Highly Dexterous Robot Hand Can Operate in the Dark — Just Like Us

Powered by DYNAMIXEL X-Series

Columbia Engineers design a robot hand that is the first device of its kind to join advanced sense of touch with motor-learning algorithms — it doesn’t rely on vision to manipulate objects

Dexterous Manipulation with Tactile Fingers


In this paper, we present a novel method for
achieving dexterous manipulation of complex objects, while
simultaneously securing the object without the use of passive
support surfaces. We posit that a key difficulty for training such
policies in a Reinforcement Learning framework is the difficulty
of exploring the problem state space, as the accessible regions
of this space form a complex structure along manifolds of a
high-dimensional space. To address this challenge, we use two
versions of the non-holonomic Rapidly-Exploring Random Trees
algorithm; one version is more general, but requires explicit
use of the environment’s transition function, while the second
version uses manipulation-specific kinematic constraints to attain
better sample efficiency. In both cases, we use states found via
sampling-based exploration to generate reset distributions that
enable training control policies under full dynamic constraints via
model-free Reinforcement Learning. We show that these policies
are effective at manipulation problems of higher difficulty than
previously shown, and also transfer effectively to real robots.
A number of example videos can also be found on the project
website: sbrl.cs.columbia.edu

All credit goes to Columbia University: Gagan Khandate∗†, Siqi Shang∗†, Eric T. Chang‡
, Tristan Luca Saidi†
, Yang Liu‡
Seth Matthew Dennis‡
, Johnson Adams‡
and Matei Ciocarlie‡
†Dept. of Computer Science ‡Dept. of Mechanical Engineering ∗
joint first authorship
Columbia University, New York, NY 10027, USA
Corresponding email: gagank@cs.columbia.edu

For Full Paper: https://arxiv.org/pdf/2303.03486.pdf