Mmastodon TechnologyRobotics first seen 1 d ago, last 1 d ago, peak #4

Bipedal robot learns to score goals through trial and error

Original: A longstanding goal for robot reinforcement learning is that robots learn from their own physical experience, as humans

Researchers report progress on a longstanding goal in robot reinforcement learning: machines improving through their own physical experience, the way humans and animals do. The team studied the problem using a contact-rich task, training a bipedal robot to score soccer goals through direct trial and error rather than explicit programming or simulation alone.

Why now: New arXiv research shows robots learning motor skills from real physical experience, a milestone robotics watchers follow closely.

arXivbipedal robot

Open on mastodon →

Evidence

• A longstanding goal for robot reinforcement learning is that robots learn from their own physical experience, as humans and animals do. We study this challenge using the contact-rich task of a bipedal robot learning to score soccer goals. Source: arXiv cs.RO https://… · sipirtu

API: https://socialmediatrends-api.osmike.com/v1/trends/2219689