Kinematic Control of Redundant Robot Arms Using Neural Networks

8Neural Network Based Learning and Control Co‐Design for Stewart Platform Control

8.1 Introduction

Kinematically redundant manipulators [39] are those manipulators that prove to have sufficiently higher degrees of freedom (DOFs) than required for positioning and for orientation of the platform. With the advances in the field of robotic technologies, robotic manipulators are widely used in the applications of factory automation which are required to carry out continuous and delayed work, such as lifting and transporting radioactive substances and executing the work in hazardous, scattered or packed environments. In comparison with nonredundant manipulators, redundant ones offer extra DOFs [75], and are often used to improve the dexterity, in order to work efficiently by avoiding collisions with obstacles. Research in the field of kinematically redundant manipulators has gained popularity due to their ability to avoid obstacles, their internal singular configurations, and their ability to optimize the performance of the workspace and the end‐effector motion task. Among various types of redundant manipulators, parallel ones, which usually feature higher rigidity, higher precision and higher response speed than serial ones, have received popular applications in flight simulators, electrostatic magnetic lenses, etc. However, how to efficiently control the motion of redundant parallel manipulators, especially in the situation with parameter uncertainties, or even unknown, is of ...