Co-VLA: Coordination-Aware Structured Action Modeling for Dual-Arm Vision-Language-Action Systems

2026-06-18

AI Summary

This work addresses the lack of explicit coordination mechanisms in existing vision-language-action (VLA) models for tightly coupled dual-arm tasks, which undermines behavioral reliability, interpretability, and stability. To overcome this limitation, the authors introduce the Structured Action Experts (SAE) module—the first such integration within a VLA framework—employing shared and residual latent variables to model task-level coordination intent and per-arm execution adjustments, respectively. Coupled with a Latent-Aware Controller (LAC), the approach enables synchronized regulation of dual-arm behavior. The method supports real-time modulation of synchronization strength, execution asymmetry, motion smoothness, and safety constraints. Experiments demonstrate a 27% improvement in success rate on tightly coordinated tasks, a doubling of out-of-distribution performance in real-world scenarios (from 13% to 27%), and up to a 25% reduction in task completion time.

Abstract

Vision-language-action (VLA) models show strong capabilities in single and dual-arm robotic manipulation. Prior works show coordinated bimanual behaviors can emerge from end-to-end learning, leveraging large vision-language backbones with continuous action prediction. However, as bimanual tasks become tightly coupled and execution constraints become critical, implicit coordination alone is insufficient to ensure reliable, interpretable, and stable behavior. In this work, we propose Co-VLA, a coordination-aware bimanual manipulation framework introducing explicit structural priors into VLA models. We instantiate our method on a state-of-the-art vision-language backbone by replacing its monolithic action head with a Structured Action Expert (SAE) designed for bimanual coordination. Specifically, we introduce explicit structure at the action generation level with a modular coordination-aware loss that shapes shared and residual latents according to task-specific structures. The shared latent encodes task-level coordination intent, while residual latents capture execution adjustments for each arm. At deployment, a Latent-Aware Controller (LAC) interprets the learned representations to modulate synchronization strength, execution asymmetry, smoothness, and safety constraints in real time. LAC operates at the joint-command level and remains compatible with standard control pipelines without requiring force or impedance control. Experiments across simulation and real-world benchmarks show Co-VLA significantly outperforms monolithic baselines, achieving a 27% success rate gain in tight-coordination tasks, more than doubling performance in OOD real-world scenarios (from 13% to 27%), and reducing task completion time by up to 25%.

Problem

• bimanual manipulation
• coordination
• vision-language-action models
• structured action modeling
• dual-arm robotics

Innovation

• structured action modeling
• bimanual coordination
• vision-language-action
• latent-aware control
• modular coordination loss