The control of a motion stage is key for the throughput and accuracy of many industrial systems. Examples of such systems are wafer stages, robotic arms, pick and place machines, printing systems, and many more. In order to achieve maximum performance, advanced control techniques are applied. Most systems operate in a repetitive manner, i.e. they repeat the same task over and over. For such systems, superior performance can be achieved by use of Iterative Learning Control (ILC). ILC is a control strategy which exploits the repetitive behavior of the system to learn the optimal command signal. However, the command signal is only optimal for one particular reference. As a result, the performance deteriorates dramatically when a different trajectory is tracked. This research focuses on enhancement of the extrapolation properties of ILC, while maintaining the same level of performance. In this research ILC is extended with rational basis functions to obtain excellent extrapolation properties. Several pre-existing approaches are analyzed and a new solution is proposed. The results are validated in simulation and applied on a complex industrial printing system from Océ.
Iterative Learning Control with a Rational Feedforward Basis: a new Solution Algorithm
van Zundert, J. (Author). 31 Aug 2014
Student thesis: Master