Robot Manipulator Redundancy Resolution

Chapter 15QP Based Joystick Control

15.1 Introduction

A robot manipulator is said to be redundant when it has more DOF than necessary to perform a given end-effector task [1, 107]. The fundamental issue to operate such a redundant system is the redundancy-resolution problem [1, 36, 254]. In recent decades, a favorable manner based on online optimization techniques has been applied to solving such redundancy-resolution problem [1, 61] and most of these optimization techniques can be formulated as a (time-varying) quadratic program (QP) subject to equality and inequality constraints [61, 90]. Furthermore, the constrained QP can be transformed into a system of piecewise-linear equations (PLE) [136, 168, 255] and then solved by a number of methods and techniques [1, 36, 61, 65, 97, 104, 128, 136, 168, 255, 256].

In this chapter, to achieve the real-time joystick control of the redundant manipulator, a cosine-aided position-to-velocity mapping subscheme is presented to map the position of the joystick in the motion range to generate the desired velocity of the robot end-effector in real time. On the other hand, a real-time joystick-controlled motion planning (JCMP) subscheme based on minimum-velocity-norm (MVN) is presented to resolve the velocity-specified redundancy-resolution problem of a joystick-controlled manipulator. In addition, such subschemes are implemented on the six-DOF planar robot manipulator hardware system presented in the previous chapters. By using the cosine-aided ...