cs.RONov 15, 2026

Prompt Re-Specification and Target-Mask Availability in Mask-Conditioned Visual Reinforcement Learning for Moving-Platform Landing

Authors: Ho Tim Tsui, Jongyun Kim

Abstract

Visual reinforcement learning (RL) policies that land a drone on a moving platform can fail when the landing pad looks different from the one used in training. Using binary target masks as the visual input removes most pixel-level appearance information while preserving the spatial representation of the target, and lets a text prompt specify the target. In simulation, we test whether a frozen (fixed-weight) mask-based DrQ-v2 policy can handle a new pad when only the prompt of its fine-tuned SAM 3 detector changes. A banana-themed pad lowered landing success from 75.6% to 12.0%, and SAM 3 rarely detected it under the original prompt. A revised prompt, chosen after testing several candidates, restored success to 68.4% with no weight updates. In a separate experiment with a policy trained on ground-truth masks, filling in masks that were missing while the pad was visible helped landing far more than the tested mask-centroid and area corrections. Tracking the pad over time supplied masks more often and also kept producing them after the pad was confirmed absent from the rendered image, but it did not establish a reliable improvement in landing success. Whether a usable mask reaches the policy at all thus appears to be an important limiting factor for this approach.

Figures & tables

Appendix figures & tables6 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Robust Bimanual Vision-Language-Action Models via Embarrassingly Simple Modality Masking

    Aug 23, 2026Dongzhou Cheng, Ziang Li, Yixiao Zhou +6Vision-Language-Action ModelsRobustness of VLA Models

  2. RoboMP-DINOv2: Prompts, Not Filters for Robust Robot Manipulation

    Sep 22, 2026Han Qi, Heng YangGeneralization in Robotic ManipulationVisuomotor Policy Learning

  3. MaskWAM: Unifying Mask Prompting and Prediction for World-Action Models

    Jun 11, 2026Hanyang Yu, Haitao Lin, Jingbo Zhang +4World Model LearningLanguage-Conditioned Robot Manipulation