AsyncVLA: An Asynchronous VLA for Fast and Robust Navigation on the Edge

TL;DR

AsyncVLA introduces an asynchronous control framework for robotic navigation that combines high-level vision-language reasoning with fast reactive actions, significantly improving success rates in real-world scenarios with communication delays.

Contribution

The paper presents AsyncVLA, a novel asynchronous control architecture that decouples semantic reasoning from reactive control, enabling robust and fast navigation on edge robots.

Findings

Achieves 40% higher success rate than baselines in real-world tests.

Effectively handles communication delays up to 6 seconds.

Bridges the gap between large model reasoning and real-time control.

Abstract

Robotic foundation models achieve strong generalization by leveraging internet-scale vision-language representations, but their massive computational cost creates a fundamental bottleneck: high inference latency. In dynamic environments, this latency breaks the control loop, rendering powerful models unsafe for real-time deployment. We propose AsyncVLA, an asynchronous control framework that decouples semantic reasoning from reactive execution. Inspired by hierarchical control, AsyncVLA runs a large foundation model on a remote workstation to provide high-level guidance, while a lightweight, onboard Edge Adapter continuously refines actions at high frequency. To bridge the domain gap between these asynchronous streams, we introduce an end-to-end finetuning protocol and a trajectory re-weighting strategy that prioritizes dynamic interactions. We evaluate our approach on real-world vision-based navigation tasks with communication delays up to 6 seconds. AsyncVLA achieves a 40% higher success rate than state-of-the-art baselines, effectively bridging the gap between the semantic intelligence of large models and the reactivity required for edge robotics.