Adapting robotic manipulation to new objects and tasks often requires additional demonstrations or manual engineering. Reusable manipulation skills can reduce this effort, but adapting these skills to new scenes remains challenging. We present ManiSkillFormer, a framework for demonstration-free and compositional manipu...
Pei-Qi Yu, Mosam Dabhi, Shan Li et al.· 1 citation
ARSTAG is presented, an agentic Real2Sim2Real system that turns a single RGB image and a natural-language instruction directly into robot policy-learning data and shows that task-consistent randomization substantially improves robustness, and policy performance increases with generated dataset size.
This work introduces Model-Based Diffusion via Constraint Optimization and Adaptive Scheduling (MD-COAS) for SRMP that unifies the inexact Augmented Lagrangian Method (iALM) soft diffusion prior with a Convex Feasible Set (CFS)-based hard projection operator, and adaptively schedules and co-optimizes safety enforcement...