Event
G1 Dribble End-to-End Reinforcement Learning Dribbling Control Real-World Test
1 reports · 1 sources
Guyuehome introduced a method for achieving humanoid robot soccer dribbling using end-to-end reinforcement learning, called G1 Dribble: the walking task does not adjust weights, and the velocity command simultaneously drives the body and the soccer ball. The author states that they conducted real-world tests locally on MuJoCo-Warp and recorded eighteen cloud-based sessions, providing the ball-touch rhythm, the re-touchable distance band, and failure modes, including no touch when crossing the ball from a centered start and a first touch only when offset to the side, as well as the ball being knocked away directly at 1.5 meters per second. These are local real-world test records, not verified general effects.
Report timeline
- G1 Dribble: End-to-End Reinforcement Learning for Humanoid Robot Soccer Dribbling Control
G1 Dribble uses end-to-end reinforcement learning to achieve humanoid robot soccer dribbling; the walking task's weights are not adjusted, and velocity commands drive both the body and the ball. The author tested it locally in MuJoCo-Warp and recorded eighteen cloud sessions, reporting touch rhythm, the distance band for re-touching the ball, and failure modes, including that a centered start steps over the ball without touching it, a first touch occurs only when offset to the side, and at 1.5 meters per second it knocks the ball away directly.
Guyuehome ·
Related topics
Interest trend
- Current
- 3
- Peak
- 4Nov 4 01:00
- Past 24 hours
- –
The trend compares participants observed throughout the same period, so its range may be smaller than the current interest count. Move or tap to inspect each hour; use the arrow keys on a keyboard.