Application of Adam to Iterative Learning for an In-Hand Manipulation Task

Abstract

This paper proposes an iterative learning scheme for in-hand manipulation systems by utilizing the learning gain adaptation concept of deep learning. The advantages of the proposed method are that (1) there is no need to generate theoretical analytical models for the learning process and (2) the proposed method is robust against uncertainties such as measurement errors, friction force, and contact state. Finally, the validity of the proposed method is verified through experiments.