A Humanoid Visual-Tactile-Action Dataset for Contact-Rich Manipulation

Abstract--Contact-rich manipulation has become increasingly important in robot learning. However, previous studies on robot learning datasets have focused on rigid objects and underrepre-sented the diversity of pressure conditions for real-world manipulation. T o address this gap, we present a humanoid visual-tactile-action dataset designed for manipulating deformable soft objects. The dataset was collected via teleoperation using a humanoid robot equipped with dexterous hands, capturing multi-modal interactions under varying pressure conditions. Contact-rich interaction represents a critical gateway for enabling robots to perform complex tasks in real-world environments, yet it remains one of the fundamental challenges in robotic manipulation [1].