Tac-Man: Tactile-Informed Prior-Free Manipulation of Articulated Objects

2024-03-04

IEEE Transactions on robotics

AI Summary

Robust manipulation of articulated objects (e.g., doors, drawers) in human environments is hindered by the lack of kinematic priors, especially when object structure is unknown. Method: We propose a fully model-free, touch-driven manipulation paradigm grounded in high-resolution tactile sensing. Our approach establishes a closed-loop control framework integrating real-time contact-state estimation, adaptive force regulation, and model-free policy optimization—requiring no CAD models, joint parameters, or visual reconstruction. Contribution/Results: The core innovation replaces conventional kinematic modeling with high-fidelity contact modeling, enabling dynamic, stable interaction using tactile feedback alone. Evaluated on both physical robots and large-scale simulation, our method achieves >98% success rate—significantly outperforming state-of-the-art model-based and vision-based approaches. It demonstrates exceptional generalization and robustness to unknown objects, parametric uncertainties, and external disturbances.

Technology Category

• Intelligent Robots: Manipulation
• Humans and AI: Interaction Techniques and Devices
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Integrating robots into human-centric environments such as homes, necessitates advanced manipulation skills as robotic devices will need to engage with articulated objects such as doors and drawers. Key challenges in robotic manipulation of articulated objects are the unpredictability and diversity of these objects' internal structures, which render models based on object kinematics priors, both explicit and implicit, and inadequate. Their reliability is significantly diminished by pre-interaction ambiguities, imperfect structural parameters, encounters with unknown objects, and unforeseen disturbances. Here, we present a prior-free strategy, Tac-Man, focusing on maintaining stable robot-object contact during manipulation. Without relying on object priors, Tac-Man leverages tactile feedback to enable robots to proficiently handle a variety of articulated objects, including those with complex joints, even when influenced by unexpected disturbances. Demonstrated in both real-world experiments and extensive simulations, it consistently achieves near-perfect success in dynamic and varied settings, outperforming existing methods. Our results indicate that tactile sensing alone suffices for managing diverse articulated objects, offering greater robustness and generalization than prior-based approaches. This underscores the importance of detailed contact modeling in complex manipulation tasks, especially with articulated objects. Advancements in tactile-informed approaches significantly expand the scope of robotic applications in human-centric environments, particularly where accurate models are difficult to obtain.

Problem

• Managing articulated objects without prior kinematic models
• Handling unpredictability and diversity in internal object structures
• Overcoming pre-interaction ambiguities and unexpected disturbances

Innovation

• Tactile feedback enables prior-free manipulation
• Maintains stable contact during object interaction
• Handles diverse articulated objects without models