cs.RONov 18, 2026

Homeward: No Place Like the Training View

Authors: Mingyang Lyu, Yinqian Sun, Gongce Cao, Guobin Shen, Yuhao Zhang, Lu Jia, Chenfei Yan, Qian Zhang, 

Organizations: Institute of Automation, Chinese Academy of Sciences, Beijing, China · Zhongguancun Academy, Beijing, China · University of Chinese Academy of Sciences, Beijing, China · ByteDance, China · 

Abstract

Vision-language-action (VLA) policies implicitly assume that third-person cameras remain fixed at their data-collection poses; once a camera is bumped, remounted, or swapped at deployment, task success degrades precipitously even though the underlying task and robot remain unchanged. We present Homeward, a training-free, plug-and-play observation adapter that canonicalizes observations from an uncalibrated deployment camera back into the training view for frozen VLA policies. Our design stems from a closed-loop diagnostic paradox in simulation: while re-rendering perturbed views into the training camera with oracle geometry recovers much of the lost performance, substituting an off-the-shelf feed-forward 3D foundation model can leave the policy worse off than applying no correction at all. This degradation concentrates on the thin, articulated robot arm, which feed-forward depth prediction tears and scatters - yet the arm's exact 3D state is already known from robot kinematics. Homeward exploits this asymmetry: it estimates what is unknown - camera parameters and static scene depth - from the live RGB frame and a single training-view reference, and renders what is known - the moving robot links - directly from the kinematic model. Evaluated across ∼\sim300,000 closed-loop episodes spanning four frozen VLAs, Homeward boosts success across all 1,599 camera variants of LIBERO-Plus from 62.5%, 76.5%, 68.0%, and 3.6% to 77.6%, 95.1%, 91.2%, and 26.4% for OpenVLA-OFT, π0.5π_{0.5}, π0π_0, and OpenVLA, respectively. Without any scene fine-tuning or deployment camera calibration, Homeward extends to 6-DoF random placements, focal-length changes, and, with per-step pose estimation, continuous intra-episode camera drift, while matching or surpassing a calibrated, scene-trained view-synthesis baseline on three of four policies on LIBERO-Plus.

Figures & tables

Appendix figures & tables21 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. From Fixed to Free Cameras: Calibration-Free View-Robust Vision-Language-Action Model

    Jul 6, 2026Wenhao Li, Xueying Jiang, Quanhao Qian +4Language-Conditioned Robot ManipulationHand-Eye Calibration

  2. Cross-View Action Consistency for Camera-Robust Vision-Language-Action Policies

    Aug 7, 2026Bingqi Huang, Bingchuan Wei, Xuan Wang +2Multi-View ConsistencyVision-Based Robot Control

  3. GS-VLA: Plug-and-Play Viewpoint Canonicalization for Frozen VLA Policies via Gaussian Splatting

    Aug 19, 2026Yechan Park, HyunJin KimRobustness of VLA Models