A Soft Wrist with Anisotropic and Selectable Stiffness for Robust Robot Learning in Contact-rich Manipulation

2026-02-16

AI Summary

This work addresses the challenges of contact-intensive robotic manipulation—namely, frequent damage from collisions and difficulty in policy learning—by introducing CLAW, a compliant wrist that integrates orthogonal leaf springs with a rotation joint featuring a mechanical locking mechanism. For the first time, this design combines mechanical locking with an orthogonal leaf-spring architecture to achieve large six-degree-of-freedom deformations (40 mm lateral, 20 mm vertical) and three-level tunable anisotropic stiffness, all within a lightweight and low-cost framework. Experimental results demonstrate that CLAW achieves a 76% success rate in peg-insertion tasks, significantly outperforming both the Fin Ray gripper (43%) and a rigid gripper (36%), while effectively enabling complex contact-rich operations such as precision assembly and manipulation of fragile objects.

Technology Category

• Intelligent Robots: Manipulation
• Computer Vision: Low Level & Physics-based Vision
• Humans and AI: Interaction Techniques and Devices

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Contact-rich manipulation tasks in unstructured environments pose significant robustness challenges for robot learning, where unexpected collisions can cause damage and hinder policy acquisition. Existing soft end-effectors face fundamental limitations: they either provide a limited deformation range, lack directional stiffness control, or require complex actuation systems that compromise practicality. This study introduces CLAW (Compliant Leaf-spring Anisotropic soft Wrist), a novel soft wrist mechanism that addresses these limitations through a simple yet effective design using two orthogonal leaf springs and rotary joints with a locking mechanism. CLAW provides large 6-degree-of-freedom deformation (40mm lateral, 20mm vertical), anisotropic stiffness that is tunable across three distinct modes, while maintaining lightweight construction (330g) at low cost ($550). Experimental evaluations using imitation learning demonstrate that CLAW achieves 76% success rate in benchmark peg-insertion tasks, outperforming both the Fin Ray gripper (43%) and rigid gripper alternatives (36%). CLAW successfully handles diverse contact-rich scenarios, including precision assembly with tight tolerances and delicate object manipulation, demonstrating its potential to enable robust robot learning in contact-rich domains. Project page: https://project-page-manager.github.io/CLAW/

Problem

• contact-rich manipulation
• robot learning robustness
• soft end-effectors
• anisotropic stiffness
• unstructured environments

Innovation

• anisotropic stiffness
• soft wrist
• compliant mechanism
• contact-rich manipulation
• robot learning