Skip to Main Content
In the vast majority of publications, it is noticeably claimed that parallel robots or manipulators are supposed to perform better than their serial counterparts. However, in practice, such mechanisms suffer from many problems, as theoretically provided potentials are difficult to exploit. This paper focuses on the issue of dynamics and control and provides a methodology to achieve accurate control for parallel manipulators in the range of high dynamics. The general case of a 6-DOF mechanism is chosen as the case study to substantiate the approach by experimental results. An important contribution is the emphasis on the structural properties of 6-DOF parallel robots to derive an appropriate and integrated control strategy that leads to the improvement of tracking performance by using only the available measurements of actuator positions. First, accurate and computationally efficient modeling of the dynamics is discussed. It is followed by presenting appropriate and optimal design of experimental parameter identification. The development of the control scheme begins with robust design of controller-observer for the single actuators. It is enhanced by a centralized feedforward dynamics compensation. Since systematic tracking errors always remain, a model-based iterative learning controller is designed to further increase the accuracy at high dynamics.