Exp2VLA: Enabling Vision-Language-Action for Drone Navigation from Expert Demonstrations

2026-07-03

AI Summary

Existing vision-language-action (VLA) models for unmanned aerial vehicles (UAVs) face practical deployment challenges due to high computational costs or insufficient adaptability in complex environments. This work proposes a lightweight VLA training framework based on distillation from expert demonstrations, which, for the first time, leverages reinforcement learning and teleoperation-derived expert policies as supervision signals to fine-tune a unified language-conditioned navigation model. By eliminating dedicated control modules, the approach enables end-to-end semantic instruction-driven UAV control. Experimental results demonstrate that the fine-tuned model accurately executes diverse natural language commands and exhibits strong generalization, robustness, and adaptability in unseen object combinations and multi-object scenarios.

Abstract

Vision-language-action (VLA) models open a new path toward intuitive robot control by directly linking perception, language, and action in a single end-to-end framework. Yet for UAVs, practical adoption remains difficult because existing solutions are either computationally heavy or insufficiently capable in complex environments. In this work, we propose a practical expert-distillation pipeline (Exp2VLA) for language-conditioned drone navigation. The core idea is to distill expert behavior, obtained from reinforcement learning, teleoperation, or other controllers, into training data that can be used to fine-tune compact VLA models. This allows existing control strategies to be transferred into a unified language-guided navigation model, reducing manual system integration and lowering the barrier for deploying new robot behaviors. Experiments in both sim-to-sim and simulation-in-the-loop settings across multi-object scenes show that the fine-tuned models can handle varied semantic commands and generalize to unseen target compositions. The proposed framework demonstrates how expert-policy distillation can help mechatronic systems move from specialized control modules toward more flexible and reusable robot intelligence.

Problem

• Vision-Language-Action
• Drone Navigation
• Expert Demonstrations
• VLA Models
• Language-Guided Control

Innovation

• expert distillation
• vision-language-action
• language-conditioned navigation
• UAV control
• policy transfer