex prepared a research poster and an accompanying short video presentation. The online oral presentation by the TWI team happened on the last day of the conference, May 27, as part of the ICRA 2022 Workshop on Agricultural Robotics and Automation. The poster focuses on Mushroom Detection and Pose Estimation Using Instance Segmentation and Model Fitting. More specifically, the research team aims to develop a highly accurate and low-cost vision system to estimate a mushroom’s pose. To achieve that, training and testing an algorithm as to render it capable of distinguishing different mushroom forms and precisely estimating orientations is essential. Ultimately, the final objective is to provide our soft grippers with the necessary information to handle mushrooms without blemishing or destroying them throughout the harvesting process.
For more information on…
- The ICCS research paper
- The TWI Hellas/EssexAi vision system work: watch the video presentation or check out the research poster.
