Video Friday: Strandbeest
by Evan Ackerman from IEEE Spectrum on (#6FJ41)
Video Friday is your weekly selection of awesome robotics videos, collected by your friends at IEEE Spectrum robotics. We also post a weekly calendar of upcoming robotics events for the next few months. Please send us your events for inclusion.
Enjoy today's videos!
Let's not concern ourselves with whether this beautiful monstrosity of a Strandbeest is technically a robot or not and instead just enjoy watching it move.
Since the beginning of this summer I have been trying to connect several running units (Ordissen) in succession. Animaris Rex is a herd of beach animals whose specimens hold each other as defense against storms. As individuals they would simply blow over, but as a group the chance of surviving a storm would be greater. It is 18 meters long (5 meters longer than the largest Tyrannosaurus Rex found.)
[ Strandbeest ]
It's Slightly Less Big and Significantly Bluer Hero 6!
[ Paper ]
A low-cost robot does extreme parkour including high jumps on obstacles 2x its height, long jumps across gaps 2x its length, handstand on stairs, and running across tilted ramps. We show how a single neural net policy operating directly from a camera image, trained in simulation with large-scale RL, can overcome imprecise sensing and actuation to output highly precise control behavior end-to-end. We show our robot can perform a high jump on obstacles 2x its height, long jump across gaps 2x its length, do a handstand and run across tilted ramps, and generalize to novel obstacle courses with different physical properties.
[ CMU ]
Human waiters might actually have something to worry about here.
[ LSRL ]
While traditional control methods require multiple sensory feedback for the stable and fast locomotion of quadruped robots, our recent work presents a modular neural control architecture that can encode robot dynamics for stable and robust gait generation without sensory feedback. The architecture, integrating a central pattern generator network, a premotor shaping network, and a motor-memory hypernetwork, enables a quadruped robot to walk at different walking frequencies on different terrains, including grass and uneven stone pavement.
Thanks, Poramate!
Visual control enables quadrotors to adaptively navigate using real-time sensory data, bridging perception with action. Yet, challenges persist, including generalization across scenarios, maintaining reliability, and ensuring real-time responsiveness. This paper introduces a perception framework grounded in foundation models for universal object detection and tracking, moving beyond specific training categories.
[ ARPL ]
As always, performing a live robot demo is no small feat, but KIMLAB members embraced the challenge! MOMO (Mobile Object Manipulation Operator) stole the Demo Expo at IROS 2023 by charming everyone as it handed out candies and swept the floor with a broom. We also unveiled our armor-controlled robotic backpack.
[ KIMLAB ]
X30 quadruped robot, a flagship product designed to meet core industry needs in multiple fields, including inspection of power stations, factories, pipeline corridors, emergency rescue, fire detection, scientific research and more.
[ DeepRobotics ]
Robots operating in close proximity to humans rely heavily on human trust to successfully complete their tasks. But what are the real outcomes when this trust is violated? Self-defense law provides a framework for analyzing tangible failure scenarios that can inform the design of robots and their algorithms.