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

• URL: http://arxiv.org/abs/2603.25038v1
• Date: Thu, 26 Mar 2026 05:19:54 GMT
• Title: $π$, But Make It Fly: Physics-Guided Transfer of VLA Models to Aerial Manipulation
• Abstract summary: We introduce AirVLA, a system that investigates the transferability of manipulation-pretrained VLAs to aerial pick-and-place tasks.<n>To bridge this "dynamics gap," we introduce a Payload-Aware Guidance mechanism that injects payload constraints directly into the policy's flow-matching sampling process.<n>We evaluate our method through a cumulative 460 real-world experiments which demonstrate that this synthetic data is a key enabler of performance.
• Score: 14.43820301560473
• License: http://creativecommons.org/licenses/by/4.0/
• 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.

Related papers

• Learning to Throw: Agile and Accurate Cable-Suspended Payload Delivery with a Quadrotor [17.676294984292433]
  Quadrotors offer the agility needed to rapidly transport suspended payloads during time-critical applications.<n>State-of-the-art approaches typically rely on model-based trajectory optimization and tracking.<n>We propose a hybrid simulation framework that couples a high-fidelity analytical quadrotor model with a physics solver for complex rope and payload interactions.
  arXiv  Detail & Related papers  (2026-06-25T23:30:52Z)
• Towards Precise Intent-Aligned VLA Aerial Navigation via Expert-Guided GRPO [10.368821738661707]
  We introduce an efficient reinforcement learning framework for VLA-based aerial navigation.<n>At its core, we propose EG-GRPO to augment online rollouts with few-shot expert data.<n>Across multiple tasks specified by complex human intents, EG-GRPO improves the success rate to 2.13x that of the SFT baseline, while improving intent alignment performance by 60.9%.
  arXiv  Detail & Related papers  (2026-06-01T14:31:35Z)
• EXPO-FT: Sample-Efficient Reinforcement Learning Finetuning for Vision-Language-Action Models [84.73890225707264]
  We present EXPO-FT, a system for stable, sample-efficient RL finetuning of pretrained VLA policies.<n>Our system achieves perfect task performance (30/30 successes) across all evaluated tasks within an average of 19.1 minutes of online robot data.<n>We release an open-source robustness with the aim of facilitating broader adoption of RL finetuning of VLA models in robotics.
  arXiv  Detail & Related papers  (2026-05-25T06:31:03Z)
• Towards Human-Like Manipulation through RL-Augmented Teleoperation and Mixture-of-Dexterous-Experts VLA [62.16042475700567]
  We introduce IMCopilot, a shared-autonomy assistant to simplify teleoperation data collection.<n>We present MoDE-VLA, an architecture that seamlessly integrates heterogeneous force and tactile modalities into a pretrained VLA backbone.<n>We validate our approach on four tasks of escalating complexity, demonstrating doubled success rate improvement over the baseline in dexterous contact-rich tasks.
  arXiv  Detail & Related papers  (2026-03-09T09:02:30Z)
• AIR-VLA: Vision-Language-Action Systems for Aerial Manipulation [15.824903784302562]
  Vision-Language-Action (VLA) models have achieved remarkable success in ground-based embodied intelligence.<n>Their application to Aerial Manipulation Systems (AMS) remains a largely unexplored frontier.<n>We propose textbfAIR-VLA, the first VLA benchmark specifically tailored for aerial manipulation.
  arXiv  Detail & Related papers  (2026-01-29T12:09:00Z)
• Steering Vision-Language-Action Models as Anti-Exploration: A Test-Time Scaling Approach [78.4812458793128]
  We propose textbfTACO, a test-time-scaling framework that applies a lightweight pseudo-count estimator as a high-fidelity verifier of action chunks.<n>Our method resembles the classical anti-exploration principle in offline reinforcement learning (RL), and being gradient-free, it incurs significant computational benefits.
  arXiv  Detail & Related papers  (2025-12-02T14:42:54Z)
• Align-Then-stEer: Adapting the Vision-Language Action Models through Unified Latent Guidance [63.33213516925946]
  We introduce textbfAlign-Then-stEer (textttATE), a novel, data-efficient, and plug-and-play adaptation framework.<n>Our work presents a general and lightweight solution that greatly enhances the practicality of deploying VLA models to new robotic platforms and tasks.
  arXiv  Detail & Related papers  (2025-09-02T07:51:59Z)
• AnyPos: Automated Task-Agnostic Actions for Bimanual Manipulation [24.199522837278128]
  We present a new notion of task-agnostic action paradigm that decouples action execution from task-specific conditioning.<n>ATARA is a scalable self-supervised framework that accelerates collection by over $ 30times $ compared to human teleoperation.<n>We propose AnyPos, an inverse dynamics model equipped with Arm-Decoupled Estimation and a Direction-Aware Decoder.
  arXiv  Detail & Related papers  (2025-07-17T03:48:57Z)
• Enhancing Training Data Attribution with Representational Optimization [57.61977909113113]
  Training data attribution methods aim to measure how training data impacts a model's predictions.<n>We propose AirRep, a representation-based approach that closes this gap by learning task-specific and model-aligned representations explicitly for TDA.<n>AirRep introduces two key innovations: a trainable encoder tuned for attribution quality, and an attention-based pooling mechanism that enables accurate estimation of group-wise influence.
  arXiv  Detail & Related papers  (2025-05-24T05:17:53Z)
• An Experimental Study on Exploring Strong Lightweight Vision Transformers via Masked Image Modeling Pre-Training [51.622652121580394]
  Masked image modeling (MIM) pre-training for large-scale vision transformers (ViTs) has enabled promising downstream performance on top of the learned self-supervised ViT features. In this paper, we question if the textitextremely simple lightweight ViTs' fine-tuning performance can also benefit from this pre-training paradigm. Our pre-training with distillation on pure lightweight ViTs with vanilla/hierarchical design ($5.7M$/$6.5M$) can achieve $79.4%$/$78.9%$ top-1 accuracy on ImageNet-1
  arXiv  Detail & Related papers  (2024-04-18T14:14:44Z)
• DiffsFormer: A Diffusion Transformer on Stock Factor Augmentation [36.75453713794983]
  We introduce the Diffusion Model to generate stock factors with Transformer architecture (DiffsFormer) When presented with a specific downstream task, we employ DiffsFormer to augment the training procedure by editing existing samples. The proposed method achieves relative improvements of 7.2% and 27.8% in annualized return ratio for the respective datasets.
  arXiv  Detail & Related papers  (2024-02-05T03:54:36Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.