$TF^2$-Steering: Action Refinement with Tactile Forecasts and Feedback via Velocity-Field Steering
Keeps a vision-based flow-matching policy and learns to steer its velocity field from tactile forecasts and live feedback, so touch can correct actions. Reports gains over baselines in six sim and six real tasks.保留视觉流匹配策略，用触觉预测与实时反馈学习对其速度场的残差引导，让触觉能修正动作。报告在六个仿真和六个真实任务上优于基线。
BibTeX
@misc{chen2026tf,
title = {{$TF^2$-Steering: Action Refinement with Tactile Forecasts and Feedback via Velocity-Field Steering}},
author = {Yiming Chen and Xiaofei Wei and Chun Gu and Wei Xiao and Peidong Liu and Li Zhang},
year = {2026},
eprint = {2611.01483},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2611.01483},
}Tactile & Visuotactile Sensing触觉与视触觉感知Manipulation & Dexterous Hands操作与灵巧手flow-matchingresidual-policytactilerealsim