cs.ROOct 30, 2026

TRACE: Target-State Risk-Aligned Conditional Execution for Embodied Tracking

Authors: Mingyang Lyu, Gongce Cao, Yiyang Jia, Lu Jia, Erliang Lin, Yuhao Zhang, Yinqian Sun, Pin Lyu, 

Organizations: Institute of Automation, Chinese Academy of Sciences · University of the Chinese Academy of Sciences · Zhongguancun Academy · University of Science and Technology Beijing · 

Abstract

Embodied visual tracking requires a robot to follow a language-specified person under occlusion, look-alike distractors, and ambiguous instructions. We show that improving tracking persistence can simultaneously increase safety violations. Using a compact 0.6B vision-language-action (VLA) tracker with a frozen backbone, we update only the planner head with proximal policy optimization (PPO). PPO improves success and following rates across all three EVT-Bench settings, but increases the collision rate on distracted tracking because episodes that previously ended in target loss can instead end in contact. We introduce TRACE, an execution-time target-state risk gate that preserves the learned policy and intervenes selectively. A lightweight risk head processes 84 cues, including candidate boxes, identity margins, a learned range-and-bearing probe, and the policy's commands. When the calibrated risk exceeds a threshold, TRACE rescales the translational component of the next waypoint while leaving yaw unchanged. The executed action is identical to PPO's on more than 98% of steps. In paired evaluations on identical episodes, TRACE reduces the distracted-tracking collision rate by 3.96 percentage points, while further improving success and following rates. Matched slowdown controls without target-state timing do not reproduce this improvement. Concurrent evaluation reproduces a collision reduction of 3.77 points, compared with 0.29 points for identical-arm null pairs, indicating that the effect exceeds evaluation variability. These results show that target-state-aware execution can recover the safety cost of policy post-training with minimal intervention.

Figures & tables

Appendix figures & tables26 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. VLX-Go: Robust Embodied Tracking through Multi-Stage Learning and On-Policy Refinement

    Oct 14, 2026Kyusong Lee, Heting Ying, Tiancheng ZhaoVision-Based Robot ControlMultimodal Visual Object Tracking

  2. ReferTrack: Referring Then Tracking for Embodied Visual Tracking

    Jul 22, 2026Hanjing Ye, Tianle Zeng, Jiazhao Zhang +6Visual Object TrackingVisual Grounding

  3. DobotTrack: Target-Grounded Vision-Language-Action Learning for Embodied Person Tracking

    Nov 6, 2026Shengjun Deng, Yinghan Zhang, Chengrui Wei +8Multimodal Visual Object TrackingSpatial Grounding for VLA Models