A potential field grid representation enables RL-based robotic throwing policies to generalize across arbitrary obstacle configurations, achieving up to 90% success on a real robot.
A solution to adaptive mobile manipulator throwing,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Learning to Throw Objects Safely in Multi-Obstacle Environments
A potential field grid representation enables RL-based robotic throwing policies to generalize across arbitrary obstacle configurations, achieving up to 90% success on a real robot.