Source note2 min read
DreamFly: Causal Memory and Receding-Horizon Diffusion Planning for Aerial Vision-Language Navigation
Source
arXiv
arxiv.org/abs/2608.12308v1
Source
- Source name: arXiv
- Source kind: paper
- Authors / organization: Yan Deng, Fei Xu
- Venue: arXiv: cs.CV
- Published: 2026-08-12
- Collected: 2026-08-13
- URL: http://arxiv.org/abs/2608.12308v1
- PDF: https://arxiv.org/pdf/2608.12308v1
- DOI: UNKNOWN
- Registry group: academic_apis_and_preprints
- Evidence role: canonical_academic_index
- Language: en
- Country / region: UNKNOWN
Plain-Language Brief
The authors propose DreamFly, a framework for aerial vision-language navigation that uses a causally aligned historical memory to reason about the past without leaking future information. It also uses receding-horizon diffusion planning and LiteStop to improve performance on the OpenFly benchmark.
What the Document Says
- DreamFly is a framework for aerial vision-language navigation.
- It uses causally aligned historical memory.
- The historical memory augments the current visual representation using only observations preceding the current decision step.
- Navigation is formulated as receding-horizon diffusion planning, where the policy predicts a K-step action chunk but executes only the first action before replanning.
- It uses future actions as auxiliary planning targets while preserving closed-loop visual feedback.
- LiteStop estimates the stop probability directly from action logits at the initial all-mask state.
- Experiments on the OpenFly benchmark show consistent improvements in seen and unseen environments.
Key Concepts Mentioned
- aerial vision-language navigation
- DreamFly
- causally aligned historical memory
- receding-horizon diffusion planning
- LiteStop
- OpenFly benchmark
Claims Made by the Document
- DreamFly improves aerial vision-language navigation.
- It outperforms all compared methods on the OpenFly benchmark.
Evidence Used
- Experiments on the OpenFly benchmark demonstrate consistent improvements in seen and unseen environments.
- Results show that DreamFly achieves the lowest navigation error.
Methods / Approach
- causally aligned historical memory
- receding-horizon diffusion planning
- LiteStop
Limitations Stated by the Authors
Not stated in source
Limitations Noted by the Agent
Agent interpretation — not from the source.
The paper does not provide details about the compared methods or the specific navigation errors.
Open Questions
- What are the specific compared methods mentioned in the text?
- What is the exact definition of the navigation error?
Notes for Future Q&A
May include interpretation. Labeled where it does.
The paper does not provide detailed information about the compared methods or the specific navigation errors.
Raw Abstract / Excerpt
Aerial vision-language navigation (VLN) requires an embodied agent to integrate visual evidence over time, plan future actions, and determine when it has reached a navigation goal under partial observability. Although recent VLA models offer a promising perception-to-action paradigm, adapting them to aerial navigation remains challenging due to limited historical context, short planning horizons, and unreliable implicit termination. To address these challenges, we propose DreamFly, a diffusion-based aerial VLN framework built on Dream-VLA. DreamFly introduces a causally aligned historical memory that augments the current visual representation using only observations preceding the current decision step, enabling temporal reasoning without future information leakage. We further formulate navigation as receding-horizon diffusion planning, where the policy predicts a $K$-step action chunk but executes only the first action before replanning. This plan-$K$, execute-one strategy uses future actions as auxiliary planning targets while preserving closed-loop visual feedback. Finally, LiteStop estimates the stop probability directly from action logits at the initial all-mask state, decoupling explicit termination from action generation. Experiments on the OpenFly benchmark demonstrate consistent improvements in seen and unseen environments. DreamFly achieves 32.04%/29.46% SR and 28.22%/23.54% SPL on the test-seen/test-unseen splits, respectively, outperforming all compared methods on both metrics while attaining the lowest navigation error. These results demonstrate the effectiveness of jointly modeling historical context, future action structure, and explicit termination for aerial VLN.
