TacCoRL: Integrating Tactile Feedback into VLA via Simulation
Adds touch to a VLA via sim-real co-training and simulated RL in a real-aligned simulator, learning how contact should change actions near failure. Reports 72.5% vs 50.0% on four bimanual tasks.通过仿真-真机协同训练和在对齐真实的模拟器中做强化学习，为 VLA 加入触觉，学习接近失败时触觉应如何调整动作。报告四个双臂任务成功率 72.5%，基线 50.0%。
BibTeX
@misc{ma2026taccorl,
title = {{TacCoRL: Integrating Tactile Feedback into VLA via Simulation}},
author = {Siyu Ma and Yuqi Liang and Chang Yu and Yunuo Chen and Hao Su and Yixin Zhu and Yin Yang and Chenfanfu Jiang},
year = {2026},
eprint = {2606.11743},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2606.11743},
}