Soccer skills with onboard vision, unlike many recent learning-based systems that rely on motion capture or external cameras to track the ball. Next year's RoboCup is going to be on a whole different level.
Introducing Joga, a full-stack humanoid soccer system with active vision. Agile dribbling. Tight direction changes. Receive → dribble → shoot/pass. An actuated neck enables ball tracking with onboard vision. Residual models for perception + actuation close the sim-to-real gap. Paper coming soon! Work done in collaboration with @Avi_Narula14 @khai_ngx @venki1899 @JohnMarangola @martinpeticco @pulkitology

Sep 29, 2026 · 5:01 AM UTC

5
8
69
9,318
Sort replies: Relevant Recent Liked
Replying to @TheHumanoidHub
onboard vision chasing a ball without mocap babysitters is the honest flex. curious how often the residual still dies when the pitch lights go weird…
18
Replying to @TheHumanoidHub
Football, my Padawan, this⬆️ is not.⚽️
1
22
Replying to @TheHumanoidHub
An actuated neck just to keep eyes on the ball is a smart fix. Curious how onboard vision holds up on a wet, muddy pitch. Let's connect, I post robotics clips too.
3
Replying to @TheHumanoidHub
Talk about a game-changer! Can't wait to see Joga's skills put to the test next year 🚀
5