Neural network-Based hand posture control of a humanoid Robot Hand

Abstract

This paper presents a novel method to provide Neural Network Based Control of a Multi-finger Robot Hand. There are several challenges known from the literature that researchers are facing when they are trying to produce a human-like trainable robotic hand due to the complexity of building it and controlling its 3D movement. The authors of this article are providing an improved solution to this problem by developing a framework that enables easy training and control of a Robotic Hand by using Artificial Neural Networks.