1
Return

A hierarchical approach to imitation learning for manipulation tasks requiring time varying forces

delete2026-04-12
delete0
PRE
AI
R
Rishabh Shukla *
A
Adithya Santhosh
S
Shaili Gandhi
S
Samrudh Moode
S
Satyandra K. Gupta
DOI：10.1016/j.rcim.2026.103309delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
• Imitation learning based on Latent Diffusion Models (LDM). • Diffusion policy selects discrete skills for contact-rich manipulation. • A low-level force-reactive trajectory generator executes high-frequency tasks, e.g., chiseling and peeling. • Achieves 100% in-distribution task success while reducing peak interaction forces compared to baselines. • Zero-shot generalization to unseen geometries.
Keywords:
Imitation learning
Latent Diffusion Models
Contact-rich manipulation
Force-reactive control
Zero-shot generalization

Journal

R
Robotics and Computer-Integrated Manufacturing
IF:
11.4
Papers:
3.3K
Citations:
1.3W

Organization

No organization information available
Cited Papers

Cited Papers

Citing Papers

Citing Papers