Cartesian nonlinear model predictive control of redundant manipulators considering obstacles

Abstract

For redundant fixed or mobile manipulators in shared human-robot workspaces, control algorithms are necessary that allow the robot to perform a task defined in the Cartesian space and that simultaneously realize additionally desired robot behaviors like avoiding collisions with humans or other obstacles. In this contribution, a Nonlinear Model Predictive Control (NMPC) approach is proposed to move the end-effector of the robot to a desired pose (3D position and orientation), along a geometric reference path or along a reference trajectory while considering obstacles. Due to the underlying general robot model, the control algorithm is applicable to both fixed and mobile manipulators, which is shown by means of simulations of a 7 DoF fixed manipulator and of a 10 DoF mobile manipulator.