New work from the lab on robust sim2real transfer of multifinger grasps using the strategy of grasp evaluation being presented at #Corl2024!
We're releasing a dataset of 3.5M labeled grasps with corresponding perceptual data - see the 🧵 from @albert_h_li below for details!
There have been many recent big grasping datasets, but few demos of real-world grasping using generative models. How do we achieve this?
Introducing: Get a Grip (#corl2024)!
We show that instead of generative models, discriminative models can attain sim2real transfer!
👀🧵👇
It was wonderful to work with Preston Culbertson (@pdculbert), Vince Kurtz, and my advisor, Aaron Ames🤝
🖥️Project website: https://t.co/2XGNqyr0TX
📰Paper: https://t.co/zTD507y0hs
🧑💻Code: https://t.co/NjmoSY3sbn
🛠️Open-source hardware setup coming soon!
Awesome new work from our lab member @albert_h_li! Sampling-based control strategies like CEM alone are viable for contact-rich tasks like in-hand cube reorientation in the real world (no reinforcement learning 🤯).
For more details, check the thread below👇
Excited to share our new📰, DROP: Dexterous Reorientation via Online Planning!
Overview:
🔹We tackle cube rotation🧊♻️on hardware
🔹DROP is the first 🧊♻️sampling-based MPC demo. No reinforcement learning!
🔹Median 30.5 rotations w/o dropping, max of 81👑🦾
See 🧵below👇
I'm so excited to announce I will be joining @Cornell_CS as an assistant professor next fall!
I can't wait to launch my lab and start building robots with human-level dexterity and adaptability. 🤖🏃🤹
And AMBER lab post-doc Jeeseop took home the “Outstanding Paper” award for work that he did during his PhD (with Randy, Leila, and Kaveh) on multi-agent collaboration. Congrats to all!!
AMBER lab graduate students Wyatt and Noel took home the “Outstanding Student Paper” and “Outstanding Dynamics and Control Paper” awards, respectively, for their work on planning through motion primitive graphs and MPC for 3D hopping
I am so happy to share that our paper entitled ``Distributed data-driven predictive control for multi-agent collaborative legged locomotion" by my Ph.D. students R. Fawcett, L. Amanzadeh, and J. Kim, in collaboration with Dr. A. Ames (@AMBER_lab), has been selected as the outstanding paper of the 2023 IEEE International Conference on Robotics and Automation (#ICRA2023).
https://t.co/LBCcJGfN7Q
Excited to share our review paper that just came out on control methods for lower-limb powered prostheses!
@maegan_tucker1@AMBER_lab@GT_EPIC_Lab
https://t.co/OQHhEossUD
So happy to share our survey paper on state-of-the-art control techniques for lower-limb powered prostheses!!
https://t.co/ohR0XYXahg
@RachelGehlhar @AMBER_lab@GT_EPIC_Lab
Looking forward to giving a seminar at the University of Minnesota today! I'm excited to join the Mechanical Engineering faculty here in Spring 2024. @UMNME@UMNCSE https://t.co/sMoAYwcIkj
Today I officially signed an offer to become an Assistant Professor at the Georgia Institute of Technology starting January 2024!!! I will be joining both the @GeorgiaTech_ECE and @MEGeorgiaTech departments. @gatechengineers@GeorgiaTech
We had such a blast hosting the FTC Team 25 Rock N’ Roll Robots — a local all-girls robotics team! Can’t wait to see what you all accomplish this year!🤖💪🏼💁🏻♀️
@maegan_tucker1 @RachelGehlhar @noel_c_s
@inseungkang This “active mode” allows patients to contribute to the walking motion (also known as partial or variable assistance), so the assumption is that the patients would have either incomplete motor paraplegia or would be stroke patients. But we’re still exploring 🧐