Abstract
Flow-based vision-language-action (VLA) models are highly effective for generalist robot manipulation, yet their reliance on computationally expensive VLM encoding and multi-step iterative action generation imposes a significant latency bottleneck. The resulting inference latency makes it difficult for robots to respond quickly, especially in dynamic environments. We address this limitation with RAVEL (Rolling Asynchronous VLA Enabling Low-Latency Control), an asynchronous inference framework that addresses the computational bottlenecks of both the VLM backbone and the action expert. To reduce the delay from multi-step action denoising, RAVEL allows near-term actions to be executed after a single denoising step by carrying partially denoised future actions forward in a rolling buffer. To avoid blocking on slow VLM encoding, RAVEL decouples VLM encoding from rolling action generation, allowing the action expert to operate continuously using the latest available VLM context, while a lightweight Fast Observation Pathway (FOP) directly conditions the action expert on current observations. Across simulated and real-world manipulation tasks, RAVEL consistently achieves substantially lower response latency while maintaining the task capability of the underlying VLA, enabling high-frequency and responsive closed-loop control.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Chen, Y., Yu, K., Liu, P., Wang, S., Yang, Y., & Zhu, L. (2026). RAVEL: Asynchronous Rolling Inference for Flow-Based Vision-Language-Action Models. https://omanscience.com/en/articles/ravel-asynchronous-rolling-inference-for-flow-based-vision-language-action-models
MLA 9
Chen, Yuhan, et al. "RAVEL: Asynchronous Rolling Inference for Flow-Based Vision-Language-Action Models." https://omanscience.com/en/articles/ravel-asynchronous-rolling-inference-for-flow-based-vision-language-action-models.
Chicago (author–date)
Chen, Yuhan, Ke Yu, Pengfei Liu, Shuxun Wang, Yi Yang, and Linchao Zhu. 2026. "RAVEL: Asynchronous Rolling Inference for Flow-Based Vision-Language-Action Models." https://omanscience.com/en/articles/ravel-asynchronous-rolling-inference-for-flow-based-vision-language-action-models.
Harvard
Chen, Y., Yu, K., Liu, P., Wang, S., Yang, Y. and Zhu, L. (2026) 'RAVEL: Asynchronous Rolling Inference for Flow-Based Vision-Language-Action Models', Available at: https://omanscience.com/en/articles/ravel-asynchronous-rolling-inference-for-flow-based-vision-language-action-models.
Vancouver
Chen Y, Yu K, Liu P, Wang S, Yang Y, Zhu L. RAVEL: Asynchronous Rolling Inference for Flow-Based Vision-Language-Action Models. https://omanscience.com/en/articles/ravel-asynchronous-rolling-inference-for-flow-based-vision-language-action-models
IEEE
Y. Chen, K. Yu, P. Liu, S. Wang, Y. Yang, and L. Zhu, "RAVEL: Asynchronous Rolling Inference for Flow-Based Vision-Language-Action Models," https://omanscience.com/en/articles/ravel-asynchronous-rolling-inference-for-flow-based-vision-language-action-models.