# MILD Dataset and AprilVINS: End-Effector Localization for UMI-Style Manipulation Teaching

- Source: arXiv cs.RO
- Published: 2026-10-28T04:00:00.000Z
- RoboHot: https://robohot.news/en/items/t6m56lkb6svn2jfe2wh4mso1j
- Original: https://arxiv.org/abs/2610.09857

## Summary

The authors propose the MILD manipulation-interface localization dataset, comprising 86 real sensor sequences collected with Insta360 X5 and Insight9 (covering 15 repeated tabletop tasks) and the MILD-Sim simulation subset in Isaac Sim, along with a robot end-effector reference trajectory for each execution.