AdaDexGrasp: Adaptive Dexterous Grasping via 3D Visuo-Tactile Representation Fusion
This work is significant for robotics researchers and engineers working on dexterous manipulation, as it offers a more robust and adaptive grasping solution by incorporating tactile feedback, addressing the limitations of vision-only approaches.
This paper addresses the challenge of replicating human-like stable and adaptive grasps in robotic systems by integrating visual perception and tactile feedback. The proposed method introduces a visuo-tactile representation that fuses object geometry with tactile signals, enabling contact-aware grasp pose generation and tactile-guided refinement, leading to significantly enhanced grasp success rates and generalization across diverse objects.
Humans achieve stable and adaptive grasps by seamlessly integrating visual perception and tactile feedback, a capability that remains challenging to replicate in robotic systems. Existing robotic grasping approaches predominantly rely on visual inputs and lack mechanisms for tactile-guided adaptation after contact, limiting robustness and generalization. To address this challenge, we propose a unified visuo-tactile-fusion grasping framework that integrates grasp generation, feasibility prediction, and adaptive refinement. At its core, our method introduces an efficient visuo-tactile representation that tightly fuses object geometry with tactile feedback by associating tactile signals with finger identities. This unified representation supports contact-aware grasp pose generation during planning and tactile-guided refinement after contact, enabling the system to reason about fine-grained finger-object interactions and adjust grasps dynamically. Comprehensive experiments in both simulation and real-world environments demonstrate that our approach significantly enhances grasp success rates and generalization across diverse objects.