Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

$π$, But Make It Fly: Physics-Guided Transfer of VLA Models to Aerial Manipulation

Johnathan Tucker, Denis Liu, Aiden Swann, Allen Ren, Javier Yu, Jiankai Sun, Brandon Kim, Lachlain McGranahan, Quan Vuong, Mac Schwager

Abstract

Vision-Language-Action (VLA) models such as $π_0$ have demonstrated remarkable generalization across diverse fixed-base manipulators. However, transferring these foundation models to aerial platforms remains an open challenge due to the fundamental mismatch between the quasi-static dynamics of fixed-base arms and the underactuated, highly dynamic nature of flight. In this work, we introduce AirVLA, a system that investigates the transferability of manipulation-pretrained VLAs to aerial pick-and-place tasks. We find that while visual representations transfer effectively, the specific control dynamics required for flight do not. To bridge this "dynamics gap" without retraining the foundation model, we introduce a Payload-Aware Guidance mechanism that injects payload constraints directly into the policy's flow-matching sampling process. To overcome data scarcity, we further utilize a Gaussian Splatting pipeline to synthesize navigation training data. We evaluate our method through a cumulative 460 real-world experiments which demonstrate that this synthetic data is a key enabler of performance, unlocking 100% success in navigation tasks where directly fine-tuning on teleoperation data alone attains 81% success. Our inference-time intervention, Payload-Aware Guidance, increases real-world pick-and-place task success from 23% to 50%. Finally, we evaluate the model on a long-horizon compositional task, achieving a 62% overall success rate. These results suggest that pre-trained manipulation VLAs, with appropriate data augmentation and physics-informed guidance, can transfer to aerial manipulation and navigation, as well as the composition of these tasks.

$π$, But Make It Fly: Physics-Guided Transfer of VLA Models to Aerial Manipulation

Abstract

Vision-Language-Action (VLA) models such as have demonstrated remarkable generalization across diverse fixed-base manipulators. However, transferring these foundation models to aerial platforms remains an open challenge due to the fundamental mismatch between the quasi-static dynamics of fixed-base arms and the underactuated, highly dynamic nature of flight. In this work, we introduce AirVLA, a system that investigates the transferability of manipulation-pretrained VLAs to aerial pick-and-place tasks. We find that while visual representations transfer effectively, the specific control dynamics required for flight do not. To bridge this "dynamics gap" without retraining the foundation model, we introduce a Payload-Aware Guidance mechanism that injects payload constraints directly into the policy's flow-matching sampling process. To overcome data scarcity, we further utilize a Gaussian Splatting pipeline to synthesize navigation training data. We evaluate our method through a cumulative 460 real-world experiments which demonstrate that this synthetic data is a key enabler of performance, unlocking 100% success in navigation tasks where directly fine-tuning on teleoperation data alone attains 81% success. Our inference-time intervention, Payload-Aware Guidance, increases real-world pick-and-place task success from 23% to 50%. Finally, we evaluate the model on a long-horizon compositional task, achieving a 62% overall success rate. These results suggest that pre-trained manipulation VLAs, with appropriate data augmentation and physics-informed guidance, can transfer to aerial manipulation and navigation, as well as the composition of these tasks.

Paper Structure

This paper contains 33 sections, 15 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Vision-Language-Action Policies for Manipulation
  4. Vision-Language and Vision-Language-Action Models for UAV Navigation
  5. Aerial Manipulation with Learned Policies
  6. Method
  7. System Overview
  8. Hardware
  9. Gripper Design
  10. Observation and Action Spaces
  11. Policy Architecture:
  12. Physics-Aware Guidance for Action Generation
  13. General Tracking-Error Guidance
  14. Payload-Aware Vertical Guidance
  15. Gaussian Splat Data Pipeline
  16. ...and 18 more sections

Figures (4)

  • Figure 1: Custom gripper installed on the Starling 2 Max drone. The design prioritizes low weight for extended flight time while maintaining sufficient grip strength for the target objects.
  • Figure 2: Examples of synthetic trajectories used to generate training data. Nominal rollouts and corrective rollouts overlay sampled trajectories in the gate-manipulation scene.
  • Figure 3: Out-of-distribution objects used for generalization probes. From left to right: goomba, box, chips, and sandwich. The policy was fine-tuned on a penguin but evaluated on these novel objects to test object-level generalization beyond training instances.
  • Figure 4: Trajectory visualization for the left gate task.The top three rows display real-world RGB observations from external, forward, and downward cameras. The bottom two rows show corresponding synthetic views rendered from Gaussian Splats.