Video Friday is a weekly compilation of cool robotics videos collected by our friends. IEEE spectrum Robotics. We also publish a weekly calendar of upcoming robotics events over the coming months. Please send us your events for inclusion.
HRI 2024: March 11-15, 2024, Boulder, Colorado.
Eurobot Open 2024: 8-11 May 2024, La Roche-sur-Yon, France.
ICRA 2024: May 13-17, 2024, Yokohama, Japan
RoboCup 2024: July 17-22, 2024, Eindhoven, Netherlands.
Enjoy today’s video!
We present Human to Humanoid (H2O), a reinforcement learning (RL)-based framework that enables real-time whole-body teleoperation of large humanoid robots with only an RGB camera. We have successfully achieved remote manipulation of dynamic whole-body movements in real-world scenarios, including walking, jumping backwards, kicking, turning, shaking, pushing, boxing, etc. To the best of our knowledge, this is the first demonstration to achieve learning. – Real-time full body humanoid remote control based.
[ CMU ]
Thanks to their ability to carefully select their footing and flexibly adjust their body posture while walking, legged robots have the potential to traverse complex terrain and access confined spaces that traditional platforms cannot reach. However, robust deployment in real-world applications is still an open challenge. In this paper, we present a leg movement control method using reinforcement learning and 3D volumetric representation to enable robust and versatile movement in limited unstructured environments.
[ Takahiro Miki ]
Of course, a speed of 3.3 meters per second is fast for a humanoid, but seeing it spin around while walking downstairs is even more impressive.
[ Unitree ]
To improve the safety of cooperative manipulators, the inertia of the moving parts must be reduced. We introduce a novel approach in the form of a passive 3D wire aligner that acts as a lightweight, low-friction power transmission mechanism to achieve the desired low inertia in manipulator actuation.
[ SAQIEL ]
Thank you, Temma!
Robot Era recently released Humanoid-Gym, an open source reinforcement learning framework for bipedal humanoids. As you can see in the video, the RL algorithm gave a robot named Xiao Xing (XBot) the ability to climb and descend haphazardly stacked boxes with relative stability and ease.
[ Robot Era ]
“Double-handed grasping for impact recognition of high-momentum objects.” Need I say more?
[ SLMC ]
More than 80% of stroke survivors have difficulty walking, which has a significant impact on their daily activities, independence and overall quality of life. Now, new research from the University of Massachusetts Amherst is pushing the boundaries of stroke recovery using a unique robotic hip exoskeleton designed as a training tool to improve walking function. This raises the possibility of new treatments that are more accessible and easier to translate from practice into everyday life compared to current rehabilitation methods.
[ UMass Amherst ]
Thank you, Julia!
The controls here are pretty impressive, but without knowing how much the video has been sped up, it's hard to know just how impressive they are.
[ Somatic ]
DJI drones strive to make the world a better place, and one of the ways we do this is through conservation work. We partnered with Halo Robotics and OFI Orangutan Foundation International to show how these drones can make an impact.
[ DJI ]
The purpose of the test is to demonstrate how to remove and replace a satellite module in a 27U CubeSat format using augmented reality control of the robot. In this use case, the “client” satellite is upgraded and refueled using modular components. The robot then removes the failed computer module and places it in a fixture. Then do the same with the propellant tank. The robot then places it correctly back on the satellite.
[ Extend Robotics ]
This video contains highlights and favorite moments from the CYBATHLON Challenges 2024 held on February 2, demonstrating the different types of assistive technologies that perform training tasks and demonstrating the tenacity and determination of the pilots. The Challenge saw new teams, new tasks and new formats for various CYBATHLON disciplines.
[ Cybathlon ]
It has been a long way before electrically powered robots were realized.
[ ABB ]
Small drones to prepare for catastrophic forest fires ( [40,470 hectares]) is like bringing a flashlight to light up a soccer field. This short video explains the main uses for drones of all sizes, why and when to use them, and why not to use them.
[ CRASAR ]
It probably won't surprise you to learn that there are a lot of robots involved in building Rivian trucks and vans.
[ Kawasaki Robotics ]
DARPA's Learning Introspective Control (LINC) program is developing machine learning methods that show promise in bringing such scenarios closer to reality. LINC aims to fundamentally improve the safety of mechanical systems, especially in ground vehicles, ships, drone swarms and robotics, using a variety of methods that require minimal computing power. The result is an AI-based controller the size of a cell phone.
[ DARPA ]