Robust and Resilient Soft Robotic Object Insertion with Compliance-Enabled Contact Formation and Failure Recovery

2025-09-22

AI Summary

To address insufficient robustness in object insertion tasks caused by pose uncertainty and environmental disturbances, this paper proposes an autonomous recovery method based on a passively compliant soft wrist. Methodologically: (1) the soft wrist absorbs contact impact through elastic deformation, enabling a stepwise compliant contact sequence that progressively constrains degrees of freedom; (2) a pre-trained vision-language model is leveraged to diagnose failure modes and replan skills using only end-effector images and pose estimates—without requiring explicit force sensing or high-frequency closed-loop control; (3) the approach ensures safe, iterative trial-and-error execution. Experiments demonstrate an 83% success rate in simulation, with tolerance to ±5° grasping pose errors, ±20 mm hole localization errors, fivefold variations in friction, and unseen square/rectangular pegs. The method is further validated on a real robotic platform.

Technology Category

• Intelligent Robots: Manipulation
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Computer Vision: Biometrics, Face, Gesture & Pose

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Object insertion tasks are prone to failures under pose uncertainties and environmental variations, traditionally requiring manual finetuning or controller retraining. We present a novel approach for robust and resilient object insertion using a passively compliant soft wrist that enables safe contact absorption through large deformations, without high-frequency control or force sensing. Our method structures insertion as compliance-enabled contact formations, sequential contact states that progressively constrain degrees of freedom, and integrates automated failure recovery strategies. Our key insight is that wrist compliance permits safe, repeated recovery attempts; hence, we refer to it as compliance-enabled failure recovery. We employ a pre-trained vision-language model (VLM) that assesses each skill execution from terminal poses and images, identifies failure modes, and proposes recovery actions by selecting skills and updating goals. In simulation, our method achieved an 83% success rate, recovering from failures induced by randomized conditions--including grasp misalignments up to 5 degrees, hole-pose errors up to 20mm, fivefold increases in friction, and previously unseen square/rectangular pegs--and we further validate the approach on a real robot.

Problem

• Addressing object insertion failures under pose uncertainties and environmental variations
• Eliminating need for manual tuning or controller retraining in robotic insertion
• Enabling robust insertion with compliance-enabled contact formation and failure recovery

Innovation

• Passively compliant soft wrist for safe contact absorption
• Compliance-enabled contact formations to constrain degrees of freedom
• Vision-language model assesses failures and proposes recovery actions