Application of a nonlinear adaptive controller to a 6 DOF parallel manipulator

Abstract

Nonlinear adaptive control schemes for robotic manipulators or new types of high speed machine tools with nonlinear dynamics are well suited to keep the tracking errors low, even at high speeds. Their application however requires fast microprocessors and an operating system that allows implementing the control algorithms as hard real-time tasks. The paper presents the implementation of a nonlinear model-based manipulator controller with adaptation algorithms on a new generation real-time operating system (RTOS) called XOberon. This RTOS goes beyond the established concepts, like interruptor priority-driven scheduling, bringing in new programming paradigms, such as deadline-driven scheduling. The superiority of this controller when applied to a fast 6 DOF parallel manipulator is finally shown by comparing its tracking performance with the one obtained using conventional linear controllers.