Blog
Robótica & RL
TRACT: Temporally Routed Action Chunks with Chronological Phase Authority for Contact-Rich Manipulation
arXiv:2607.29285v1 Announce Type: new Abstract: Action chunking shortens the effective decision horizon of robot imitation learning by predicting multiple future actions, while conventional phase conditioning describes the current control instant. When a predicted horizon crosses a procedural boundary, assigning the current phase to the entire chunk creates a structural temporal mismatch. We present TRACT, which factorizes phase-structured action chunking into an accepted current phase and a sin...
arXiv cs.RO
·Jiahao Liu, Kento Kawaharazuka, Tasuku Makabe, Kei Okada
·
// relacionados