Predefined-time convergence in task-based inverse dynamics using time base generators

In this paper, we propose a state-feedback controller that is capable to achieve predefined-time convergence for a class of nonlinear second order systems that can be linearized to a perturbed double integrator. The controller relies on the use of time base generators. Also, we study the link between the proposed predefined-time controller and the widely applied task-based inverse dynamics framework to control kinematically redundant robots. We show that positioning tasks assigned to the robot subject to joint-limits constraints are successfully completed in a user defined time regardless of the robots initial state. Moreover, the proposed task-based controller is robust to parametric uncertainties and matched disturbances since it includes a term given by a super-twisting control. This is verified through simulations with a kinematically redundant mobile-manipulator KUKA youBot with eight degrees of freedom. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.