X-Fly: Learning to Reach and Stop in Full-Chunk Aerial Vision-Language Navigation
X-Fly: Learning to Reach and Stop in Full-Chunk Aerial Vision-Language Navigation
This research presents X-Fly, a staged training framework for full-chunk aerial vision-language navigation. By executing complete predicted action chunks before replanning, the method reduces model inference time from 195 ms to 42 ms per waypoint and improves recovery from non-expert states. The approach combines Recovery DAgger labeling with Flow-GRPO refinement to jointly optimize a chunked flow controller and a Bernoulli STOP policy. On the OpenFly test-seen dataset, X-Fly achieves a 74.18% success rate compared to 16.45% for full-chunk behavior cloning and 32.04% for DreamFly.
The paper introduces X-Fly, a staged training framework for aerial vision-language navigation that uses full-chunk action execution to reduce inference latency and improve success rates on test-seen and test-unseen datasets compared to baseline methods.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
state. Flow-GRPO then jointly refines the chunked flow controller and a Bernoulli STOP policy from same-task closed-loop outcomes, supplemented by outcome-verified stopping supervision. With eight controls per decision, X-Fly reaches 74.18% SR on test-seen and 51.73% on test-unseen, compared with 16.45% and 18.43% for full-chunk BC and 32.04% and 29.46% for DreamFly.
Open source S4
supervise neither recovery from these states nor termination at the learner's own arrival. On OpenFly test-seen, no-stop failures account for 47.78% of full-chunk behavior cloning (BC) outcomes. We present X-Fly, a staged training framework for full-chunk aerial VLN. Recovery DAgger labels learner-visited decision states with smooth chunks that reconnect to future expert segments and with STOP at the first in-region
Open source S3
arXiv:2610.21450v2 Announce Type: replace Abstract: Aerial vision-language navigation (VLN) agents built on large vision-language models are expensive to query. Executing a complete predicted action chunk before replanning amortizes this cost, reducing model inference from 195 to 42 ms per executed waypoint for our policy. It also moves the states at which the policy is queried away from expert demonstrations, which
Open source S2
Source：arXiv Robotics — research abstracts · arxiv.org