Quadrupedal Parkour
Written by Lucy Black   
Sunday, 31 March 2024

What is it with robots and parkour? First Atlas and now ANYmal want to impress us with their prowess. For the roboticist, however, emulating the skills of free running can enhance the capabilities of autonomous robots.

There are two teams of researchers working with ANYmal at ETH Zurich and they take different approaches to giving the robot the skills needed for a search and rescue role such as negotiating perilous unfamiliar terrain. Both teams rely on reinforcement learning and we have already reported on the model-based control approach adopted by doctoral student Fabian Jenelten. Now we are looking at an alternative strategy which uses trial and error and comes from Nikita Rudin, also a doctoral student, who does parkour in his free time.

The video shows ANYmal's fluency as it negotiates an obstacle course:


Rudin explains that before the current project, which is detailed in  started, several of his colleagues thought that legged robots had already reached the limits of their development potential but, he says:

I had a different opinion. In fact, I was sure that a lot more could be done with the mechanics of legged robots.”

Having set out to further push the boundaries of what ANYmal can do, based on his own experience of parkour, Rudin has seen the quadrupedal robot learn new skills. ANYmal can now scale obstacles and perform dynamic maneuvers to jump back down from them. When presented with an obstacle, ANYmal uses its camera and artificial neural network to determine what kind of impediment it’s dealing with. It then performs movements that seem likely to succeed based on its previous training.

In the report in Science, Rudin and his co-authors conclude that their approach has enabled ANYmal to navigate complex scenes where it has to overcome large obstacles while selecting a nontrivial path toward its target location, commenting:

"by aiming to match the agility of free runners, we can better understand the limitations of each component in the pipeline from perception to actuation, circumvent those limits, and generally increase the capabilities of our robots."

 anymal parkour

More Information

ANYmal parkour: Learning agile navigation for quadrupedal robots, Science Robotics, Vol 9, Issue 88 

Related Articles

ANYmal Evolution and Locomotion

Atlas Does Parkour And The Team Explains Why And How

A Tale of Three (Robot) Dogs

Spot The Robot Dog Learns New Tricks

Should We Beware The Unitree Go2

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.



Pg_lakehouse Makes PostgreSQL Quack

Pg_Lakehouse from ParadeDB is an extension that turns PostgreSQL into the analytical engine of DuckDB. Why is that useful? How do you use it?

Researchers Use AI To Decode Dog Language

Scientists from the University of Michigan have used AI to decode what dogs mean by different types of bark. Wav2Vec2 succeeded at four classification tasks - dog recognition, breed identification, ge [ ... ]

More News

kotlin book



or email your comment to: comments@i-programmer.info

Last Updated ( Sunday, 31 March 2024 )