Performance-driven control of nano-motion systems

Abstract

The performance of high-precision mechatronic systems is subject to ever increasing demands regarding speed and accuracy. To meet these demands, new actuator drivers, sensor signal processing and control algorithms have to be derived. The state-of-the-art scientific developments in these research directions can significantly improve the performance of high-precision systems. However, translation of the scientific developments to usable technology is often non-trivial. To improve the performance of high-precision systems and to bridge the gap between science and technology, a performance-driven control approach has been developed. First, the main performance limiting factor (PLF) is identified. Then, a model-based compensation method is developed for the identified PLF. Experimental validation shows the performance improvement and reveals the next PLF to which the same procedure is applied. The compensation method can relate to the actuator driver, the sensor system or the control algorithm. In this thesis, the focus is on nano-motion systems that are driven by piezo actuators and/or use encoder sensors. Nano-motion systems are defined as the class of systems that require velocities ranging from nanometers per second to millimeters per second with a (sub)nanometer resolution. The main PLFs of such systems are the actuator driver, hysteresis, stick-slip effects, repetitive disturbances, coupling between degrees-of-freedom (DOFs), geometric nonlinearities and quantization errors. The developed approach is applied to three illustrative experimental cases that exhibit the above mentioned PLFs. The cases include a nano-motion stage driven by a walking piezo actuator, a metrological AFM and an encoder system. The contributions of this thesis relate to modeling, actuation driver development, control synthesis and encoder sensor signal processing. In particular, dynamic models are derived of the bimorph piezo legs of the walking piezo actuator and of the nano-motion stage with the walking piezo actuator containing the switching actuation principle, stick-slip effects and contact dynamics. Subsequently, a model-based optimization is performed to obtain optimal drive waveforms for a constant stage velocity. Both the walking piezo actuator and the AFM case exhibit repetitive disturbances with a non-constant period-time, for which dedicated repetitive control methods are developed. Furthermore, control algorithms have been developed to cope with the present coupling between and hysteresis in the different axes of the AFM. Finally, sensor signal processing algorithms have been developed to cope with the quantization effects and encoder imperfections in optical incremental encoders. The application of the performance-driven control approach to the different cases shows that the different identified PLFs can be successfully modeled and compensated for. The experiments show that the performance-driven control approach can largely improve the performance of nano-motion systems with piezo actuators and/or encoder sensors.