Dexterous Manipulation using Tactile Feedback

Humans manipulate objects dexterously, even when the objects are unknown and are in cluttered environments. This adaptive and versatile dexterity could not be achieved without numerous tactile receptors in the skin. TML aims to develop human-like robotic tactile skins and enable adaptive robotic dexterity.

Multimodal tactile sensor for in-hand sliding manipulation

We designed a tactile sensor that measures both local normal and shear forces by using its “nib” arrays. Modulating the sampling rate, the sensor measures static forces as well as dynamic directional slip events simultaneously. Using this sensor, the robot successfully reorients objects with unknown surface properties and weights.

Related Paper: Huh, Tae Myung, Hojung Choi, Simone Willcox, Stephanie Moon, and Mark R. Cutkosky. “Dynamically reconfigurable tactile sensor for robotic manipulation.” IEEE Robotics and Automation Letters 5, no. 2 (2020): 2562-2569. (URL)