Generative Action-Chunk Sampling for Adaptive Stiffness Control in Physical Human-Robot Collaboration
产业动态
来源：arXiv cs.RO发布时间待核实
arXiv:2608.25284v1 Announce Type: new Abstract: Physical human-robot collaboration requires a robot to provide assistance when human intention is clear while remaining compliant when several future motions are plausible. We present an adaptive stiffness framework based on generative action-chunk sampling. Conditioned on an RGB image and external joint-torque estimates, the policy samples multiple future action chunks from an observation-conditioned prior.