cs.RONov 11, 2026

What Makes a VLA Reactive? A Study of Execution Strategies for Dynamic Manipulation

Authors: Zhuoyue Li, Zhendong Han, Pankhuri Vanjani, Wenting Li, Rudolf Lioutikov

Organizations: Intuitive Robots Lab (IRL), Karlsruhe Institute of Technology (KIT) · Robotics Institute Germany (RIG)

Abstract

Vision-language-action (VLA) models predict chunks of future actions, and when objects move, those chunks go stale before they are executed. We study what makes a pretrained VLA reactive, using \pihalf{} on two dynamic simulation benchmarks and three real-robot tasks. We vary the prediction and execution horizons, compare synchronous and asynchronous inference, test methods that deliver actions sooner or keep successive chunks continuous, and adapt the policy with motion cues and latent-space reinforcement learning. The best prediction horizon is about one second, and the best execution horizon depends on the checkpoint. Short horizons help more with asynchronous execution, and methods that keep the trajectory continuous outperform those that only reduce latency. Adapting the policy repairs targeted failures, but on the conveyor task changing the execution horizon gained as much. Reactivity depends on the execution pipeline as much as on the policy, so both should be evaluated together.

Figures & tables

Appendix figures & tables5 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. React When You Need To: Event-Triggered Asynchronous Inference for VLA Policies

    Sep 18, 2026Yansong Wu, Huaqing Li, Tianding Hou +2Vision-Language-Action ModelsVLMs for Robotics

  2. Intercepting the Future: Latent-Space Predictive World Model for Dynamic VLA Manipulation

    Jun 1, 2026Shahram Najam Syed, Arthur Jakobsson, Haoran Hao +1World Models for RoboticsRobotic Manipulation

  3. Dynamic Execution Commitment of Vision-Language-Action Models

    May 12, 2026Feng Chen, Xianghui Wang, Yuxuan Chen +4VLM RobustnessEfficient VLM Inference