ScoutVLA: UAV-Centric Active Perception via a Dual-Expert VLA Model for Open-World Embodied Question Answering

2026-06-09

AI Summary

This work addresses the limitations of existing outdoor embodied question-answering systems, which typically terminate upon target detection and struggle with evidence-seeking tasks requiring fine-grained viewpoint adjustments. Inspired by honeybee “waggle dances,” the authors propose ScoutVLA—a decoupled dual-expert vision-language-action model that identifies missing evidence through semantic intent recognition and actively perceives the environment by generating high-degree-of-freedom, continuous viewpoint optimization trajectories. Key contributions include the introduction of FG-EQA, a fine-grained active perception benchmark; a knowledge-isolation mechanism enabling synergistic optimization of semantic reasoning and continuous control; and a flow-matching trajectory generation approach integrated with multimodal fusion. Experiments demonstrate that ScoutVLA substantially outperforms current methods in both simulation and real-world settings, achieving an average strict success rate 10.48 times higher and a question-answering accuracy 7.72 times greater.

Abstract

Aerial Embodied Question Answering (EQA) requires Unmanned Aerial Vehicles (UAVs) to actively perceive the environment and answer natural language questions. Existing outdoor EQA systems usually stop once the target enters the UAV's field of view, leaving the fine-grained viewpoint adjustment needed for evidence-seeking questions largely unresolved. To address this issue, we introduce FG-EQA, a fine-grained active perception EQA benchmark with more than 40K simulated trajectories and 1K real-world trajectories. Drawing inspiration from the ``waggle dance'' of scout bees, which iteratively adjust their flight paths to verify target information, we propose ScoutVLA, an evidence-driven Vision-Language-Action model for outdoor EQA. To emulate this active exploration behavior, ScoutVLA features a decoupled dual-expert architecture: a vision-language expert infers the semantic intent to identify missing evidence, while an independent action expert employs high-DoF flow matching to generate continuous viewpoint-refinement trajectories. To balance the competing demands of continuous control and semantic reasoning, we devise a decoupled training strategy with a knowledge insulation mechanism that prevents the action gradients from erasing the model's multimodal reasoning ability. Extensive simulated experiments and a qualitative real-world field study both verify the superiority of ScoutVLA over the state-of-the-art baselines, demonstrating a 10.48$\boldsymbol{\times}$ higher average strict success rate and a 7.72$\boldsymbol{\times}$ higher average QA correctness.

Problem

• Embodied Question Answering
• UAV
• Active Perception
• Fine-grained Viewpoint Adjustment
• Evidence-seeking

Innovation

• dual-expert architecture
• active perception
• Vision-Language-Action model
• flow matching
• knowledge insulation