DAM-VLA: Decoupled Asynchronous Multimodal Vision Language Action model
For robotics manipulation, this work addresses the misalignment between synchronous VLA models and asynchronous real-world sensor streams, achieving large gains in success rate and control frequency.
DAM-VLA decouples temporal processing per modality in vision-language-action models, enabling each modality to update at its own sensor rate. Across seven real-world manipulation tasks, it achieves 95.2% average success rate, more than doubling the 40.95% of the strongest synchronous baseline, while sustaining 100 Hz control.
Vision-language-action (VLA) models inherit a shared synchronous clock from vision-language pretraining, processing every input at one rate. This is misaligned with physical interaction, where a high-frequency modality changes at hundreds of hertz, vision evolves more slowly, and language stays constant across an episode. A synchronous VLA oversamples slow modalities, undersamples fast ones, and caps action generation at the lowest effective frequency. We hypothesize that decoupling temporal processing per modality, letting each update and retain information at its own sensor rate, yields stronger representations and more robust control. We present DAM-VLA, which maintains per-modality latent buffers refreshed at sensor rates and read continuously by the action head, integrating new high-frequency modalities through gated cross-attention that leaves the pretrained backbone intact. Across seven contact-rich real-world manipulation tasks, DAM-VLA more than doubles the average success rate of the strongest synchronous baseline (95.2\% vs.\ 40.95\%) while sustaining smooth, reactive 100\,Hz control. Project website: \href{https://intuitive-robots.github.io/DAM-VLA/}{intuitive-robots.github.io/DAM-VLA/}