End-to-End Dexterous Arm-Hand VLA Policies via Shared Autonomy: VR Teleoperation Augmented by Autonomous Hand VLA Policy for Efficient Data Collection

TL;DR

This paper introduces a shared autonomy framework combining VR teleoperation and autonomous hand control to efficiently collect high-quality data for training dexterous manipulation policies, achieving high success rates with minimal human effort.

Contribution

It presents a novel shared control system for data collection and an end-to-end VLA policy with a new feature enhancement module for natural coordination.

Findings

Achieved 90% success rate across diverse objects.

Reduced human workload in data collection.

Validated effectiveness through comprehensive experiments.

Abstract

Achieving human-like dexterous manipulation remains a major challenge for general-purpose robots. While Vision-Language-Action (VLA) models show potential in learning skills from demonstrations, their scalability is limited by scarce high-quality training data. Existing data collection methods face inherent constraints: manual teleoperation overloads human operators, while automated planning often produces unnatural motions. We propose a Shared Autonomy framework that divides control between macro and micro motions. A human operator guides the robot's arm pose through intuitive VR teleoperation, while an autonomous DexGrasp-VLA policy handles fine-grained hand control using real-time tactile and visual feedback. This division significantly reduces cognitive load and enables efficient collection of high-quality coordinated arm-hand demonstrations. Using this data, we train an end-to-end VLA policy enhanced with our novel Arm-Hand Feature Enhancement module, which captures both distinct and shared representations of macro and micro movements for more natural coordination. Our Corrective Teleoperation system enables continuous policy improvement through human-in-the-loop failure recovery. Experiments demonstrate that our framework generates high-quality data with minimal manpower and achieves a 90% success rate across diverse objects, including unseen instances. Comprehensive evaluations validate the system's effectiveness in developing dexterous manipulation capabilities.