LiMA cuts delay in robot hand tasks with fast and slow planning
LiMA: Bridging Long-term Imagination to Real-time Dexterous Manipulation via Asynchronous Diffusion
Robotics
Summary
Controlling robot hands to do tricky tasks needs both careful long-term planning and quick reactions. The authors created LiMA, a new way to split these two parts so the robot can plan slowly but act fast when needed. LiMA uses two systems working together asynchronously and a special math method to link slow plans with quick movements. This makes robot hands work better with less waiting time and higher success on complex tasks.
What this means in practice
- •For industrial robot engineers: Improve control of robotic hands performing complex assembly by reducing delay between planning and reacting.
- •For robotics software developers: Build responsive robot control systems that separate long-term action goals from fast motion adjustments.
Authors
Ning Chen, Yankai Fu, Junkai Zhao, Qianpu Sun, Guocai Yao, Pengwei Wang, Zhongyuan Wang, Shanghang Zhang
Abstract
Dexterous manipulation demands long-term foresight and rapid reactive control. Vision-Language-Action (VLA) models, while proficient in high-level reasoning, often lack a fine-grained understanding of physical dynamics and spatial perception. Conversely, World-Action Models (WAMs) typically suffer from high inference latency due to iterative generation. These deficiencies result in a critical temporal misalignment where the model's intent fails to adapt to rapid physical contact changes. To overcome this fundamental bottleneck, we propose LiMA, an asynchronous dual-system generative framework that systematically decouples intent planning from reactive execution. LiMA organizes computation into a multi-scale hierarchy: a slow system handles sparse long-horizon spatiotemporal intent generation, while a fast system focuses on dense high-frequency motion refinement. To align sparse intent predictions with dense action trajectories, we introduce a Latent Schrödinger Bridge Coupling mechanism that formulates refinement as an entropy-regularized probabilistic transport process. LiMA reduces inference latency by 45.8% compared with Cosmos-Policy via asynchronous decoupling. Evaluated across six bimanual dexterous manipulation tasks spanning multiple horizons, LiMA achieves an overall success rate of 70.8% and an average subtask success rate of 78.9%, while maintaining performance in unseen scenarios. The project website is available at https://ccdcs.github.io/LiMA_repo/