Multi-Resolution Tactile Imitation Learning for Contact-Rich Robotic Manipulation
Fuses an RGB camera, a GelSight Mini and a high-rate event-based tactile sensor for a flow-matching policy. Reports 80% average success on five contact-rich tasks vs 35.5% vision-only.融合 RGB 相机、GelSight Mini 与高频事件型触觉传感器的多分辨率表示，驱动流匹配策略。报告五个富接触任务平均成功率 80%，纯视觉仅 35.5%。
BibTeX
@misc{krohn2026multi,
title = {{Multi-Resolution Tactile Imitation Learning for Contact-Rich Robotic Manipulation}},
author = {Rickmer Krohn and Erik Helmut and Niklas Funk and Jan Peters and Vignesh Prasad and Georgia Chalvatzaki},
year = {2026},
eprint = {2606.06281},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2606.06281},
}