Estimating Deformable-Rigid Contact Interactions for a Deformable Tool via Learning and Model-Based Optimization

2025-05-16

AI Summary

This work addresses the challenge of modeling contact mechanics in interactions between soft tools and rigid objects. We propose a unified learning-and-optimization framework for deformable-rigid hybrid contact. Methodologically, it integrates data-driven joint estimation of contact forces and object motion with physics-based static equilibrium modeling. Crucially, we introduce the Contact Quadratic Program (CQP), the first formulation to explicitly encode Coulomb friction and static equilibrium constraints within a differentiable optimization layer. The framework leverages simulation-based pretraining followed by domain adaptation to ensure robust transfer to real-world settings. Experiments demonstrate significant improvements over existing baselines across diverse pushing and rotating tasks involving multi-material and multi-geometry rigid bodies. Validation on a physical soft-tool platform confirms high accuracy in predicting both contact forces and object motion, as well as strong generalization across manipulation tasks.

Technology Category

• Intelligent Robots: Learning & Optimization for ROB
• Search and Optimization: Learning to Search
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Dexterous manipulation requires careful reasoning over extrinsic contacts. The prevalence of deforming tools in human environments, the use of deformable sensors, and the increasing number of soft robots yields a need for approaches that enable dexterous manipulation through contact reasoning where not all contacts are well characterized by classical rigid body contact models. Here, we consider the case of a deforming tool dexterously manipulating a rigid object. We propose a hybrid learning and first-principles approach to the modeling of simultaneous motion and force transfer of tools and objects. The learned module is responsible for jointly estimating the rigid object's motion and the deformable tool's imparted contact forces. We then propose a Contact Quadratic Program to recover forces between the environment and object subject to quasi-static equilibrium and Coulomb friction. The results is a system capable of modeling both intrinsic and extrinsic motions, contacts, and forces during dexterous deformable manipulation. We train our method in simulation and show that our method outperforms baselines under varying block geometries and physical properties, during pushing and pivoting manipulations, and demonstrate transfer to real world interactions. Video results can be found at https://deform-rigid-contact.github.io/.

Problem

• Estimating deformable-rigid contact interactions for deformable tools
• Modeling motion and force transfer in deformable manipulation
• Recovering environment-object forces with quasi-static equilibrium constraints

Innovation

• Hybrid learning and first-principles approach
• Learned module estimates motion and forces
• Contact Quadratic Program for force recovery