9Zeroing Neural Networks for Robot Arm Motion Generation

9.1 Introduction

In recent decades, robotics has received more and more attention in scientific areas and engineering applications. Many research studies have focused on this topic, and various kinds of robots have been developed and investigated [1, 15, 63, 6568, 98101]. For example, a multiple collaborative manipulators system is investigated in [101], where the velocity consensus state can be achieved with the aid of the presented consensus control method. Redundant manipulators have been playing an increasingly important role and are an appealing topic in engineering fields [1, 15, 63, 98100]. Redundant manipulators can achieve subtasks easily and dexterously such as repetitive motion and optimization of various performance criteria, since they possess more degrees of freedom than the minimum number required to execute a given primary task. One of the fundamental issues in operating such a robot system is the inverse‐kinematics problem (also termed the redundancy‐resolution problem). That is, with the desired trajectories of the manipulator's end‐effector being provided in Cartesian space, trajectories in joint space should be generated accordingly [1, 15, 63, 98100]. Rather than individual manipulators, multi‐manipulator systems are used in many scenarios in order to improve performance and reliability. For example, in applications such as exploration, surveillance, tracking, or payload transport, in order to ...

Get Kinematic Control of Redundant Robot Arms Using Neural Networks now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.