Communications of the ACM
Benchmarking Robots with Dog-Inspired Barkour
A lack of standardized tasks for these types of robots has made it difficult to compare their progress, said Google DeepMind Atil Iscen.
Credit: Google
Scientists from Google DeepMind evaluated the agility of quadripedal robots with a dog-agility-inspired course called Barkour.
DeepMind's Atil Iscen said Barkour features assorted obstacles requiring "a combination of different types of behaviors such as precise walking, climbing, and jumping," while a timing-based metric rewards faster behavior.
Barkour's obstacles include a pause table, weave poles, climbing an A-frame, and a jump to test “multiple axes of agility, including speed, acceleration, and balance," according to Iscen.
The researchers used a specialist training strategy that focused on each type of skill required for the course and a generalized framework that analyzed models run using the specialist program.
From IEEE Spectrum
View Full Article
Abstracts Copyright © 2023 SmithBucklin, Washington, D.C., USA
No entries found