Robo-Insight #1 – Robohub

Source: DALL E 2 OpenAI quickly “a hyperrealistic image of a robot reading the news on a laptop in a coffee shop”

Welcome to the premiere issue of Robo-Insight, the bi-weekly robotics news update! In this post, we’re excited to present some of the great advances in this field, highlighting robotics advances in terrain traversability, shape morphing, object avoidance, mechanical memory, physics-based AI techniques, and new home robotics kits. These developments demonstrate the continuous evolution and potential of robotics technology.

The quadrupedal robot traverses tricky terrain thanks to enhanced 3D vision

Recently, researchers from the University of California San Diego have given the quadrupedal robots a forward-facing depth camera to enable them to clearly analyze the environment around and beneath them. The researchers used a model that derives 3D information from short 2D frame videos. This data can also be compared with previous images to estimate possible 3D transformations. In addition, their system is also self-checking, as it combines information to provide a kind of short-term memory. While the model doesn’t guide the robot to a specific location, it does allow the robot to traverse challenging terrain. Full paper, more videos, and code (coming soon) can be found Here.

Neural Volumetric Memory for Visual Driving Control

Mori3: A polygon shapeshifting robot for space travel

In line with appearing in difficult environments, Mori3, a robot that can change shape and interact with objects and people, was created by researcher at the EPFL School of Engineering. The Mori3 modular robot can transform from 2D triangles to various 3D shapes by combining digital polygons that blend with flock behavior. This study helps highlight how modular robotics can be used for tasks such as space exploration. Robots show a great deal of versatility thanks to their adaptability and ability to assemble and disassemble. The Mori3 robot will be used by the crew to communicate with the spacecraft and perform exterior repairs.

Mori3, the shape-shifting and modular origami robot

A step towards a safe and reliable autopilot for flying

And speaking outside the field, a machine learning method recently devised by MIT researcher to overcome challenging stabilization-avoidance problems in autonomous aircraft. This method offers a tenfold increase in stability and outperforms the previous technique in terms of security. Researchers can achieve a steady trajectory while avoiding hurdles by redefining the problem as constrained optimization and using deep reinforcement learning techniques. This method avoids crashing into a simulated jet when flying in a narrow space. This method can be used to create dynamic robot controllers and maintain stability and security in mission critical systems. Correction of accounting uncertainty and hardware testing will be carried out in the future.

This video shows how the researchers used their technique to fly a simulated jetliner effectively in a scenario where it had to stabilize to a target near the ground while maintaining a very low altitude and remaining within a narrow flight corridor. Thank you researchers.

Metamaterials with inherent frustration have a mechanical memory

Breakthroughs in the development of materials with mechanical memory have been achieved by researchers from the University of Amsterdam and ENS de Lyon. They created materials that could remember how they were previously bent or stretched and which had special sections or lines that would not deform when pushed or pulled. These developments in metamaterials can be used in mechanical and quantum computers, as well as in robotics and photonics. To create this mechanical memory effect, the researchers used the idea of ​​a non-orientable sequence, which is present in items such as the Möbius strip.

Metamaterials with inherent frustration have a mechanical memory

Hybrid AI powered computer vision combines physics and Big Data

On the topic of improving computer vision technology, researchers from UCLA and the United States Army Research Laboratory has developed a hybrid strategy that integrates physics-based awareness into data-driven algorithms. This article presents several approaches to integrating physics and data in AI such as physics-based AI data sets, network design, and network loss functions. Hybrid techniques have shown promising results in image enhancement, motion prediction, and object tracking. AI systems based on deep learning may eventually be able to independently master the rules of physics, according to the researchers.

Achuta Kadambi/UCLA
The graphic shows two techniques for incorporating physics into the machine learning pipeline — residual physics (above) and physical fusion (below). Source.

myCobot 320 AI Kit 2023

On the industry front, myCobot 320 AI Kit 2023, a breakthrough robot arm built for user-programmable development, has just been released by Elephant Robotics. It offers flexibility for business, research and creative endeavors due to its increased working radius, higher payload capacity and smart gripping ability. The kit features considerable improvements over previous designs, supports five advanced sight recognition algorithms, including grippers, and comes with easy-to-use visualization software.

Bowl bot

Finally, Bowl Bot is a self-cleaning robot recently created by Nala Robotics which can prepare a variety of individual food bowls. It offers a wide variety of 28 ingredients for bases, proteins, garnishes, and sauces in a small footprint. Bowl Bot, which is equipped with the latest AI and vision technologies, works at high speed while maintaining cleanliness and eliminating cross-contamination with its self-cleaning system.

Source: Nala Robotics

This extraordinary breakthrough is just a glimpse into the vibrant and dynamic world of robotics. This field continues to inspire and push boundaries, propelling us towards a future in which robotics technology plays an increasingly important role. Look forward to more exciting updates in the next issue!


  1. “A Quadrupedal Robot Crosses Complex Terrain Thanks to Enhanced 3D Vision.” Accessed July 1, 2023.
  2. Christoph H. Belke, Kevin Holdcroft, Alexander Sigrist, Jamie Paik. Morphological flexibility in robotic systems through the splicing of physical polygons. Natural Machine Intelligence, 2023; DOI: 10.1038/s42256-023-00676-8
  3. “A Step Towards a Safe and Reliable Autopilot to Fly.” MIT News | Massachusetts Institute of TechnologyRetrieved 12 June 2023
  4. Amsterdam, University. “Metamaterials with Innate Frustration Have a Mechanical Memory.” University of Amsterdam. Accessed July 1, 2023.
  5. Hybrid AI-Powered Computer Vision Combines Physics and Big Data. Accessed July 1, 2023.
  6. Empowering Research and Development: Introducing the MyCobot 320 AI Kit 2023 by Elephant Robotics. Accessed July 1, 2023.
  7. “The Bowl Fully Automated Robot Salad Bowl Maker – Nala Robotics.”. Accessed July 1, 2023.

Shaunak Kapur is part of the Robohub volunteer team, and soon to be a senior at High School (Texas). Shaun has been fascinated by robotics from a young age.

Shaunak Kapur is part of the Robohub volunteer team, and soon to be a senior at High School (Texas). Shaun has been fascinated by robotics from a young age.

Source link

Related Articles

Back to top button