Home > Robotics > Surprisingly Robust In-Hand Manipulation: An Empirical Study

Surprisingly Robust In-Hand Manipulation: An Empirical Study

A recent paper published on arXiv.org proposes a robotic hand with dexterous in-hand manipulation skills. Contact-rich movements like finger-gaiting, pivoting, and the exploitation of gravity are achieved without sensing, hand or object models, or machine learning.

Robotic hand used in the study. Image credit: RBO TU Berlin (still image from the YouTube video)

A highly compliant hand demonstrates skills which transfer to objects of diverse shapes, weights, and sizes unmodified. The demonstrated skills go beyond the state of the art in their robustness, generality, and fluidity of motion.

Researchers identify principles that lead to robust in-hand manipulation. Firstly, contact dynamics should be transferred to the hand’s morphology. Also, the exploitation of constraints to limit the motion leads to the simplification of perception and control. Thirdly, it is possible to exploit the compositionality of manipulation funnels to produce complex manipulation programs.

We present in-hand manipulation skills on a dexterous, compliant, anthropomorphic hand. Even though these skills were derived in a simplistic manner, they exhibit surprising robustness to variations in shape, size, weight, and placement of the manipulated object. They are also very insensitive to variation of execution speeds, ranging from highly dynamic to quasi-static. The robustness of the skills leads to compositional properties that enable extended and robust manipulation programs. To explain the surprising robustness of the in-hand manipulation skills, we performed a detailed, empirical analysis of the skills’ performance. From this analysis, we identify three principles for skill design: 1) Exploiting the hardware’s innate ability to drive hard-to-model contact dynamics. 2) Taking actions to constrain these interactions, funneling the system into a narrow set of possibilities. 3) Composing such action sequences into complex manipulation programs. We believe that these principles constitute an important foundation for robust robotic in-hand manipulation, and possibly for manipulation in general.

Research paper: Bhatt, A., Sieler, A., Puhlmann, S., and Brock, O., “Surprisingly Robust In-Hand Manipulation: An Empirical Study”, 2022. Link: https://arxiv.org/abs/2201.11503
Link to the complete video playlist: https://youtube.com/playlist?list=PLb-CNILz7vmt6Ae_yD9i15TrCw0S8bKCn


Notify of
Inline Feedbacks
View all comments
Would love your thoughts, please comment.x