1 paper · 1 filter
Yu Deng, Yufeng Jin, Xiaogang Jia +3
Robot manipulation often fails in the final millimeters: a policy may recognize the right object yet miss the pose offsets, boundaries, or pre-contact alignments needed for action.…