LiveVLN: Breaking the Stop-and-Go Loop in Vision-Language Navigation

TL;DR

LiveVLN is a framework that enables more continuous and smoother vision-language navigation by overlapping perception and action, significantly reducing idle waiting time during deployment.

Contribution

It introduces a training-free method to augment pretrained VLM navigators with multi-step action continuation for continuous online execution.

Findings

Reduces average episode waiting time by up to 77.7% in real-world deployments.

Shortens wall-clock episode time by 12.6% on StreamVLN and 19.6% on NaVIDA.

Preserves benchmark performance while enabling more continuous navigation.

Abstract

Recent navigation systems achieve strong benchmark results, yet real-world deployment often remains visibly stop-and-go. This bottleneck arises because the sense-inference-execution loop is still blocking: after each new observation, the controller must wait for sensing, transmission, and inference before motion can continue. Reducing action-generation cost alone therefore does not remove redundant waiting. To address this issue, we present LiveVLN, a training-free framework for more continuous embodied navigation by augmenting pretrained VLM navigators with multi-step action continuation. Instead of pausing for each full sense-and-inference round, LiveVLN overlaps execution with the processing of newly arrived observations, allowing refreshed future actions to be handed off before the current executable prefix is exhausted. This design keeps actions continuously available during motion, reducing idle waiting and enabling smoother online execution. The framework operates at runtime and can be integrated with compatible pretrained VLM navigators. Across R2R and RxR, LiveVLN preserves benchmark performance while reducing waiting time and improving action availability. In real-world deployments, it cuts average episode waiting time by up to $77.7\%$ and shortens wall-clock episode time by $12.6\%$ on StreamVLN and $19.6\%$ on NaVIDA, yielding more coherent execution during deployment. Code is available at https://github.com/NIneeeeeem/LiveVLN.